University of Notre Dame

CBE 358 (Junior Class) Chemical & Biomolecular Engineering Laboratory I

LABORATORY MANUAL Spring 2014

Salma Saddawi
ssaddawi@nd.edu

(574) 631-3324

TABLE OF CONTENTS Page I. GENERAL INFORMATION A. Safety B. Laboratory Format and Procedures 1. 2. 3. 4. 5. 6. 7. 8. 9. Organization of Student Groups and Laboratory Projects Laboratory Session 1 Laboratory Session 2 and the Progress Report Session 3 Final Project Report and Oral Defense. Oral Presentations Laboratory Notebook Student Responsibilities in the Laboratory Grading 2 2 2 2 2 3 3 3 3 3 4 4 4 4 5 5 5 5 5 5 6 6 6 6 6 7 7 9 142 142 146

II. Guidelines for the Preparation of Written Reports 1. Title Page 2. Abstract 3. Table of Contents 4. Introduction 5. Theory 6. Experimental A. Apparatus B. Procedure 7. Results 8. Discussion 9. Conclusions and Recommendations 10. Literature Cited 11. Nomenclature 12. Appendices III PROJECT STATEMENT (Page 9 gives page reference to the individual statements.) IV. Appendix A A NOTE ON STATISTICS Sample Output

2 I. GENERAL INFORMATION A. Safety Laboratory safety is the top priority and this requires all people in the lab to be observing safe practices at all times! Safety glasses must always be worn by everyone in the laboratory. Make sure you understand how the experimental apparatus works and what all of the adjustments do before you attempt to operate it. Be sure you have asked, and received an answer, from the Professor or the TA, about any possible hazards related to your experiment before attempting to operate it. Care must be used in the handling of chemicals to avoid spills and to avoid contact with the skin. B. Laboratory Format and Procedures 1. Organization of Student Groups and Laboratory Projects Students will organize into groups of three persons. Each group is to perform four projects during the semester. (A roster of the groups and a schedule of projects will be supplied separately.) A group leader, who is in charge of directing the work for the lab, should be selected by, and from among, the members of the group. (This responsibility should rotate among the members.) All group members must be prepared for the laboratory and contribute equally to the laboratory work and preparation of the reports. However, the group leader is in charge of assigning and coordinating tasks for the laboratory period and maintaining the group notebook. He or she is ultimately responsible for making sure that everything is done to ensure a successful experiment. Each project consists of three laboratory sessions and a series of three reports --a preliminary report, a progress report and a final report. A description of each of these components follows. 2. Laboratory Session 1 At the beginning of the first session for a given experiment, a paragraph describing the experimental plan and procedure should be submitted to the TA who is in charge at that

3 time. A discussion between the TA and the students will take place to ensure that students have an accurate plan of action.

3. Laboratory Session 2 A one page progress report of results obtained in the first session, and plans for the second session are due at this time. Graphs and tables of results obtained should be appended. Failure to submit the report will result in a 10% reduction of the final report grade. The progress report should be submitted to the TA in charge of the experiment. A summary discussion of the report with the instructor will be conducted in the laboratory. 4. Session 3 No specific format or report requirements are specified for this session. The laboratory will be open to gather additional data if needed. The instructor will be available for consultations during the first hour of the laboratory period. During this session examination of the experimental apparatus for the next assigned project should be performed by each group. 5. Final Project Report and Oral Defense. The final project report is due at the beginning of the next scheduled laboratory period following Session 3. There are no exceptions to this deadline. The reports are to be submitted to one of the department secretaries in the Chemical Engineering office. During the week following the day on which the final report was submitted, the group should schedule a meeting with the instructor for an oral defense/discussion of the written report. Each member of the group should be prepared to defend and/or discuss any part of the final report. 6. Oral Presentations For the last project of the semester, each group will make a formal 20-minute oral presentation in addition to the final report on that project. The oral presentations, which will take place during the final scheduled laboratory period, will be followed by a 10minute question-and-answer session. The location and other specifics will be announced later in the semester. All students are required to be present for all presentations by the groups in their section.

7. Laboratory Notebook Part of the purpose of the chemical engineering laboratories is to learn good laboratory and research practices. An important aspect of this is safety. Another important aspect is record-keeping and documentation. In industry you will find that all experiments have to be carefully recorded in an official laboratory notebook and signed by the investigator on a daily basis. To help foster these professional practices, each group is required to keep a laboratory notebook documenting the group's work. Each group should purchase the Engineering & Science Notebook, available at the Bookstore. In the notebook should be kept a neat, labeled and dated record of all work associated with the experiment, including a copy of the precis, all raw data, the settings on the experimental controls, any problems encountered in the experiment and what was done to fix them and why, all calculations, a copy of your progress report, etc. The laboratory notebooks will be handed in at the end of the semester and will contribute to the laboratory participation portion of your grade.

8. Student Responsibilities in the Laboratory Condition of Working Area. Students are responsible for the condition of their working area at the end of each laboratory period. All power to the equipment and instruments should be turned off, and steam and cooling water flows should be shut off. Glassware used should be cleaned and dried. Any equipment or instrumentation malfunctions should be reported promptly to the instructor or the TA. Checkout before Leaving Laboratory. The students must have their notebooks initialed by a faculty member or the teaching assistant prior to leaving at the end of the laboratory period. At that time the faculty member or teaching assistant will check the working area and take information about any equipment or instrumentation problems. 9. Grading Report grading is done by the Professors who are in charge of a given experiment. This grade will be based on the written report, the oral defense and other pertinent factors (e.g., if you are totally unprepared to do an experiment, you will be docked.) Grades for this course will be determined by the grades on the four experiments as well as your laboratory participation. The laboratory participation portion of your grade in will include how well you followed laboratory safety guidelines (did you wear safety glasses at all times in the lab? did you follow the special safety precautions required for each experiment?), attendance, tardiness, participation, professionalism, how effective a group

5 leader you were, and the quality of your laboratory notebook. Both laboratory instructors and teaching assistants will contribute towards this portion of your grade. II. Guidelines for the Preparation of Written Reports A technical report is a medium commonly used by scientists and engineers to communicate the results of their work. Frequently the report is the only tangible product and thus the only evidence for evaluation of the work. Consequently, it deserves careful attention to quality, packaging, and distribution. It is important that the writer(s) of an engineering report keep in mind the needs and interests of the anticipated readers of the report. The laboratory report should be written with the same professionalism that would be used to present the results of a major industrial project. The people who will read it, and need to draw conclusions from it, can be expected to have technical training, but probably would not be familiar with the details of the work. A good report of technical work quantitatively states significant results of experiments and computations and explains how they were obtained, what they mean, and how they are useful. The report should be clear, concise, and accurate. Often the structure of the report must conform to specific conventions. A format for laboratory reports that is to be used in this course is given below. 1. Title Page The title of the report is followed by names of the authors and laboratory group, the date of submission, and identification of the institution or organization supporting the work (University of Notre Dame, Department of Chemical Engineering, CHEG 358). 2. Abstract The abstract is a tightly written summary, typically 100 to 300 words long. This section is important because it is the first impression your report will make to a reader, and it could very well be the only part of it he or she will read! (Because of its importance, it is a significant part of the overall grade.) The abstract should be written as stand alone section of just text. Its independence means that the use of symbols, tables, and graphs as well as literature references should be avoided. A good abstract states the principal objective of the investigation, describes the methodology used and summarizes the results and conclusions in statements as quantitative and as general as possible. The abstract should provide ranges of the experimental parameters (e.g. the Reynolds number was varied from 100 to 10000), report the most important results and state how these values compare to expected (i.e., literature) ones. (e.g., values for the friction factor in the laminar flow regime were consistently 15% higher that the predictions of Poiseuille flow). If the value of a single variable or a short list of numbers is given, the

6 numbers should give the uncertainty (e.g., solubility at 25C was 25 3 moles/liter) and of course units should be included. We recommend that you write the abstract last, when your thoughts are most clearly in focus (i.e., you know all the answers and thus know what to say!). 3. Table of Contents A Table of Contents should be included in the report, including a listing of the Abstract. Appendices should also be listed. All pages should be numbered, including tables, figures, and appendices. 4. Introduction The purpose of the Introduction is to place the work in the perspective of prior work including key literature references, demonstrate its importance, and state the specific objectives. The Introduction should not exceed two pages.

5. Theory This section is a short, concise statement of the essential empirical and theoretical relations to be used in interpreting the data or to be tested by the data. Equations are usually stated with a reference, along with the pertinent assumptions and limitations. Brief manipulations may be appropriate, but long derivations are relegated to an appendix. The physical significance of equation parameters should be pointed out. 6. Experimental A. Apparatus The objective of this section of the report is to describe the experimental set-up in enough quantitative detail to enable the reader to completely understand the experiment. Ranges of independent variables are cited. The model and supplier of any unique equipment should be cited. Also, a schematic diagram of the experimental apparatus should be included. B. Procedure The objective of this section of the report is to describe the materials and methods used to obtain the experimental data. Emphasis is placed on general procedures that are not routine

7 7. Results The data, or a representative fraction of them, must be included in this section. They should be presented graphically. If there are only a few (i.e. 2-3) numbers, these could put into a table if they can be understood. Data are often not presented in raw form, but are reduced and shown in the way most clearly supporting the conclusions. Representation of scatter in data is essential. For example, the experimentally determined heat transfer coefficient is meaningless unless it is accompanied by units and an estimated uncertainty. Comparison of data with theoretical predictions and/or previously published values should be included whenever possible. This may require searching for information in reference books or research articles. Comment briefly on unique aspects of the results, in particular its accuracy. Also comment on the range of the variables covered. Each graph or diagram is assigned a number (e.g., Figure 1) and should have a caption that is descriptive of the information contained in the figure. A restatement of the information on the axes is not an acceptable title. 8. Discussion All important interpretations which follow from the results and the underlying theory are logically and quantitatively compared in the Discussion section. The positive conclusions, comparison with literature data, and the significance applicability, and reproducibility of the results are stressed. Quantitative statements about the accuracy and precision of the results are required. However, when a detailed error analysis is essential to the work, it should be relegated to an Appendix.

9. Conclusions and Recommendations This section is a summary of the most significant conclusions developed in the preceding section. Quantitative statements are best. Useful recommendations to improve the experiment and to extend the work to other systems, should be included here. 10. Literature Cited Only references cited in the report are to be listed is this section since it is not a bibliography covering all references but only the most pertinent ones. Footnotes on individual pages of the report are not to be used. References cited in the text of the final project report should give the last name of the author (both authors when only two; first author et. al. when more than two) and the corresponding page numbers. An example is given below.

8 The Reynolds number can be interpreted as the ratio of inertial to viscous forces at work in the fluid (Denn 37-39). References are to be listed in alphabetical order according to author or equivalent and should not be numbered. Use Chemical Abstracts Service Source Index journal abbreviations. For the previous example the citation would be the following: Denn, M. M. Process Fluid Mechanics; Prentice-Hall; New Jersey, 1980 Typical citations for a journal are given below. Danckwerts, P. V.; Sharma, M. M. Chem. Eng. (London) 1966, 202, 244. Danckwerts, P. V. Chem. Eng. Sci. 1979, 34, 443. 11. Nomenclature Symbols used in the report are defined immediately after they are presented the first time. This section of the report lists all of the symbols used. Units should be included. 12. Appendices The appendices contain material of secondary importance: sample calculations (a sample of all calculations done for the experiment must be included in the report), error analysis, derivation of theoretical relations, and perhaps graphs, calibration curves and/or schematics. Note that the appendices should be named in the order of which they appear in the final project report. In other words Appendix A should be the first appendix referred to in the text of the report.

9 III PROJECT STATEMENT 1. FM = Pipe Flow 2. HT2 = Heat Exchanger 3. HT1 = Natural Convection 4. MT = Diffusivity Measurement 5. MT2 = Leaching Rate Measurement 6. TD1 = Phase Equilibrium 7. FM2 = Fluidization and fluid Bed Heat Transfer 8. TD2 = Liquid mixture viscosities 10 18 36 44 50 56 72 86

9. HT3 = Unsteady State Heat Transfer Thermal Diffusivities And Heat Transfer Coefficients 93 10. PI-plus-Feedforward Water Level control for Coupled Water Tanks 109

10 EXPERIMENT FM1 FLOW IN CIRCULAR PIPES OBJECTIVE To measure the pressure drop in the straight section of smooth, rough, and packed pipes as a function of flow rate, to correlate this in terms of the friction factor and Reynolds number and to compare results with available theories and correlations. To determine the influence of pipe fittings on pressure drop. To determine the discharge coefficients of the Orifice plate and the Vernturii meter. APPARATUS Pipe Network Rotameters Manometers The pipe network consists of seven flow loops. The first three flow loops are smooth round pipes with inside diameters (ID) of 0.269", 0.622" and 0.824". The fourth loop is 0.824" ID and has rough walls. The fifth loop is 0.824" ID. In it there is a 12" long section packed with 0.92 mm spheres that is 0.8" ID. The last two loops have a gate valve, a globe valve, a strainer, an orifice plate, and a venturi meter. The flow of distilled water is directed through a particular loop by valves located at the entrance to each loop. The flow rate through the system is controlled by adjustments to the valve in front of the rotameters. The two rotameters allow for a wide range of flow rates. The pressure differences are measured with heavy liquid manometers and electronic manometers. The ends of the pressure taps are connected to the manometers and to different locations along the pipe resulting in varying lengths. Note: Pipe Network diagrams are included. THEORETICAL DISCUSSION a. Pipe Flow Fluid flow in pipes is of considerable importance in process plants, long distance pipelines and has applications to circulation systems in animals and plants. Fluids could be, for example, a single phase liquids or gases, mixtures of gases, liquids and solids, nonNewtonian fluids such as polymer melts, mayonnaise or potato salad. For any of these flows a key issue is the relation between flow rate and pressure drop. Recall from CBE 355 that for a Newtonian fluid such as water, which will be used in this experiment, the Navier-Stokes equations will govern the flow field. If the flow is steady, fully developed, laminar (straight streamlines), the velocity distribution and the average velocity are (Bird, Stewart, Lightfoot, 2002)

11

vz v z,max

vz # r &2 1 = 1 " % ( and = $ R' v z,max 2

(1)

then, the Navier-Stokes equations can be solved exactly to obtain the Hagen-Poiseuille relation (Middleman, 1998),!

Q=

!R 4 "P 8 L

(2)

where Q is the volumetric flow rate, R is the tube radius, is the dynamic viscosity, and !P/L is the pressure drop. It will be interesting to verify this result with experiments. At higher flowrates, the streamlines are not steady and straight and the flow is not laminar. Generally, the flow field will vary in both space and time with fluctuations that comprise "turbulence". For this case almost all terms in the Navier-Stokes equations are important and there is no simple solution. Fortunately dimensional analysis of the pertinent variables (Middleman, p233), !P = F(D, , !, L, U, plus some further arguments about a fullydeveloped flow not depending on distance, L, yields an expected relation between two dimensionless groups. The first is the Reynolds number,

Re !

DU"

(3)

which gives the ratio of inertial to viscous forces. The second one is the friction factor,

f !

"P D L 2 #U2

(4)

relating pressure drop to inertia forces. The expected function is thus

f = F( Re)
For laminar flow it can be shown that equation (1) yields

(5)

f =

16 Re

" Re < 2100 stable % $ ' #Re > 2100 unstable &

(6)

For turbulent flows, the appropriate function is obtained through experiments. A relation that fits data reasonably well is (Bird, Stewart, Lightfoot, 2002) !

f = 0.079 Re !0.25 2.1*103 < Re < 105

(7)

12 If the pipe is not smooth, there is (at least) one additional length scale, ", that is needed to characterize the roughness. This leads to an additional dimensionless group, "/D, and thus more than one curve on the friction-Reynolds number plot (see Bird, Stewart, Lightfoot, 2002 page 182). b. Flow through fittings. Any interesting piping system will have a variety of fittings. These could include valves, couplings, expansions, and contractions. Denn (1980), notes that pressure losses caused by these devices are usually correlated by an equation of the form "p K 2 (8) Ev = = U , # 2 where Ev are the losses in terms of "velocity heads" and K is a coefficient depending upon the fitting. For a piping system, the total losses, in terms of, say, a Bernoulli equation are obtained by summing contributions from eq(7) for all of the fittings and adding to all of the straight ! sections of pipe. Flow measuring devices such as orifice plates or Venturi meters can also be present.

& 2"p ) + (9) Q = Cd Ao( 2 # 1 $ % ( )+ ' ( * where Cd is the discharge coefficient for these devices, and " is the ratio between the throat and the pipe cross sectional areas (Ao/Ap) (see Middleman page 476).
c. Flow in a packed region ! Chemical Engineers often deal with packed-bed reactors where the "pipe" is filled with solid catalyst particles. The equations for empty pipe flow do not work without considerable modification. What does work is an empirical relation called the Ergun equation (Middleman, 1998),

1/ 2

150(1 % # ) + 1.75 . Uo Dp $ It is common to define the Reynolds number for a packed bed flow as f ! = 2 L$Uo (1 % # )

"PDp# 3

(10)

Re =

Uo Dp ! (1 " # )

(11)

In these equations, Dp is the particle diameter, # is the volume fraction that is not occupied by particles (void fraction and Uo is the superficial velocity. Equation (10) contains the interesting behavior that the pressure drop varies as the first power of Uo for small Re and as Uo2 for higher Re (see Bird, Stewart, Lightfoot, 2002 page 189-192).

13

Experimental Procedure Make sure the back valve is open before turning on the pump. This will prevent pressure buildup in the somewhat fragile acrylic pipe. Open all valves leading to the flow loops in the pipe network. Turn on the valves used to adjust the rotameters flow rate to get a maximum flow and allow the water to flow for 10 minutes to deaerate the pipe network form the air. Close all valves except one of the valves leading to the chosen flow loop, and then adjust the rotameters flow rate to get a desired flow rate then. Calibrate the rotameters using the graduated cylinder and a stopwatch. To prevent damage to the manometers, connect the pressure taps of the high scale manometer (channel 4) at the desired positions. If the differential pressure is too low, use the low scale manometer. Measure the pressure drop with the liquid manometers and the electronic manometers for the first four pipes over various lengths of pipe and at various flow rates. Make sure that the gate valve and the globe valve are fully open before you open the valve leading to the flow loop containing them. Measure the pressure drop across the gate valve, globe valve at fully, half and quarter opining and the strainer at various flow rates. Measure the pressure drop across the packed pipe, the orifice plate, and the venturi meter at various flow rates as well. Select conditions enabling you to cover laminar and turbulent flow regimes. Then, establish the reproducibility of your measurements.

14

Pipe Network
1. a smooth pipe of ID = 0.269" 2. a smooth pipe of ID = 622"

3. a smooth pipe of ID = 0.824"

4. a rough pipe of ID = 0.824"

5. a packed pipe of ID = 0.8"

6. a smooth pipe of ID = 0.824" has a gate valve, a globe valve and a strainer

7. a smooth pipe of ID = 0.824" has an orifice plate and a venturi

15 Pipe Flow program I. II. III. IV. V. VI. Create a new folder on the computer desktop for your work group. Launch LabView which located on the disktop. Open the file called Flow Through. Select a channel to be monitored before starting the run. Start the program by clicking on the arrow located in upper left-hand corner. Choose a name for the file that will be created by the run and save it to the new folder you created in the first step. Note: Changes during the run 1. The graph view can be altered. Holding down the mouse button with the hand cursor positioned over the graphing area will allow the graphing view to be moved to a desired position. 2. The X, Y axis can also be changed while the program is running. Highlight the value that is to be changed then type in the new value and hit-enter. 3. Changing the channel during a run will result in faulty data. The correct channel must be selected before the run begins. VII. VIII. IX. When the run is complete hit the Stop button. Launch Microsoft Excel from the Apple menu and open the new file created from the run. The first column of data is the time in seconds. The second column contains pressure differences data. May be you need to calibrate the data with the liquid manometers data for various flow rates.

Suggested Report Items 1. 2. 3. 4. Verify the universality of your friction factor- Reynolds number relations for a smooth pipes. Verify the form of the equations for a packed bed and rough pipes. Determine the head losses due to fittings and K coefficient values. Provide an assessment of the quantitative agreement between your data and established correlations. Use a systematic error analysis technique to justify the extent of agreement.

More details can be found on the Saddawi webpage DYE DEMONSTRATION In 1883 Osborne Reynolds made an important contribution to the field of fluid mechanics when he demonstrated with his famous dye stream experiment that two flow regimes are expected to exist. For sufficiently small flow velocities the dye stream did not disperse radially. This type of flow came to be known as "laminar" flow. For faster flow the dye stream mixed very rapidly; this is the so-called "turbulent" flow regime. It is expected, therefore, that some sort of transition region exists between laminar and turbulent flows. (See the table below.)

16

Similar transitions can frequently be observed in a rising smoke stream as illustrated below. Or from the Notre Dame power-plant gas turbine stack on a very cold, still day.

Your report should include a description of your observations.

17 REFERENCES 1. Bird, R. B., Stewart, W. E., Lightfoot, E. N., Transport Phenomena, Wiley, 2002. 2. Bennett, C. O., J. E. Myers, Momentum Heat and Mass Transfer, McGraw - Hill, 1982. 3. Denn, M. M., Process Fluid Mechanics, Prentice-Hall, New Jersey, 1980 4. Geankoplis, C., Unit Operations and Transport Processes, 3rd Ed., Prentice Hall, 1993. 5. Middleman, S., An Introduction to Fluid Dynamics, Wiley, 1998 5. Nevers, N, Fluid Mechanics For Chemical Engineers 2nd Ed, McGraw-Hill, Inc., 1976 6. Welty, J. R., Wicks, C. E., Wilson, R. E., Fundamentals of Momentum, Heat and Mass Transfer, 3rd Ed., John Wiley & Sons, 1984.

S. Saddawi, January, 2014

18 EXPERIMENT HT3 HEAT EXCHANGER Heat Transfer and Heat Exchangers Heat transfer is one of the most studied subjects of engineering because virtually every industrial process involves at least some heat exchange. A device whose primary purpose is the transfer of energy between two fluids is called a heat exchanger. Traditionally, equipment was designed with the aid of correlations that were expected to be accurate for the specific flow geometry and conditions. Recent advances in numerical techniques combined with greater speed and memory in computers has lead to an increasing reliance on calculated results that may be "exact" if the flow is laminar or obtained from models for turbulent flow.
T10 T1 T4 T5 T2

T3

T6

T8 T7 Parallel Flow

T9

T10 T1 T3 T4 T5 T2

T6

T7

T8 Counter - Current Flow

T9

Figure 1. A double-pipe heat exchanger The closed-type exchanger is the most popular one. (One example of this type, the double-pipe exchanger, is shown above in Figure 1.) In this type the hot and cold fluid streams do not come into direct contact with each other but are separated a tube wall or a surface which may be a flat or curved in some manner. Energy exchange is thus accomplished from one fluid to a surface by convection, through the wall or plate by conduction, and then by convection from the surface to the second fluid.

19

When a fluid flowing through a conduit is at a temperature different from that of the walls of the conduit, heat is transferred between the surrounding walls and the fluid. This flow of heat can be expressed as being proportional to the product of a characteristic area for the system and a characteristic temperature-difference for the system . The proportionality factor is known as the heat-transfer coefficient . It is evident that the magnitude and nature of the heat-transfer coefficients are directly related to the definitions of the characteristic area and characteristic temperature-difference. For the fully developed and steady flow of fluids through circular tubes of uniform cross-section, the heat-transfer coefficient is a function of the diameter and length of the tube and the density, viscosity, heat capacity, thermal conductivity and average velocity of the fluid. For circular tubes of uniform cross-section, which are completely filled with a flowing fluid, the characteristic area is defined as the wetted surface through which the heat is transferred A = !DL where: D = tube diameter L = tube length The characteristic temperature-difference can be determined in a number of ways, depending on the application. In heat exchangers of the type described here, it is appropriately defined as the logarithmic-mean temperature difference: ("T)lm = ("T1 - "T2) / ln("T1 /"T2) where "T1= (T3-T7) "T2 = (T6-T10) (See figure 2.)
1 T3 T4 !"1 T1 Metal T7 T8 !# A T9 T5 T6 T2 T10 !"2 2

(1)

(2)

Figure 2. Variation of !T in a counter-current heat exchanger A = area= !DdL (where dL distance from point 1) For steady flow the rate of heat transfer (gained by the cold fluid or lost by hot fluid) "Q" is QH = #<v>! Ri2 CP (T3-T6) (3)

20

Where # fluid density, Cp heat capacity (for hot fluid determined at its mean temp (T3+T4+T5+T6)/4), Ri internal hot tube radius, <v> time-average velocity (for hot fluid) and

m = " < v > #Ri2

QH = m C p (T3 " T6 )
Another common way of expressing the rate of heat transfer for a situation involving a ! composite material or combination of mechanisms is with the overall heat-transfer coefficient : ! Q = U A !Tlm (4) The value of the overall heat-transfer coefficient, U, can be calculated from Equations (3) and (4). U = [(T3-T6)/ !Tlm] [#<v>Cp/4] [Di/L] Or U = [(T3-T6)/ !Tlm] [m0 Cp/Alm] (5)

Where Alm mean surface area (logarithmic mean for cylindrical conduits). Alm = (Ao-Ai)/ln(Ao/Ai) The rate of heat transfer can also be evaluated from the heat flux at the wetted surface of the tube if the temperature gradient in the flowing fluid can be determined at the solid-fluid interface. For steady and fully-developed flow of fluids where forced convection exists: Nu = ( Re, Pr, L/D) where Nu = (hi Di)/ k; Nusselt number Pr = (Cp )/k; Prandtl number Re = (# D < v >) /; Reynolds number h i= individual heat-transfer coefficient Di = internal tube diameter k = thermal conductivity of fluid = fluid viscosity Cp = heat capacity All above physical properties should be determined at hot fluid mean temperature Tm Tm = (T3+T4+T5+T6)/4) (6)

21

When the temperature gradient in the fluid is known, the exact functional form of Equation (6) can be determined. For laminar flow in a tube of circular cross section, where the velocity profile is parabolic, the dimensionless, independent variables given in Equation (6) appear in combined form as the single, dimensionless, independent variable, [ ( Re Pr D)/L]. In particular, a limiting exact form of Equation (6) for conditions of constant temperature at the interface and short tube-lengths or high flow-rates is Nu = 1.62 [ (Re Pr D) /L]1/3 (7)

Owing to the complicated nature of the temperature distribution in turbulent flow, which in turn requires a knowledge of the rate of momentum transfer at the solid-fluid interface, a precise, functional form of Equation (7) cannot be determined in the same exact manner as for laminar flow. It has been possible, however, to predict the local Nusselt number as a function of the independent variables given in Equation (7) in a semi-theoretical manner. For highly turbulent flow (Re > 10, 000) the function is of the form Nu = a Reb Prc where a, b, and c are constants. The overall coefficient, U, can be expressed in terms of the individual fluid coefficients, hi and ho, and the thermal resistance of the separating wall as follows: 1/Uo = Ao/hiAi + $Ao/kwAm + 1/ho (9) (8)

where Uo = overall coefficient based on area Ao. Ao and Ai = surface areas on the respective sides of the separating wall. Am = mean surface area (logarithmic mean for cylindrical conduits). $ = wall thickness. kw = thermal conductivity of the wall material.

22

The Experiment
This experiment is intended to demonstrate some of the basic principles of single phase force convective heat transfer in closed conduits. Students will use a double-pipe heat exchanger to determine individual and overall coefficients for heat transfer from hot to cold water over a range of flow rates. The experimental plan should include an observation of the effects of cocurrent versus countercurrent operation.

Equipment
Copper-Constant thermocouples and an electronic meter are used to measure the temperature. The water flow rates are measured by rotameters. An electric heater is used to adjust the temperature of the hot water. Students should understand the principles of operation of all of these devices before beginning the experiment. See figure 3.

General Objectives
Develop one or more correlations for the experimental heat-transfer coefficients and compare them to appropriate literature correlations. For this purpose, give careful thought to the type of experimental data required and to the methods of analyzing and plotting the data. Be sure to explain the theoretical basis of the literature correlation and why it is expected to describe the present experiment. Give reasons for any disagreement between theory and experiment. Use the experimental results in the simulation program provided in the computer with Microsoft Excel, Macro sheet. Observe the influence of each parameter on the temperature profile and compare it with the experimental data.

Some useful information Heat Exchanger: Core tube

Material - Copper External diameter Internal diameter Length Material - Copper External diameter Internal diameter

(do) = 9.5mm (di) = 7.9mm = 3*350mm =12.7mm =11.1mm

Outer tube

The physical properties of the water - see appendix

23

Observation Sheet Co - Current Flow* Counter - Current Flow** Test 1 Metal Wall at inlet T1 / oC Metal Wall at inlet T2 / oC Hot stream at inlet T3 / oC Hot stream 1st intermediate T4 / oC Hot stream 2nd intermediate T5 / oC Hot stream at outlet T6/ oC Cold stream at entry/exit T7 / oC Cold stream intermediate T8 / oC Cold stream intermediate T9/ oC Cold stream at entry/exit T10 / oC Hot water flowrate l min-1 Hot water actual flow kg s-1 Cooling water flowrate kgs-1 oC Mean hot water temperature Specific heat at mean temp. Cp /kJkg-1K-1 Density at mean temp. (hot) # /kgm-3 Thermal Conductivity at mean temp. k /Wm-1K-1 Viscosity at mean temp. /106Nsm-2 Heat transfer from hot water Qh /Watts Heat transfer to cold water Qc /Watts Local heat transfer coefficient at inner surface of core tube hi/Wm-2K-1 Local heat transfer coefficient at outer surface of core tube ho /Wm-2K-1 Over all heat transfer coefficient U / Wm-2K-1 Over all heat transfer coefficient Uo / Wm-2K-1 Prandtl No. at mean temp. Pr Reynolds No. at mean temp. Re Nusselt No Nu

24

Figure 3.

25

References
1. Geankoplis, C., Unit Operations and Transport Processes, 3rd Ed., Prentice Hall, 1993. 2. Welty, J. R., Wicks, C. E., Wilson, R. E., Fundamentals of Momentum. Heat and Mass Transfer, 3rd Ed., John Wiley & Sons, 1984. 3. Bird, R. B., Stewart, W. E., Lightfoot, E. N., Transport Phenomena, Wiley, 1960. 4. Bennett, C. O., J. E. Myers, Momentum. Heat and Mass Transfer, McGraw - Hill, 1982. 5. Foust, A. S. et al., Principles of Unit Operations, Wiley, 1960. 6. McCabe, W. L., Smith, J. C.~ Unit operations in chemical engineering, McGraw-Hill, 1985.

Saddawi January, 2014

26 Appendix

Figure 4-1.

27

Figure 4-2.

28

Figure 4-3.

29

Figure 4-4.

30

Figure 4-5

31

Figure 4-6

32

Figure 5.

33

Figure 6.

34

Figure 7.

35 EXPERIMENT HT2 NATURAL CONVECTION AND RADIATION HEAT TRANSFER FROM AN ELECTRIC LIGHT BULB Objective The aim of this experiment is to measure local natural convection heat transfer coefficients from spheres. The Nusselt number is obtained and compared with an existing literature correlation. Apparatus . Infra-red Electric Bulb and Protractor Support Stand . Portable Potentiometer . A. C. Auto. Transformer . Lab View computer program

Introduction Although mean natural convection heat transfer coefficients from spheres have been reported in the literature, we know of no studies which have reported local natural convection heat transfer coefficients from spheres. In this experiment, local natural convection heat transfer coefficients from a horizontally mounted electric light bulb are measured. A 500 watt infra-red "spherical" bulb is employed. The bulb is obviously not a true sphere since it has a stem; however, by mounting a fine thermocouple as a temperature sensor in a certain location on the bulb, it is hoped that the effect of the stem of the bulb on the thermal measurements will be largely eliminated. The essential apparatus making up the experiment is shown below. The electric light bulb is mounted on a stand and the power to the lamp is supplied by a variable autotransformer which has a voltmeter, ammeter, and watt meter built into it. Eight thermocouples attached to the electric light bulb, one located in the front of the light bulb and the others will distributed in an angle of 30o starting from the top of the light bulb, the lead wires are attached to a thermocouple interface. The thermocouple interface is hooked to the computer which contain a Lab View program. The data will be saved in a spreadsheet format for farther analysis. The electric light bulb is regarded to be a true sphere of diameter 3.783 inches. The voltage delivered to the tungsten filament of the sphere is assumed to be radiated in an uniform intensity in all directions. However, of the wattage which is delivered to the filament, a small amount is radiated as light which is transmitted through the glass; a larger amount is radiated as short-wave infra-red radiation which is also transmitted through the glass - the rest is absorbed by the glass and then lost by direct radiation and natural convection. The fraction of the wattage, which is absorbed by the glass, is converted to Joules, and this is divided by the area of the idealized sphere to give a heat flux. This heat flux is assumed to be of uniform magnitude in all directions. However, it is likely that convection heat transfer from the filament by argon gas within the bulb causes local overheating near the top of the bulb.

36

37 Procedure Start the computer program which is described below. Turn on the auto transformer by turning the lower toggle switch on the right-hand side of the transformer to 120 or 140 V. Adjust the wattage to the desired value (0-300) watts. The pens on the recorder should move accordingly. Check all thermocouple graphs on the computer Lab-View program. When no further temperature variation is observed, (generally at least 5 minutes is necessary) the following readings may be taken: Voltage, amperage, and wattage and thermocouple readings (i.e. at 0, 30, 60, 90, 120, 150, 180, and 270 degrees). Note: Please do not exceed 300 watts Thermocouple Tracking Program 1. Create a new folder on the computer desktop for your work group. 2. Launch LabView from the Apple pull down menu. 3. Open the file called Natural Convection. 4. Start the program by clicking on the arrow located in upper left-hand corner. 5. A configuration subroutine will launch automatically called Instrunet. 6. Click the Restore button. This will configure all analog inputs to monitor 'K' type thermocouples. 7. Once the inputs are monitoring readings in degrees C, close the Instrunet window by clicking in the upper left-hand corner of that window. 8. Choose a name for the file that will be created by the run and save it to the new folder you created in the first step. 9. End the run after equilibrium has been reached by pushing the Stop button. 10. Launch Microsoft Excel from the Apple menu and open the new file created from the run. The first column of data is the time in seconds. The other eight columns contain temperature data. 11. The file can be copied to your AFS space or placed on floppy. 12. To mount your AFS space go to the Apple menu and select AFS Logon and follow the prompts. Just remember to drag your AFS space icon to the Trash once you are done. Note: Changes during the run 1. The graph view can be altered. Holding down the mouse button with the hand cursor positioned over the graphing area will allow the graphing view to be moved to a desired position. 2. The X, Y axis can also be changed while the program is running. Highlight the value that is to be changed then type in the new value and hit-enter. Suggestion: Keep in mind that controlling the power to the bulb is done manually. This program is just a data gathering program. So feel free to start and stop the program at any time. To keep down the size of the data file, the last 3 to 5 minutes of the experiment may be all you need to save.

38

Analysis Calculate the local natural convection heat transfer coefficients ( hc ) for both wattages at all angular positions, the plot hc versus angular position. Determine the mean ( h c ) by means of Simpsons rule. Compute the mean Nusselt numbers Nu and compare with the equation suggested by McAdams.

h r Nu = c k where r is the radius of the bulb.

1

Nu = 0.53 (GrPr)4 Gr = Grashof Number =

g"# 2$TR 3

C Pr = Prandtl Number = p k !
The thermal properties of air for use in the above dimensionless numbers are taken at the arithmetic mean temperature of the air film, and may be interpolated from Table 3. ! Determine the light efficiency of the electric light bulb (% of energy delivered as light) at both wattages used. It has been determined with an optical pyrometer that when 256 watts is delivered to the tungsten filament that its black body temperature TBB is 3400R. Since the hemispherical emissivity of tungsten is low (at % = 0.230 Table 2) the true temperature of the tungsten coil is
4 TBB = !Ts4

or

Ts =

TBB
1 4

! Ts is the temperate of radiation source filament (tungsten coil)

The glass bulb transmits all of the radiation between 0.35 micron and 2.70 micron. &1TBB = 0.35 x 3400= 1190 &2TBB = 2.7 x 3400 = 9200

3400 0.693

39

From the Table of Planck radiation functions (Table 1) determine f values (the fraction of watts delivered to the filament) f1 = 0.0000 f2 = 0.6477 at &1TBB = 1190 at &2TBB = 9200

Thus 0.6477 is the fraction of the watts delivered to the filament, which is transmitted by the glass. The remainder watts which is absorbed on the glass and is lost by convection and radiation from the glass = (1-0.648) * 256 = 90.1 The area of the ideal glass bulb is 0.315 ft2 so that the thermal flux leaving the bulb is

q 90.1 # 3.412 = 975.94 Btu/hr x ft -2 ="= A 0.315

The flux is transported out both by radiation and natural convection. Thus

!
where

' = hT (T - T#) hT = total heat transfer coefficient hT = hc + hR

where hc = the convection transfer coefficient and hR = the radiation transfer coefficient T = the localized bulb temperature which is a function of position (in Rankine ) and T# is a the ambient temperature in the laboratory

)! T $4 ! T $4 , +# & '# ( & . " + * 100 % " 100 % . hR = 0.173* ! glas * T ' T(

%glass = 0.876 (independent of temperature for all practical purposes)

40

For different power settings: less than or more than 256 watts is delivered to the bulb, the black body temperature of the filament may be determined by a simple computation. Let TBB2 be the black body temperature of the tungsten coil when the power delivered to it is P2 watts. Then
1

" P %4 TBB 2 = 3400 * $ 2 ' # 256 & In this experiment, Prandtal Number, Nusselt Number, and Grashofe Number may be calculated. However, to calculate these numbers, it is necessary to determine the mean heat transfer coefficient. This can be done by applying the Simpsons Rule using the heat transfer coefficient !positions on the bulb. (hc) at several
Simpsons Rule is

I=

B A

f(x) dx =

h 3 [f(A) + 4f(A+h) + 2f(A+2h) + 4f(A+3h +...2f(B-2h) + 4f(B-h) + f(B)]

#

hc =

"

1 2

$h
0

&# ) cos( " % +d% '2 *

(h =
Thus !

" ) 6
# *[0 + 4hc(30) x cos 60 + 2hc(60) x cos 30 2 * 3* 6

hc =

"

! + 4hc (90) cos (0 ) + 2hc(120) x cos 30 + 4hc (150) cos (60) +0]

! The mean temperature difference (glass surface temperature-ambient temperature) may be computed by dividing the total heat flux by hT obtained from Simpsons rule as above.

"T = # hT = # (hc + hR )
The mean air film temperature to use in the Grashof, Prandtl and Nusselt numbers is thus

Tf = T! +

"T 2 where T# is the ambient temperature.

41 Although natural convection is mostly turbulent flow of fluid, there is always a layer near the surface that is laminar (Welty 209). The approximate thickness of the boundary layer may be found by the following equation k != hc where k is the thermal conductivity and $ is the boundary layer thickness. Table I Planck Radiation Functions f 0 0.0003 0.001 0.0025 0.0053 0.0098 0.0164 0.0254 0.0368 0.0507 0.0668 0.0851 0.1052 0.1269 0.1498 0.1736 0.1982 0.2232 0.2483 0.2735 0.2986 0.3234 0.3477 0.3715 0.3948 0.4174 f 0.4607 0.4812 0.501 0.5201 0.5384 0.5561 0.573 0.5892 0.6048 0.6197 0.634 0.6477 0.6608 0.6733 0.6853 0.6968 0.7078 0.7183 0.7284 0.738 0.7472 0.7561 0.7645 0.7726 0.7803 0.7878 f 0.8017 0.8082 0.8145 0.8205 0.8263 0.8318 0.8371 0.8422 0.8471 0.8689 0.8869 0.9018 0.9142 0.9247 0.9335 0.9411 0.9475 0.9579 0.9657 0.9717 0.9764 0.9891 0.9941 0.9965 0.9977 0.9984

&TBB(R) 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 4200 4400 4600 4800 5000 5200 5400 5600 5800 6000 6200 6400 6600

&TBB(R) 7000 7200 7400 7600 7800 8000 8200 8400 8600 8800 9000 9200 9400 9600 9800 10000 10200 10400 10600 10800 11000 11200 11400 11600 11800 12000

&TBB(R) 12400 12600 12800 13000 13200 13400 13600 13800 14000 15000 16000 17000 18000 19000 20000 21000 22000 24000 26000 28000 30000 40000 50000 60000 70000 80000

42 Table II Emissivity of Tungsten Temp. F0 100 1000 2000 3000 4000 5000 6000 6190 % 0.03 0.092 0.162 0.23 0.281 0.32 0.35 0.355

(melting point of tungsten)

Silica glass transmits all of the incident radiation in the wave-length range between 0.35 and 2.7 microns and is opaque at longer and shorter wave-lengths. The emissivity of glass for long wave-length radiation is 0.876. The optical spectrum is from 0.4 to 0.7 microns The average thickness of the glass bulb is 0.027".

Table III Properties of Air ____________________________________________________________ T(F) k(Btu/hrxftxF) Pr g

"# 2 2

1 F " ft 3

_____________________________________________________________ 100 200 300 0.0154 0.0174 0.0193 0.72 0.72 0.71

! 1.76 !x 106
0.850 x 106 0.444 x 106

43 References 1. 2. Incropera, F. P. and D. P. Dewitt, Fundamentals of Heat and Mass Transfer, John Wiley and Sons, NY (1985). Rohsenow, W. M. and H. Y. Choi, Heat Mass and Momentum Transfer, Prentice Hall, NY (1961). Revised by Saddawi January 2014

44 EXPERIMENT MT1 MEASUREMENT OF THE BINARY DIFFUSIVITY OF DIETHYL ETHER IN AIR

Objective Use a Stefan diffusion cell to obtain values of the diffusivity of diethyl ether in air at temperatures in the range 25-32C. Report on the sensitivity of the measurements as a function of temperature. Apparatus: . Capillary Diffusion Cell . Constant Temperature Bath . Cathetometer

Introduction Whenever a concentration gradient exists, species are transported by diffusion from a region of high concentration to a region of low concentration, in a similar way heat is conducted from regions at high temperature to regions of low temperature. Because of the similarities between heat and diffusion mass transport, Adolph Fick proposed that the diffusional mass flux is proportional to the concentration gradient, i.e., JA ( - ) cA [1]

where the proportionality factor in Eq. [1] is prefixed with a minus sign to indicate that the net flux of material is in the direction of decreasing concentration. When the fluid is a binary mixture of ideal gases, it can be shown using kinetic theory that this proportionality factor is equal to the binary diffusivity, DAB of the gas mixture. A consideration of Eq. [1] indicates that DAB may be determined experimentally if the molar flux and the concentration gradient are known. In the latter part of the nineteenth century, Stefan devised a convenient technique for determining the diffusivities of vapors of volatile liquids by placing the liquid in the lower part of a vertical capillary. As the fluid A evaporates, its vapor diffuses through a nonvolatile vapor, B, to the mouth of the capillary where it is swept away by a stream of vapor, B. Under these conditions the flux is determined by following the descent of the liquid-vapor meniscus due to evaporation from the liquid phase. The boundary conditions on the vapor phase in the capillary are then set by assuming that the vapor is in equilibrium with the fluid at the fluid-vapor interface and that the concentration of the diffusing vapor at the mouth of the capillary is identical to its concentration the purge vapor.

45 Theory Since the change in height in the liquid-vapor meniscus is slow, it can be assumed that the gas column is at steady state. For a unidirectional molar flux steady state mass balances for A and B are

dN A =0 dz

dN B =0 dz

or upon integration

NA = C1

NB = C2.

[2]

where C1 and C2 are constants According to Fick's law JA = - cDAB and by definition JA = NA - yA(NA + NB) then NA = -c DAB [4]

dy A dz

[3]

dyA + yA (NA + NB) dz

[5]

But since B (air) is insoluble in the liquid, at the gas-liquid interface, NB = 0, since NB = constant, then NB is zero through the gas column and the NA flux is therefore given by

NA =

! cDAB dy A . (1 ! yA ) dz

[6]

The boundary conditions on the capillary are: at z = 0

o yA = yA

[7]

at z = z*

yA = y* A

[8]

where y is the mole fraction of A in equilibrium with the liquid phase at the meniscus. Assuming the mixture is ideal so that c = P/RT, Eq. [6] is integrated to yield

46

PD AB NA= ln RTz*
or in terms of partial pressure

1 - yA 1 yA

. [9]

P DAB PA1 - P A2 N A = - RT z* PB LM

[10]

where PA1 and PA2 are the partial pressures of ethyl ether at the vapor-liquid interface and at the mouth of the capillary respectively. (PB)LM is the log mean partial pressure of air between the mouth of the capillary and the meniscus. A mass balance for A in the liquid gives:

d !A "r 2 z* )= # N A"r 2 ( dt MwA

or

dz* Mw A N A =! dt "A

[11]

Combining Eqs. [10 and 11] and integrating over time yields
*2 Z*2 ! Zo =

2tPMw A DAB {PA1 ! P A2 } ( PB ) LM RT" A

[12]

which gives the height of the vapor column in the capillary as a function of time. The slow flow of air at the top of the capillary flushes the vapor from the top, hence one can assume that PA2 = 0. The partial pressure of A at the interface is calculated from the vapor pressure of A.

Procedure Anhydrous Diethyl Ether (Mallinckrodt, Mw = 74.12) is placed in the lower two-thirds of a glass capillary which has an inner diameter of 0.1 cm. The capillary is then placed in a cell which has been inserted in the thermostatted bath. Dry air (Mw ~ 29.) is provided at bath temperatures by passing air through an in-line gas filter (Gelman), and a silica gel moisture trap (American Scientific). The air should be admitted at a rate which is sufficient to flush the vapor from the tip of the capillary without inducing mixing inside the mouth of the capillary. The optimum rate may be determined by trial and error. The capillary is readily filled with the syringe inserting the needle into the capillary until it just touches the closed end. When the liquid level is near the mouth of the capillary, withdraw the

47 syringe. The syringe is now used to withdraw fluid to a specified level. An initial vapor column height of about 1.1cm affords a good balance between accuracy and duration of the experiment. Insert the capillary into the glass holder and measure the height of the Top of the CAPILLARY FIRST. The value of Z* is relative to the top of the capillary, consequently without an accurate reading of this first point the experiment is useless. After allowing approximately ten minutes for the fluxes to come to a pseudo steady-state,2,3 the height of the vapor column from the meniscus to the capillary tip is recorded as a function of time using the cathetometer. Readings of the cathetometer should be taken at five minute intervals and the experiment should run for at least one hour. Repeat this experiment at other temperatures in the range of something around 1 - 33C in order to obtain the binary diffusivity* as a function of temperature. Using ice obtained from the TA, make at least one run below room temperature. Compare this functionality with that predicted by theory (e.g., the Chapman-Enskog theory). NOTE: PLEASE CLOSE THE FLOW VALVE (located at the top of the water bath feed line) WHEN YOU SHUT DOWN THE WATER BATH. Data Analysis The diffusivity can be determined by a linear least squares analysis of the experimental data. Plot the square of the column height, z*2 - zo , against time elapsed from the start of the experiment and determine the slope of this curve. Report your results as the value of the diffusivity at the conditions under which it was measured and then give your results corrected to a pressure of 760 mmHg.** Plot log DAB vs. log T(K) to obtain the temperature exponent with those predicted by theory. Typical experimental results at low temperatures give an exponent in the range, 1.5 - 2.0. Consider the sources of experimental error and deviations between the experimental results and the theory used in analysis of the experiment. _______________________________ * Strickly speaking, air is a two component gas with trace impurities of argon, carbon dioxide, etc., and therefore the proportionality constant is a multi-component diffusion coefficient. However, since oxygen and nitrogen do not separate to any appreciable extent in this experiment, the coefficient measured may be considered a pseudo-binary diffusivity. ** Do not forget to record the barometric pressure and room temperature during this experiment.
*2

48 Notation c cA DAB JA - Total molar concentration of gas. - Molar concentration of volatile vapor. - Binary diffusivity of gas A in gas B. - Vector form of the diffusive molar flux of gas A. - Molar flux of gas A or gas B. - Pressure. - Gas constant. - Absolute temperature. - Mole fraction of gas A. - Mole fraction of gas A at the vapor-liquid interface. - Distance measured along capillary starting at the tip. - Distance from the capillary tip to the vapor-liquid inter-face. - Height of the vapor column at the start of the experiment. - Vapor from the volatile liquid. - Nonvolatile gas. - Density of A in the liquid phase. - Grad operator.

N A,N B
P R T yA

y* A
z z*
* Zo

A B #A )

References 1. 2. Cussler, E. L., Multicomponent Diffusion, Elsevier Scientific Publishing Co., New York 1976. Bird, R. B., W. E. Stewart and E. N. Lightfoot, Transport Phenomena, John Wiley and Sons, Inc., New York 1960.

49 3. Welty, J. R., C. E. Wicks, and R. E. Wilson, Fundamentals of Momentum, Heat Mass Transfer, Third Edition, John Wiley and Sons, New York 1984. Revised By Saddawi January 2014 and

50

Leaching Rate Measurement Phosphate Rock in 2% Citric Acid OBJECTIVE To determine the leaching rates in a rock phosphate 2% citric acid system, the mass transfer coefficient from experimental data, and to compare the results obtained with theoretical values for laminar and turbulent agitation, and to study the factors that influence the leaching rate such as particle size and temperature. APPARATUS Constant temperature bath Mixer Mixer speed measuring device Jacketed batch reactor) Sampling equipment and glass ware Chemicals Lab-view computer program Leaching rate Theoretical discussion Phosphates are essential ingredients in the fertilizers. Rock phosphate is the only source as raw material in the production of phosphate fertilizer. The demand for phosphate is proportional to the world populations increasing rate. Therefore the leaching of rock phosphate using 2% citric acid solution is important for two reasons: 1) the leaching of phosphates contained in phosphate rock is the process that simulate the digestion of phosphate rock with soil by rain solution when rock phosphates are applied as a fertilizer, 2) it is often used for testing the reactivity of raw materials containing phosphorus compounds. The leaching process refers to the extraction of soluble constituents from a solid by means of a solvent. The process may be employed either for the production of a concentrated solution of a valuable solid material, or in order to free an insoluble solid from a soluble material with which it is contaminated. In the mass transfer controlled leaching process, the fluids are always in motion e.g. batch processes with continuous mixing thus means that the fluid flows in a turbulent state past a solid surface, however, because the fluid velocity is zero at the surface of the solid particles, there must be a film of fluid adjacent to the surface. Hence, from this point of view there are two forms of mass transfer from the particle surface to the fluid. 1) It is controlled by molecular diffusion in the laminar flow region, and 2) it is controlled by turbulent transport in the turbulent region core. Using the idea that a thin film is responsible for the resistance of transfer, one can write the equation for mass transfer as dM k !A( cs " c ) 1 = dt b where

51 A, is the area of solid -liquid interface, b, is the effective thickness of the liquid film surrounding the particles, c, is the concentration of the solute in the solution bulk at time t, cs, is the concentration of the saturated solution in contact with the particles. M , is the mass of solute transferred in time t, and k', is the diffusion coefficient. For a batch reactor in which the total volume (V) of a solution remains constant, dM = Vdc dc k !A(c s " c ) 2 = dt bV Assuming both b and A remain constant, the time taken to increase the concentration of the solution from its initial value co to c can be determined by integration of the above equation

" cs ! c
ln

dc

" Vb dt

k #A

cs ! co k "A = t cs ! c Vb

For pure solvent co =0, therefore

1"

c = e cs

" k #A t Vb

5
t

! k "A

c = cs (1 ! e Vb )

! In the turbulent region the moving eddies transport matter from one location to another, just as they transport momentum and energy. As mentioned above, when the flow past an interface is laminar, however, the mass-transfer rate is based on molecular diffusion. Thus, when the Reynolds number is above the critical value, the usual buffer zone and turbulent core appear and the mass transfer coefficient, KL will become a resultant of both turbulent and molecular diffusion.
KL = k! + ED b

where, ED , is the eddy diffusivity Hence the mass transfer equation for turbulent region is

C = Cs (1 ! e

KL A V

52

From the mechanism of the mass transfer, it can be expected that, the mass transfer coefficient will depend on the diffusivity DL on the fluid properties and mass velocity. In a batch leaching process using a fine particle size and a continuous mixing, the effect of mixing on mass transfer is expressed by the mixing speed, which is defined by the dimensionless Nd 2 ! ) group ( where N is the number of revolutions per unit time of the mixer d is the diameter of the vessel # is the density of the fluid, and is its viscosity Nd 2 ! ) is less than 67,000 the mass transfer When the value of the dimensionless group ( coefficient will be
2 KL d 0. 5 "5 Nd # 1. 4 = 2.7 ! 10 ( ) ( ) DL #DL

and for higher values of (

Nd 2 ! ),

2 KL d 0.5 " 5 Nd # 0.62 = 0.16 ! 10 ( ) ( ) DL #DL

10

where KL the mass transfer coefficient, equals

k! in equation 1. b

Analytical method of phosphate determination The Molybdenate - Vanadate Phosphate method This method is based on the reaction of phosphate with ammonium vanadate and molybdenate in an acidic solution to form a yellow complex, who's color intensity in the solution is proportional to the phosphoric acid content. The color solution does not obey Beers law, therefore a calibration curve is necessary. Applied solutions 1. Standard phosphate calibration solutions. To prepare a standard phosphate solution with a concentration equal 1 mg/ml, 4.3900 g of KH2PO4 was dried at 110o C and dissolved in 1 liter distilled water in a calibration flask. The working calibration samples are obtained by appropriate dilution of this standard solution with distilled water. Or you can use the ready phosphate standard calibration solution.

53

2. The reference solution To prepare a reference solution, 15 ml of ammonium vanadate solution, 15 ml of ammonium molybdenate solution, and 20 ml nitric acid solution pour into a calibrated bottle of 100 ml volume, where distilled water was added to the solution until it reached the mark (100 ml). In the case you have all the three solutions in one named as (AVAM), then you need to pour 10 ml of that solution and complete the volume to 100 ml. 1. The ammonium vanadate solution (0.25% solution of NH4VO3) 2.5 g of pure NH4VO3 was dissolved in 500 ml of boiling distilled water, with intensive mixing. After the solution was cooled ,20 ml of nitric acid of density 1.4 g/cm3 was added and the volume was completed to 1 liter. The solution should be kept in a dark glass bottle 2. The 5% ammonium molybdenate solution Dissolve 50 g of pure ammonium molybdenate in 600 ml distilled water. Mix and heat the solution continuously (without boiling) until the solid is completely dissolved. Filter the solution while it is still hot. After cooling complete the volume to 1 liter. Keep the solution in dark glass bottle. 3. The nitric acid solution (HNO3) Mix one volume of nitric acid of density equal to 1.4 g/cm3 with two volumes of distilled water. Note The above three solution, ammonium vanadate, ammonium molybdenate, and nitric acid are mixed together (AVAM Solution) in one container and ready to use. Calibration Curve To obtain a calibration curve, calibration solutions need to be prepared first. Into a calibrated bottle of 100 ml volume add 1 to 7 ml, of the standard calibration solution of phosphate for a desired phosphate concentration, and 10 ml of the AVAM Solution. Complete with distilled water to the mark and mix very well. Warm up the colorimeter for one hr, adjust the wave length to 450 nm. Fill one cuvette with the reference solution and transfer it to the colorimeter cell. Adjust the absorption reading to zero. Wait about 60 min. for the calibration solution to reacts and gives a constant yellow color, fill a cuvette with it and transfer it to the cell of the colorimeter, and read the intensity of the color (absorbance). Draw the concentration of phosphate as Phosphorous ions in phosphate (mg/ml) versus the intensity gotten from the colorimeter. PROCEDURE 1. Citric acid solution preparation. Insert 20.0 g of pure dry citric acid into the jacketed Batch Reactor. Pour 1 liter of distilled water into the reactor and connect it to the water bath. Adjust the bath water temperature, and

54 the mixing rate to the desired values. Wait until all citric acid crystals are completely dissolved and the reactor temperature reaches a steady state. 2. Add 10.0 g of rock phosphate of a known particle size to the reactor. Start the time at the moment you add the rock phosphate. Because the leaching process between the rock phosphate and the citric acid takes place very quickly during the first 10 min. Samples should be taken after 1, 2, 3, and 5 min and so on. The samples are taken should include solids with the withdrawn solution, so that the ratio of solid to solvent in the reactor will remain constant. 3. Use the sampling equipment to pull 7 to 8 ml from the reactor mixture at the desired time. This sample should filter directly by a filter found inside the sampling syringe. 4. Calculations of the P2O5 concentration and the conversion percent (x ) Pour the filtered solution into a small beaker, and take from it a small volume noted by (A) ml. Pour this (A) volume into a calibrated 50 ml bottle, then complete the volume to the mark with distilled water. Take from this solution a volume noted by (B) and pour it into a 50 ml calibrated bottle. Add to it 10 ml of the AVAM solution, and fill to the mark with distill water. Mix it well. This solution should give a yellow color. To obtain a stable yellow color, a time interval of more than 30 min is needed. Prepare the colorimeter at least one hr before reading. Put the wave length in the filter of the colorimeter at & = 410 nm. Fill a cuvette with the solution and transfer it to the colorimeter cell, read the color intensity. From the calibration curve, get the corresponding concentration noted by (C). The concentration of the phosphorus ion in the leaching liquid is noted by (Cp) C ! 50 CP = = mg / ml A! B

C P2 O5 = C P ! 2.29 ( mg / ml)

X =

The amount of P 2 O5 in the leaching liquid The amount of P 2 O5 in the raw phosphate rocks

X =
where

CP2 O5 ! V r m ! o.3

! 100

Vr, is the liquid volume in the reactor m, is the mass of the raw phosphate rock 0.3, is the weight fraction of P2O5 in the raw phosphate rock.

55

Sampling Device Suggested report items 1. Summarize the experimental results in a table. 2. Provide some of estimate of your experimental error 3. Plot the conversion versus time for different particle size and different temperatures. 4. Determine the mass transfer coefficient from the experimental data and the empirical equations 6. Plot on natural log-log scale the concentration of P2O5 versus time and determine the slope. Calculate the leaching rate. REFERENCES 1. Geankoplis, Christie J., Transport Process and Unit operations, 3rd Ed., Prentice Hall, 1993. 2. Coulson, J. M., Richrdson, J. F., Chemical Engineering, vol 2., 3rd ED 2., McGraw-hill, 1978 3. Perry, John H., Chemical Engineering Handbook, 4th ED, 1963 Saddawi Jan 2014

56 EXPERIMENT TD1 PHASE EQUILIBRIA AND LIQUID DIFFUSION Objective Determine the solubility, diffusion coefficient, and the enthalpy of solution of carbon dioxide in Stoddard solvent in the range of 18 - 35C and at 1.0 atmosphere pressure. Apparatus Integral Phase Equilibria Unit Vacuum Pump Digital Absolute Pressure Gage Precision Temperature Controller

Introduction When a gas and a liquid phase which are not thermodynamically in equilibrium are brought into close contact, transfer of one or more components may occur from the gas phase to the liquid or, vice versa, by the mechanism of molecular diffusion. Mass transfer by molecular diffusion is the basic physical mechanism underlying many important areas of soil science, petroleum engineering, chemical engineering, biotechnology and nuclear engineering. In this experiment, a method for determining diffusion coefficients of Carbon dioxide gas in Stoddard solvent at constant volume and temperature is developed using Integral Phase Equilibria Unit. The solubility of a gas in a liquid solvent may be represented to good accuracy at dilute concentrations of the dissolved gas by Henry's Law: f=HX (1)

where, f is the fugacity of the gas in the gas phase in equilibrium with the liquid phase of concentration X of dissolved gas. H is the Henrys law constant, which is a function of temperature. Thus, by measuring the solubility one can obtain an estimate of the Henry's law constant. By measuring the solubility of gas in a liquid at several different temperatures, one can evaluate the enthalpy of solution (heat of solution) of the gas in the specific liquid solvent. In this case the heat of solution/R is equal to the slope of the plot of ln (H) versus 1/T. A derivation of this relationship should appear in the final report. (See the Gibbs-Helmholtz equation, which can be found in many thermodynamics texts, including the Appendix of Smith & Van Ness, 3rd ed.) The diffusion coefficient of the dissolved gas in the liquid phase may be determined by measuring the rate at which the gas dissolves in the quiescent liquid phase at a constant temperature. Phase Equilibrium Computations

57 Compute the moles of carbon dioxide dissolved from stoichiometry using the pump temperature, cell pressure and corrected barometer reading and total volume of carbon dioxide delivered from the pump. The following equation gives the details
o PT ("VP # Vd ) Vcg (PT # P1 ) n= # Z pRTP Z cRTc

(2)

Where: n = = = = = = = = = = gram moles of carbon dioxide absorbed in the liquid phase corrected barometer reading vapor pressure of Stoddard Solvent at cell temperature temperature at the pump temperature of the cell (bath temperature) total gas volume delivered from the pump to the cell volume of the gas phase in the cell compressibility factor of CO2 at pump T and PT compressibility factor of CO2 at cell T and PT dead volume in the system (cc)

PT

P1
Tp Tc

! Vp
Vcg Zp Zc Vd

The fugacity, f, can be determined from the Lewis and Randall Rule, which gives f = fo y where f fo y = = = = Therefore f = fugacity of CO2 in the gas phase fugacity of pure gaseous CO2 at PT and cell T mole fraction of CO2 in gas phase (PT - Plo)/PT (fo/PT)(PT - Plo) (4) (5) (3)

and by definition: fo/PT = *, the fugacity coefficient for pure CO2 in the gas phase at cell T and PT.

58

One can compute the fugacity coefficient by using the Generalized Second virial Equation. of State. Virial coefficients for carbon dioxide can be found in the compilation by Dymond (1969).

Determination of diffusion coefficient from experimental data A number of mathematical models have been proposed to determine the diffusion coefficients from experimental volumetime profiles, however all these models are developed from the equation of continuity for the solute component:

"C
"t

+ !.Cu v = !.J v + r

(6)

where r = Rate of reaction (kg/m3s) J = Mass transfer by the mechanism of molecular diffusion (kg/m2s) v = Molar volume (m3)

Liquid phase

Gas phase

Interface

Z(t)

Z=0

Fig. 1. One-dimensional diffusion process in the diffusion cell.

Referring to Fig. 1, for a one-dimensional diffusion cell in the absence of chemical reaction, and including the movement of the interface in the boundary conditions of the system, and based upon a model proposed by Higbie (penetration theory) in which a component in the gas phase is

59 absorbed into a liquid phase starting at time zero and continuing at longer times. The liquid interface is thus always at saturation, since the molecules can diffuse in the liquid phase away from the interface only at rates which are extremely low with respect to the rate at which gaseous molecules can be added to the interface. And according to the film theory the gas and the liquid phases at the interface are thermodynamically in equilibrium, i.e. the interface concentration of the solute, Ci remains unchanged as long as temperature and pressure of the system are kept constant. However, since the liquid phase initially has no gaseous component, diffusion of the dissolved gaseous component from the interface readily occurs. For all practical purposes the distance that the dissolved gas molecules must travel before reaching the bottom of the cell is semi-infinite. Thus the unsteady-state differential equation representing concentration changes with time and position is:

D12
where C Z ! t D12 = = = =

" 2C "C = "Z 2 "t

(7)

Concentration of dissolved CO2 in the liquid phase at Z and t. Distance in cm traveled from the liquid interface. time Diffusion coefficient of species 1 in 2.

The boundary conditions are: Z = 0: C = Ci Z + #: C = 0 and the initial condition is C=0 at t = 0:

This problem can be readily solved by using a similarity transform, as you have seen for the analogous problem in heat transfer. The mathematical solution to this initial value problem is:

# Z & Ci " C = erf % %2 D t ( ( Ci $ 12 '

where Ci erf = =

(8)

Interfacial concentration of dissolved CO2 Error function, defined in standard mathematical literature.

Equation 8 correctly gives the time-dependent concentration of dissolved CO2 in the liquid phase, but the actual measured quantity is the total amount of gas fed to the interface at any time. Thus from further mathematical treatment of Equation 8, it follows that the total amount of gas fed to the interface up to time t is

60
1

NT

# D &2 = 2AC i % 12 ( t $ " '

(9)

If one plots NT versus t1/2, the slope of this line is equal to 2ACi (D12/,)1/2. Thus, one can easily determine the diffusion coefficient from the slope of the line. Of course, the volume of gas fed ! to the interface is actually being measured, so it is preferable to plot VT vs. t1/2. However, this is not a problem because one can use the simple relation that PTVT = ZpNTRTp. The slope of the line should be determined by least square methods and an estimate of the uncertainty in D12 given. A brief treatment of the diffusion problem is presented on pages 70-71 of Sherwood et al. Lab reports must include estimates of experimental errors and estimates of the uncertainty in the final values calculated. For instance, the mathematical uncertainty in the Henry's law constants, enthalpy of solution and diffusion coefficients must be included. Procedure Properties of Stoddard Solvent and Cell Dimensions Molecular Weight of Stoddard Solvent = 136 g/mole Density of Stoddard Solvent: # = 0.7726 g/cc at 20C (The isobaric expansion coefficient is 0.00104/oC.) Weight of Stoddard Solvent in the cell = 39.04 gm at room temp Volume of the Cell 150 ml Volume of the magnetic stirrer = 1.45 ml Length of the tube from the top of the cell to the bulkhead = 9.75 Length of the tube going through the panel to valve 4 = 7.5 Di of the tube 1/8 From these info please Calculate the Void volume (gas volume above the liquid phase) (A) Preparation. ----------------------------------------------------------------------------------------------------------Set the Water Bath for desired temperature. Out gas (removal of CO2 from the system). 2. Open valves V3 - Valve 3 leading to vacuum pump. V4 - Valve 4 - To remove CO2 from the cell. V5 - Valve 5 From the pressure gauge line. Valves 1A, 1B, 2A, and 2B closed, since no gas in pumps as plunger is in. 3. Connect vacuum pump to system

61 4. Start vacuum pump and start the magnetic stirrer. It helps quick degassing* Degassing period is around 20 min.* You can calculate the degassing time by taking the average convective diffusivity value and knowing the volume of the cell given to you. The Stoddard solvent is basically a mixture of solvents containing predominantly heptane. So, now degassing is done (you decide the time). 5. Close the valve 3. There wont be any harm to the vacuum pump since it is just working against valve 3 from the surface of the solvent. 6. Switch on pressure gauge. Perform this check. Look at the pressure gauge. If pressure gauge reading is, # Constant (variation in decimal is considered constant) = vacuum is alright # Change in the readings = a leak in the system. So this check is important. 7. Close valve 4 to isolate the cell. 7. Close vacuum line V3. Disconnect the pump hose*. Switch off the vacuum pump.

(C) Filling of pumps A&B with CO2. ----------------------------------------------------------------------------------------------------------There are three valves on the cylinder. I Top valve (A) (main vale that lets out CO2) To open rotate counter clockwise (direction given on knob) II Regulator valve (B) (with black cover & direction attached to it) This regulates the flow of CO2 from cylinder to pump. III Release valve (C) This is the third valve (facing downwards). This allows CO2 from entrance of cylinder to valve 1, which is closed now. RHS pressure gauge shows pressure in cylinder. LHS pressure gauge shows pressure in line. 13. Open top valve A 14. LHS gauge - Keep it at 20 psi (approx.) using regulator. (only recommended) 15. Open valve C on the cylinder slowly. One rotation of knob should be enough (clockwise). 16. Open valve 1A on the board slowly. Plunger of pump A slowly moves out indicating CO2 is going into pump. Once plunger comes out close valve 1A 17. Gently tap/twist plunger slightly so that it comes out completely if stuck. Caution Open valve 1A slowly, otherwise plunger will be thrown out of the pump due to high pressure. Note: the handle of the plunger is graduated. It is the scale you need to look at when you are doing your experiment. 18. Fill pump B in same way described for pump A, in this case the valve is 1B. (D) Setting up of pressure gauge. ----------------------------------------------------------------------------------------------------------19. Switch on the gauge and change to PSI mode. Open valves A2, and B2. The pressure in the line will be 20 PSI (greater than atm. pressure).

62 20. Open valve 3. Switch the pressure gage to mm Hg mode. 21. Allow stabilization. This helps to bring pressure inside the lines to atm. pressure. Note, since pressure inside is high, gas only goes out through line with valve 3. 22. Close V3. Do this once the gauge reading stabilizes. At this point there is CO2 in the line with V2A, V2B, V3 & V5. V1A, V1B & V4 are still closed. 23. Adjust the pressure reading to zero*. Adjust the knob on the right side to zero. and (E) Actual experiment. ----------------------------------------------------------------------------------------------------------Students: #1 should note pump readings. #2 should maintain time (using stop watch). #3 should enter readings. Because there are two pumps (A&B) in the apparatus, you should isolate one of them by closing valve 2B if you decide to use pump A. 24. Opens V4 and immediately starts stopwatch. Remember, there is vacuum after V4 So gauge reading goes down to negative value and also the piston attached to the pump goes in automatically. Note, the pressure gauge readings show negative*. Thats fine. Now, 25. Push plunger to a known quantity (any qty according to your choice). Smaller increments are recommended. Observe that pressure gauge goes to a positive value. 26. Allow the pressure to fall* to zero 27. Note the time for zero pressure. 28. Again push plunger, pressure becomes >0, 29. Allow it to fall to zero and 30. Note time for this fall. 31. Repeat this procedure till all CO2 from the pump is over. 32. Close valve 2A and open valve 2B and continue (F) Reloading gas in pump. ----------------------------------------------------------------------------------------------------------During this entire process stopwatch is on and time is measured*. 33. Close V4, V2B. V2A V1A, and V1B are closed . 34. Open V1A. Allow CO2 to fill the pump A. 35. Close V1A. Fill pump B in same way. 36. Time is measured. 37. Bring back to atmospheric pressure. (Follow steps 19 to 22). 37.5. Open V 2A and V4. 38. Go back to step 25. It takes roughly 120 200 ml of the pump volume of gas for the timed experiment. It may vary. It is just a rough estimate. You do this till pressure does not decrease. (G) Total solubility estimation. -----------------------------------------------------------------------------------------------------------

63 38. Now start magnetic stirrer. Pressure starts decreasing*. 39. Continue with addition of CO2. You may stop noting time now*. A point is reached when you see pressure does not decrease after addition*. At this time, the solvent in the cell has reached saturation point. Stop CO2 addition. This gives solubility of CO2 at infinite time*. 40. Repeat the above procedure for two more different temperatures* of your choice.

H) Winding up ----------------------------------------------------------------------------------------------------------Monitor of the group has to ensure the following. CO2 cylinder (all three valves) is closed except regulator. It is not necessary. Pump is evacuated of any CO2.i.e. plungers are inside the pumps completely. All valves (V1A, V1B, V2A, V2B, V3, V4, V5) are closed. Water Bath is switched off. Stirrer is switched off. Cooling water tap is closed. 7. Pressure gauge is switched off. Vacuum pump is off. Work place is restored to the same condition as when given. Stopwatch returned to concerned TA. For clarifications, you can contact your TA After the run is over the bath temperature may be lowered which causes additional gas to dissolve in the cell. Thus equilibrium solubilities may be determined at several lower temperatures without starting from the beginning. When all experimentation is done make sure to close the valve on the CO2 cylinder.

64

Properties of Stoddard Solvent and Cell Dimensions The viscosity of Stoddard solvent as a function of TK is (centipoises) = 0.0895 exp (631/T). The volume of Stoddard Solvent is 100 ml at room temp. Di and Hi of the cell are (49.73 mm) and (61.72 mm) respectively. Vapor Pressure of Stoddard Solvent T,F Plo (atm) 60 0.0014 70 0.0016 80 0.0024 90 0.0035 100 0.0050 References 1. Smith, J. M. and Van Ness, H. C., "Introduction to Chemical Engineering Thermodynamics," 4th Ed., McGraw-Hill, 1987. 2. Sherwood, T. K., Pigford, R. L. and Wilke, C. R. "Mass Transfer," McGraw-Hill, 1975. Revise by Saddawi January 2014

65

Figure 3.

Experimental setup

66 Note on: Phase equilibria and diffusion experiment. Davide A. Hill May 5, 2004

The measurement of the diffusion coefficient of CO2 in the Stoddard solvent using the laboratory apparatus is deceptively simple. The phenomenology of the experiment is, in fact, quite complex, involving several transport mechanisms that become active at different times during the run. In order to understand and sort out these mechanisms, one needs to examine the way the experiment is conducted as well as possible causes for discrepancies. First the cell is evacuated to desorbs the CO2 from the liquid. As it will become apparent later on, this should be done by putting the cell under vacuum while vigorously stirring the liquid. Stirring is necessary to enhance convective mass transfer from the bulk of the liquid towards the interface as much as possible. This, in turn, allows one to reduce the time needed to eliminate the CO2 from the liquid. Prior to opening the valve to let the CO2 in (at very beginning of the run), the cell is at a pressure much lower than atmospheric. Also the liquid is stagnant (not stirred), as it should be in order to allow only molecular diffusion to occur as specified by penetration theory. That the liquid must be stagnant (zero velocity) is an important point, since any convection will act in a way as to reduce mass transfer limitations in the liquid phase (there are none in the gas phase, which is pure CO2), thereby masking the pure diffusional response, which is much less effective. This point provides the key towards understanding the anomalous transport behavior observed in the experiment. As soon as the valve is opened to allow the CO2 into the evacuated cell, because of the way the cell is constructed, a jet of CO2 at a reasonably high speed will impinge onto the surface of the liquid. This results in a vigorous stirring action of the liquid. Owing to the large mass transfer coefficient during this initial time, CO2 will be absorbed at a very fast rate and the absorbed amount will exhibit a jump at small times (say, less than one minute). Thereafter, the CO2 velocity is greatly diminished, and is no longer sufficient to stir the interface. Because of viscous dissipation, eddies in the liquid (i.e., the velocity) will gradually dissipate, and the liquid

67 will again become stagnant. The time for this to occur, -, can be estimated from the liquid viscosity and density as well as the geometrical parameters of the cell. My estimate is: -~h /., where ., and h, are the fluid kinematic viscosity and liquid layer height, respectively. A conservative estimate is about 1 minute. After this time the liquid is practically stagnant, and diffusion will occur for some time as prescribed by the penetration theory. This penetration regime is short-lived, however, since a new convective mechanism ensues once the CO2 concentration in the top interfacial layer of the liquid exceed a certain level. To understand the cause for the onset of this new regime one needs to consider what happens to the average density of the top liquid layer upon absorption of the gas. The Stoddard solvent has a density of about 0.77 (??) g/ml. If the CO2 in the liquid phase is more dense than the solvent the average density of the top layer will increase as CO2 penetration occurrs. The density of CO2 in the liquid phase is quite insensitive to pressure, and as such it can be estimated from that of pure, liquid CO2 at room temperature but at a pressure higher than atmospheric. (In entering the liquid phase (Stoddard solvent), intermolecular forces stabilize the gas molecules, thereby reducing the pressure necessary to keep them in the condensed state.) Using the data in Perrys handbook (section 3) one can estimate a density between1.0 g/ml and 1.1 g/ml for liquid CO2 at 25C. This value is higher than that of the Stoddard solvent, indicating that a mixture with CO2 will be more dense than the pure solvent itself. As a consequence, once the density of the top liquid layer in the cell builds up to a sufficiently large value, the top layer will start to sink, thereby causing again a stirring action that will rejuvenate the interface, replenishing it with fresh fluid elements. This buoyancy-driven stirring enhances mass transfer, thereby invalidating the penetration-in-a-stagnant-liquid hypothesis. One should recognize, however, that this process occurs by natural convection, driven by a relatively small density difference. Hence, the associated average velocity and mass transfer coefficient will be smaller than those recorded at the beginning of the experiment under the violent CO2 jet.
2

The above phenomena can be modeled as follows.

68

I) In the first period, when stirring from the jet is important, an integral balance on the concentration of CO2 in the liquid can be written as:

dC = A K mI (C* " C) dt

where V, A, KmI, and C* are the liquid volume, interfacial area, convective mass transfer

coefficient, and equilibrium CO2 concentration, respectively. In this regime the average (mixing cup) concentration of carbon dioxide in the liquid (C) grows as (assuming zero concentration at t=0):

C = C* (1 " exp("t / #I )) ;
/

#I = V /( AK mI )

where & is characteristic time scale for mass transfer during stage I, which depends on the

fluid velocity. Following the above discussion, we expect & to be quite small initially (say, a
/

few seconds) and then become very large as the velocity drops to zero. Once the fluid is quiescent, then the penetration theory kicks in. Because the jet-dominated period is quite small (say 1 min or less) if measurements are taken every minute or so, one will miss the trends entirely. A plot of the absorbed amount against t
1/2

will then show an apparent jump at

t=0 (i.e., the curve will not extrapolate to a zero intercept). II) Penetration theory period The theory for this period is described in the manual, and shall not be repeated here. However, we should recall that during this period the amount of adsorbed gas in the liquid should increase linearly with the square root of time. Hence, plotting absorbed amount against III)

t should produce a straight line.

Buoyancy-driven convective mass transfer period.

!
During this period convective mass transfer again takes over, allowing us to use an integral balance as in period I:

69

V
and

dC = A K mIII (C* " C) dt

(1)

C = C* (1 " exp("t / #III )) ;

/// /

#III = V /( AK mIII )

(2)

where & (>>& ) is the relaxation time for buoyancy-driven mass transfer.

From the last equation, one can note that if C<<C* (as we expect from the experiments which are run only over 20 min or so, as compared to the1 hr required to reach equilibrium with the stirrer on) then the exponential in the equation can be linearized to give:

C " C* t / #III ;
will increase as ( figures below.)

#III = V /( AK mIII )
t ) 2: a plot of (CV) against

(3)

During this period, the absorbed amount (C V) will increase linearly with t, or, equivalently,

will be quadratic in this regime. (See

! ! In consideration of the above mechanisms, a plot of absorbed amount (say, Y=CV) against
X =

will exhibit a finite intercept at t=0, and combine two terms, one linear and the other

quadratic in X. Such a curve can then be fitted by a 2nd degree polynomial as

!
!

Y = A + B X + C X2

(4)

where A quantifies the initial jump during phase I, B is the true slope to be used to calculate the diffusivity during the penetration regime, and C quantifies (and thus can be used to calculate) the convective mass transfer coefficient due to the buoyancy-driven flow. Theory based on surface renewal models indicates that the latter should not be very sensitive to changes in the diffusion coefficient (and thus temperature). Usually the linear term of the polynomial (B X) is weak compared to the quadratic one. Once a polynomial fit of the original Y vs X data using the above equation is obtained, then it is advisable to replot the function [Y-(A+CX )] against X, to check that the linear relation is indeed respected and to gain some feeling for magnitude of the scatter.
2

70

Absorbed CO2 mass

Fast mass transfer from jet

Penetration regime. Stagnant liq.

Buoyancy-driven convective mass transfer

t1/2

71

CO2; 1 atm t=0

-

CO2; 1 atm valve closed t=0

+

valve open

Jet Fast mass transfer

CO2; 1 atm valve open

CO2; 1 atm valve open t>"t1 CO2: 1atm

-<t<!t1

CO2: 1atm

Penetration regime. Stagnant fluid

Buoyancy-driven convective mass

72

GAS FLUIDISED BEDS INTRODUCTION When a fluid flows through a bed of particles contained within a tube, it will exert a drag force upon them and develop a pressure drop across the bed, which increases with increase in its superficial velocity (the superficial velocity is that based on flow through the empty tube). If the bed is unrestrained and the fluid flows upwards throughout, a condition will be reached where with increasing fluid velocity, the drag force will cause the bed to expand and so offer less resistance to the fluid flow until that force is sufficient to support the weight of the particles in the bed. The bed then becomes Fluidized and the fluid/solid system shows fluid-like properties where the bed can be made to flow from one vessel to another. The pressure drop across the bed !p remains constant with further increase in the fluid velocity and equal to the effective weight of the bed per unit area.

!p =

M (" # " ) g " p Sb p

(1)

If the fluid of low density (e.g. air), then the pressure drop across the bed will approach the weight of the bed

!p "
Where:

M g Sb

! M is the mass of Fused Alumina particles of density #p have been charged into a bed of crosssectional area Sb fluidized by a fluid of density # and g is the gravitational acceleration. In this Fluidization and Fluid Bed Heat Transfer Unit, the fluid is air and the unit is operated at a slightly higher than atmospheric pressure. The porosity of the bed when true fluidization occurs is the minimum porosity for fluidization and is %mf. This minimum voidage or porosity can be found experimentally when the bed expands to this voidage before particulate motion appears.

73 L0 Sb (1 - %0) = Lmf Sb ( 1 - %mf) (2)

Where L0, Lm are the heights of the bed at zero value of superficial velocity and minimum
f

fluidization velocity respectively. %0, %mf are voidage at zero value of superficial velocity and minimum fluidization velocity

!=

volume of voids in bed total volume of bed (void + solid)

Pressure drop and minimum fluidizing velocity. As a first approximation, the pressure drop at the start of fluidization can be determined as follows. The force obtained from the pressure drop times the cross-sectional area must equal the gravitational force exerted by the mass of the particles minus the buoyant force of the displaced fluid. !p Sb = Sb Lmf (1 - %mf)(#p - #)g (3)

The pressure drop across the packed beds is a function of Reynolds number. For laminar and turbulent flow (low, intermediate and high Reynolds number), Ergun proposed the following general equation

150(1 # $) L%(v ' ) 2 (1 # $) "p = [ ' ' + 1.75] v Dp % D'p$ 3

Rearrange Equation 4

(4)

"pD'# 3 150(1 % #) = + 1.75 2 L(v ') $(1 % #) v ' D' p $

(5)

Where the left side of the equation is call the Friction Factor (f) ! Where .' is the superficial velocity based on the empty cross section. L the height of the bed at any .

74 Shape Factors: Many particles in packed beds are often irregular in shape. The equivalent diameter of particles is defined as the diameter of sphere having the same volume as this particle.

Dp = !D'p
Where Dp is the equivalent diameter D'p is the average diameter of the particle and * is the shape factor

Then equations 4&5, for friction factor and pressure drop in a packed bed becomes

f =

"p#Dp$ 3 L(v ') %(1 & $)

2

150(1 & $) + 1.75 v ' #Dp %

(6)

"p 150v # (1 % &)2 1.75 '(v #) 2 1 % & = 2 2 + L $ Dp & 3 $Dp &3

(7)

By substituting .'mf for .', %m for % and Lmf for L in equation 7 and combining the result with
f

equation 3, the minimum fluid velocity .'mf at which fluidization begins can be calculated.
3 Dp "( " p # ") g 2 2 2 1.75 Dp (' % mf ) " 3 &$mf 2

2
is the fluid viscosity

150(1 # $mf ) Dp v % mf "

3 &2$mf

(8)

Where the left hand term of Equation 8 is called Archimedes number (Ar)

For fluidized bed, defining a Reynolds number as

NRe .mf =

Dp vmf ! "

Eq. (8) becomes

75
3 Dp "(" p # ")g

Ar =

150(1 # $mf )(N Re.mf ) 1.75( N Re.mf ) 2 = + 3 3 %2$mf %$mf

(9)

On the right hand of Eq. (9) first term in NRc.mf tend to dominate under laminar flow and the second term in NRc2.mf under conditions with bed of large particles for which the flow is turbulent Fluidized bed heat transfer coefficient A vigorous bubbling fluidized bed is well mixed because of the particle circulation generated by the rising bubbles. Consequently, even when highly exothermic reactions occur in the bed, the bulk of the bed is practically at a uniform temperature. Also because of the high surface area exposed to the fluidizing gas by the particles, gas and particles come quickly to approximately the same temperature. In fluidizing bed a high rate of heat transfer can be obtained within the system. Because of the complexity of bed behavior it is not possible to predict the degree of particle circulation past the transfer surface and the bed-to-surface heat transfer coefficient and empirical correlations are of very limited value. For a Reynolds number range of 10-10,000 for gases in a packed bed of spheres, the recommended correlation with an average deviation of about 20% and a maximum of about 50% is

JH =

0.4548

NRe

"0.4069

(10)

JH is the dimensionless heat-transfer factor,

76 One reliable and straightforward correlation has been derived from simple tests in which a small calorimetric sphere was plunged into a hot bed and the heat transfer coefficient was derived from the rate of change of temperature of the sphere. This has the following form
0.2 0.2 -0.36

hmax = 35.8 #p kg

Dp

(11)

Where h is the maximum heat transfer coefficient kg is thermal conductivity to the gas

APPARATUS DESCRIPTION A bed of a chosen granular material about 70mm deep is contained in a vertical glass cylinder. At the lower end of this is a distribution chamber and air distributor, which supports the bed when de-fluidized. This distributor has been designed to ensure uniform airflow into the bed without causing an excessive pressure drop and is suitable for the granular material supplied. On leaving the bed, the air passes through the chamber and escapes to the atmosphere through a filter. The chamber, filter and distributor assembly is suspended from a bracket mounted on the panel. Installed in this bracket are probes for temperature and pressure measurement, and a horizontal cylindrical heater, all of which may be moved vertically to any level in the bed chamber. Air from the local compressed air supply is delivered through an internal filter and air inlet control. This regulates the high pressure supply to a constant low pressure suitable for introduction into the fluidized bed. From the air inlet control the reduced pressure air is brought to the flow control value and then to the distribution chamber.

77 The electrical supply to the heater is controlled by an electronic heater control and the power supplied is displayed directly in Watts on a digital Wattmeter. Two thermocouples are attached to the surface of the heater. One is used to record the surface temperature on a digital temperature indicator and the second in conjunction with an internal controller prevents the heater from exceeding a pre-set temperature of 200oC. The digital temperature indicators together with a selector switch display the surface temperature of the heater, the air supplied to the distribution chamber and the moveable temperature probe in the bed. A liquid filled manometer together with a moveable probe display the air pressure at any level in the bed material or chamber. A clean water filled tube on the front of the panel indicates the relative pressure inside the chamber: A maximum recommended pressure is marked on the panel above which the water level should not be allowed to rise. In the event that this pressure is exceeded (e.g. due to a blocked filter), the water will be forced from the tube and the excess pressure vented.

PROCEDURES: 1. Observation of the behavior of particles in a bed with upward air flow

The airflow should be turned to its maximum value and then steadily decreased. The bed movement and the nature of the bubble behavior should be continuously observed. When the airflow has been reduced to zero, the height of the bed should be noted. Then the bed chamber should be tapped by gentle shaking. If the airflow is now steadily increased the increase of bed volume to approximately its volume before tapping will be seen, just before any particle movement is observed. Further increase of the airflow rate from here will cause similar bubble behavior and movement to those seen earlier. Temperature distribution in a fluidized bed

78 With the bed well fluidized, the heater and temperature probe should be lowered into the bed. By adjusting the flow control and heater control, the bed temperature should be raised to about 20K above t3. If, without moving the heater, the probe is raised and lowered within the well fluidized bed, little variation of temperature will be noted. (This is an indication of the degree of mixing caused by the bubbles). If the probe is now raised into the air space above the bed, it will be found that the air is close to the bed temperature. (This is due to the very large surface area of the particles and the intimate contact between the air and bed material). Other bed materials may be charged into the bed chamber and their behavior observed in a similar manner. 2. Investigation of the relationship between bed height, bed pressure drop and upward air

velocity through a bed of granular material. Ensure that the manometer has been zeroed. Raise the heater to the highest position and turn the heater control to zero and then lower the temperature probe into the bed. Turn the air flow control counter-clockwise to give maximum indicated air flow and allow the bed to mix thoroughly for two or three minutes. Without changing the airflow, note (i) (ii) (iii) (iv) The air inlet temperature t3 The bed material temperature t2 The bed height The pressure drops across the bed (i.e. the difference between pressures indicated

with the probe at its highest and lowest positions). Reduce the air flow rate in steps and repeat the observations at each setting.

79

(Note: It is advisable to draw the height and !p against flow rate graphs as the experiment proceeds so that the magnitude of the decrements can be estimated). Having reduced the flow rate to zero, sharply tap the bed chamber until the bed reaches a minimum height. Record this height (L0) Now repeat all the observations from zero airflow to maximum indicated flow at similar values to the reducing flow procedure above. At all stages make notes about the behavior of the bed. The test and observations should be repeated at other bed temperatures. (Note: The bed temperature may be increased before starting the higher temperature experiments by lowering the heater into the bed and switching it on for a few minutes with a moderate air flow).

3.

Investigation of the effect of,

(a) superficial velocity

(b) depth of immersion On the surface heat transfer coefficient for a hot surface in an air fluidised bed. Ensure that the manometer has been zeroed. Position the temperature probe at the same level as the center of the heater. Set the heater to a convenient height above the distributor (say 40mm). Turn the airflow control to give a high flow rate and allow the bed to mix thoroughly. Adjust the heater control until the desired heater surface temperature (t1) is attained (about 150oC is suitable).

80

Allow condition to stabilize, then note (i) (ii) (iii) (iv) (v) Heater surface temperature t1 Bed temperature t2 Air temperature t3 Heater Input Q Air flow rate Vm

Reduce the airflow, reset the heater control to obtain the same value of t1 then repeat the observations. Repeat in convenient steps until the flow control valve is closed. The observations should be repeated with the heater at other heights above the distributor, (say 20, 60 and 100mm). Note: Provided that the heater temperature (t1) is steady, the rate of heat transfer from it to the bed is displayed directly by the digital watt indicator.

The bed has a large thermal inertia, and, when fluidized, has little temperature variation. It is therefore not necessary to obtain complete stability of bed temperature. OBJECTIVES 1. Plot the pressure drop (!p) vs superficial velocity (v) at: 1. different bed temperature 2. v start from zero to maximum value 3. v start from maximum to zero value

2. 3.

Describe the bed height as you increase . from zero to vmf and v>vmf Determine the void fractions (%o and %mf)

81 4. Calculate the pressure drop at minimum fluidizing conditions and compare its value with the observed value !p. 5. 6. Calculate the minimum fluidizing velocity vmf and Ar use equation (9) Calculate friction factor f and plot it vs Re (for both experimental and theoretical) use equation

(8) where Re = 6.

vD p " (1 # $)

Determine the heat transfer coefficient (theoretical & experimental)

USEFUL DATA Heating Element: 12.7mm dia. x 57.1mm long Surface area 24 cm2 Granular Material: Fused Alumina (White Aluminum Oxide) Density 3770 kg m3 Grit size Average particle size Minimum particle size *Pour density approx. kg m3 Bed Chamber: Nominal diameter Nominal length Cross sectional area 105mm 220mm 8.66 x 103 m2 m m 54 320 210 1720 60 250 177 1670 80 177 125 1620 100 125 74 1560

Density of air: 1.2 kg m3 (at 101.3 kN m2 and 21oC) 1mm H20 = 9.81 N m2

82

*The shape factor * for the granular material supplied with this unit is about 0.73

83

84

85

References: 1. 2. 3. 4.

Christie J. Geankoplis "Transport Processes and Unit Operations" Third edition 1993, Prentice-Hall, Inc. McCabe, Smith, Harriott "Unit Operation of Chemical Engineering" Fifth edition 1993, McGraw-Hill, Inc. Morton M. Denn "Process Fluid Mechanics" 1980, Prentice-Hall, Inc. O.S.M. Botterill, Y. Teoman and K.R. Yuregir, AICHe Symp. Series, 77 (1981) No. 208, 330.

Saddawi December 2014

86

LIQUID MIXTURE VISCOSITIES Introduction and Theory: Liquid viscosities are needed by process engineers for quality control, while design engineers need the property for fixing the optimum conditions for the chemical processes and operations as well as for the determination of the important dimensionless groups like the Reynolds number and Prandtl number. Liquid viscosity is also important in the calculation of the power requirements for the unit operations such as mixing, pipeline design, it cetera. The flow characteristics of liquids are mainly dependent on the viscosity and are broadly divided into Newtonian, when the viscosity of a liquid remains constant and is independent of the applied shear stress. And Non-Newtonian, when liquids, viscosity depend on the applied shear force and time. Viscosity is a fundamental characteristic property of all liquids. When a liquid flows, it has an internal resistance to flow. Viscosity is a measure of this resistance to flow or shear. Viscosity can also be termed as a drag force and is a measure of the frictional properties of the fluid. Viscosity is a function of temperature and pressure. In this experiment students will deal with the viscosity of pure liquids such as glycerin and water and its change as function of temperature. As well as student will exam the viscosity of mixture solution (glycerin + water) and its change as a function of mole fraction and temperature changes. Viscosity is expressed in two distinct forms absolute or dynamic viscosity and kinematic viscosity. Dynamic viscosity is the tangential force per unit area required to slide one layer against another as when these two layers are maintained at a unit distance. In another words, the viscosity of a fluid is defined as the measure of how resistive the fluid is to flow, and it can be described as: Shear stress = * (Strain or shear rate) !!!!! where is the dynamic viscosity $ is shear stress and is strain rate From the definition of the is strain rate (1)

87

! ! ! !" ! !

! !"

(2)

where x is the distance, t is the time, and dx/dt is the velocity v. Therefore, the dynamic viscosity can be written as !!!
! !

(3)

While kinematic viscosity requires knowledge of density of the liquid (% ) at that temperature and pressure and is defined as ! !!! (4) Consider a Newtonian fluid laminar flow through a glass cylindrical capillary tube (viscometer) of diameter r and length l, as shown in Fig. 1 with no slip at the wall of a pressure difference &P between the two ends and the fluid is subjected to a force F. Then the viscosity may be calculated by applying Newtons Law of viscosity.

r F R P1 l

P2

88

! ! ! !"

!"

(5)

make a force balance on the cylindrical element.

!!!"# ! !!"! ! Substitute the above Eq 6 into share stress Eq 5 yields
!" !"

(6)

! !!" !
! ! !! ! ! ! ! ! !!"

!!

(7)

Integration of Eq. (7), using the boundary condition that at the wall of the capillary v(R) = 0, gives !! (8)

Equation 8 shows that the velocity distribution through the capillary tube is parabolic, hence the total flow rate can be estimated by !! By combining Eqs 8&9 !! !!
! !! !! ! ! !!" ! ! !! ! !!" ! !!"#$# !

(9)

!! ! ! ! !"

(10) (11)

The above equation is known as Poiseuilles equation and it can be used for calculation of viscosity when using a capillary viscometer. For vertical tube arrangement which is the case for the capillary viscometer we use in this experiment, the hydrostatic pressure, %gh , depends on the height, h, of the liquid. Therefore, the pressure difference, &p , in terms of hydrostatic pressure is given by !! ! !"! (12)

If we know the volume (V) of the liquid dispensed during the experiment and the time (t) required for this volume of liquid to flow between two graduation marks in a viscometer,

!!

! !

(13)

by substituting eqs 12 & 13 into eq 11

89 !!
!"! ! ! !!"

!"

(14)

By assuming !! Eq14 will become

!"! ! ! !!"

(15) (16)

! ! !"# And the Kinematic viscosity is ! ! !" where C is the viscometer constant (mm2/sec2)

(17)

Students may calculate the viscometer constant C at a different temperatures other than those at the viscometers were calibrated by using the following equation ! ! !! !! ! !!! ! !! ! (18)

where C0 the viscometer constant when filled and tested at T0 (room temperature) B the temperature depending factor C0, T0, and B can be fond in the Viscometer certification paper. There has been no comprehensive theory on the viscosity of liquids so far because of its complex nature. However there are small number of semi-empirical and empirical methods provide reasonable results such as those proposed by Kirkwood et al.44 Where at temperatures below the normal boiling point, the logarithm of liquid viscosity varies linearly with the reciprocal of the absolute temperature as described below !" ! ! ! ! ! (19) where, A & B are constants and can be determined empirically. However, when the temperatures above the normal boiling point, the ln versus (1/T) relationship becomes nonlinear and in this case a number of semi empirical equation are needed. While for mixture viscosity the following equation proposed by Panchenkov which is basis on the mechanism of the flow of viscous liquid. !! !
! ! ! !! ! !! !! !!
!

!"#

! !"

!!

(20)

!! ! !" ! !! !! ! ! ! where m As !m T is the mixture viscosity, is an empirical coefficient, is the density of the mixture, is the absolute temperature,

90 "m R x ! Mg is the energy of bonds between the molecules in the mixture, and is the universal gas constant. is the mole fraction is the density molecular wt of glycerin = 92.09 g

Please note that The quantities As and "m of the eq (20) must be evaluated from the mixture viscosity data, and therefore at least viscosity of the mixture at two conditions must be known for the application of this method with good accuracy.

Apparatus and Method 1. Reverse Flow Viscometer (Cannon-Fenske Opaque Viscometers) 2. Water Bath controlling temperature 3. Glycerin liquid 4. Pycno meter to measure the density A schematic diagram of the Cannon-Fenske Opaque Viscometer apparatus is shown in Fig. 2. These reverse flow type viscometers wet the timing section of the viscometer capillary only during the actual measurement. The liquid sample flows into a timing bulb not previously wetted by sample, thus allowing the timing of liquids whose thin films are opaque. Reverse flow viscometers must be cleaned, dried, and refilled prior to each measurement. Clean the viscometer using suitable solvents, and dry by passing clean dry filtered air through the instrument to remove any traces of solvent. To charge the sample into the viscometer, invert the instrument and apply suction to arm L, immersing tube N in the liquid sample, and draw liquid to bulb D, up to mark G. Turn the viscometer to the normal position, wipe and clean it. When the meniscus travels through D and fills the bulb an up to its half, use a small rubber stopper to close the limb N to stop the flow of the liquid. Place the viscometer is in a thermostat by means of a proper holder and allow 10 -15 minutes to attain the required temperature. Subsequently remove the rubber stopper and record the efflux time for the liquid to pass through the bulbs C and J. By measuring the time required by the liquid to pass between the markings E, F and I, the viscosities may be calculated from the efflux times for the two bulbs. Students may compare between the values. It should be noted that any the reverse flow viscometers must be cleaned, dried, and refilled before a repeat measurement can be made. Repeat the measurement different glycerin/ water concentrations, and different temperatures like T = 10, 20, 30, 40, 50 oC and for several concentrations (in wt %)

91

N L

G I R

J F

C E

92 Figure 2. Cannon-Fenske opaque reverse flow viscometer Table 1. Various size Cannon-Fenske opaque reverse flow viscometer Size No. Kinematic Viscosity Range, mm2/s (cSt) 0.4 - 2 0.8 - 4 1.6 - 8 3-15 7-35 20-100 50-200 100-500 240-1,200 500-2,500 1,600-8,000 4,000-20,000 Capillary Diameter, Viscometer mm ( 2%) Constant (C) mm2/s2 0.31 0.42 0.54 0.63 0.78 1.02 1.26 1.48 1.88 2.20 3.10 4.00 0.002 0.004 0.008 0.015 0.035 0.1 0.25 0.5 1.2 2.5 8 20

25 50 75 100 150 200 300 350 400 450 500 600

References: 1. J. Welty, C. Wicks, and R. Wilson, Fundamentals of Momentum, Heat, and Mass Transfer John Wiley and Sons, Inc., New York Third edition (1983). 2. A. Dinsdale and F. Moore, Viscosity and its measurement, Chapman and Hall, London (1962). 3. M. R. Cannon and M. R. Fenske, Viscosity measurement, Ind. Eng. Chem. Analytical Edition. 10(6), 297-301 (1938). 4. M. R. Cannon and M. R Fenske, Viscosity measurement, Ind. Eng. Chem. Analytical Edition. 13(5), 299-300 (1938). 5. M. R Cannon, R. E Manning, and J. D. Bell, Viscosity measurement, the kinetic-energy correction and a new viscometer, Anal. Chem. 32, 355-358 (1960). 6. Y. A. Pinkevich, New viscometer for the determination of the viscosity of petroleum products at low temperatures, Petroleum (London), 8, 214-215 (1945). 7. ASTM (American Society for Testing and Materials) D 88, Standard Test Method for Saybolt Viscosity. 8. C. R. Duhne, Viscosity-temperature correlations for liquids, Chem.Eng. 86 (15), 83 (1979). 9. G.M. Panchenkov, Calculation of absolute values of the viscosity of liquids, Zhurn. Fiz. Khim. 24, 1390-1406 (1950). 10. M. J. Lee and M. C. Wei, Corresponding-states model for viscosity of liquids and liquid mixtures, J. Chem. Eng. Japan, 26(2), 159-165 (1993).

Saddawi January 2014

93

EXPERIMENT UNSTEADY STATE HEAT TRANSFER THERMAL DIFFUSIVITIES AND HEAT TRANSFER COEFFICIENTS Objective Observe unsteady state heat flow to or from the center of a solid shape (qualitative, using data acquisition) when a step change is applied to the temperature at the surface of the shape. Using analytical transient- temperature heat flow chart (graphical method), method of Separation of variables, and Polynomial approximation method, to determine the thermal conductivity of a solids and the heat transfer coefficient. Investigate the effect of shape, size, and material properties on unsteady heat flow Compare the results obtained using these methods with literature values, and determine how accurate each method is. Determine the temperature profiles as a function of position and time. Investigate the Lumped Thermal Capacitance method of transient temperature analysis. Apparatus
The unit consists a stainless steel water bath and integral flow duct with external water circulating pump. The bath has a capacity of approximately 30 liters and is heated by a thermostatically controlled 3 kW electric heating element in the base of the bath. Adjustment of the thermostat allows the bath to be set to a constant temperature (T1) before beginning the experimental procedure.

Figure 1 It also consist seven simple shapes of different materials (Stainless steel and brass) are provided as shown in Figure below and each is fitted with a thermocouples well at its geometric center and thermocouple and into the carrier to sense water temperature adjacent to the shape to allow

94
measurement of its core temperature (T3) and the surface (T2). Also there is a pump has a 3-speed selector switch, to circulate the water into the unit.

Figure 2

Procedure: 1- Turn on the power supply to the Unit and turn on the 16A bleater miniature circuit breaker
(MCB).

2- Ensure that the red power indicator adjacent to the thermostat is illuminated. 3- Turn the thermostat to position 6 for fastest heating. The water will take approximately 30 4- Having achieved the desired temperature T1, say 85 oC, reduce the thermostat setting to 5- The water bath temperature T1 should not be allowed to exceed 85-90 0C 67- Attach the required test shape to the carrier by tightening the coupling nut (finger-tight). 8- Insert the T3 and T2 probes to engage fully into the center and adjacent to the shape of the
shape. position 2. It will cycle ON/OFF to maintain the existing temperature. minutes to heat from cold. At this setting the water will boil, if left unattended.

9- Avoid touching the shape by hand to reduce thermal effects and place the shape on the bench
to reach ambient temperature. 10- Record T3, T2, and T1. When all the temperatures reach to same value as T1, 11- Repeat the procedure for other objects.

12- Repeat the experiment as many times as possible to determine the error in the measurements. Consider the quantitative effects of sampling error, particularly at short and long times, on your data analysis. Consideration should also be given to the

95 source and consequences of discrepancies between the proposed model and experimental apparatus, and the method of analysis in regard to the accuracy of your results. 13- When the work is completed, then turn off the RCCB and mains supply to the Unit. Turn off
the main switch on the Heat Transfer Service Unit and turn off the mains supply.

14- If it is necessary to drain the bath allow the bath to cool completely. Note
The heating element incorporates internal thermal protection so that power is switched off in the event of the heater being turned on when not covered by water. However, repeatedly allowing the heater to overheat in this wax should be avoided as it will eventually fail.

Running the Thermal Conductivity Program X. XI. XII. XIII. XIV. Create a new folder on the computer desktop for your work group. Open the folder called Heat Conduction on the disk top. Launch Lab View (double click on the file heat conductivity). Start the program by clicking on the arrow located in upper left-hand corner. Choose a name for the file that will be created by the run and save it to the new folder you created in the first step. Note: Changes during the run 4. The graph view can be altered. Holding down the mouse button with the hand cursor positioned over the graphing area will allow the graphing view to be moved to a desired position. 5. The X, Y axis can also be changed while the program is running. Highlight the value that is to be changed then type in the new value and hit-enter. XV. Calibrate the readings to room temperature and the bath by using a thermometer before submersing the test blocks. XVI. End the run by hitting the Stop button. XVII. Launch Microsoft Excel from the menu and open the new file created from the run. Note: The default setting for data acquisition is one data point per second.

Introduction and theory: Heat transfer generally takes place by conduction, convection, and radiation. Conduction is the transfer of thermal energy between adjacent molecules in a body due to a temperature gradient and can take place in any matter such as Plasma, Gases, Liquids, and Solids.

96 In solids, conduction is due to the combination of vibrations of molecules in a lattice and energy transfer by free electrons. For Many applications, we required to understand of complete time history of temperature variation and distribution within the body. Such as design insulation thickness materials, chip temperature calculation, heat treatment of metals, which control the characteristic of the processed materials, such as softening and hardening. In absence of internal heat generation, when a cool, solid body is placed in a warm environment, energy in the form of heat flows into the body until it attains a thermal equilibrium with its surroundings. Starting at time, t = 0, the temperature at any point in the solid increases until the body reaches an isothermal steady state. The period of time required for the body to reach steady state is a function of the size, shape, and physical properties of the body as well as the conditions imposed on the surface of the body by the experimental conditions. In another words, in unsteady transient heat conduction, temperature is a function of both time and spatial coordinates. Fouriers Equation governs the temperature response of a body. It is possible to determine the thermal conductivity of the body by recording changes in its central temperature as a function of time and comparing this experimental data with temperature profiles predicted using the Thermal Energy Equation (1st Law of Thermodynamics for closed system). ! ! !!! !! ! !! ! ! ! !!" !! !! ! !! !!..[1] !!! !! ! !! ! !! ! !" ! !
!" ! !! !!

! !

! ! ! ! !! ! ! ! !! ! ! !

..[2]

Where ! ! !! ! !! ! !! ..[3] Rate of heat transfer by conduction from the center of the body out side or in opposite direction can be describe by Fouriers Law !!!!!! ! !!!!!!!!! ..[4] For instance, when the body is a metal semi infinite (slab or cylinder) or sphere, for onedimensional case the governing energy equation is: !!! !" ! ! ! !" ! ! ! !" ..[5] Where for slab m=0 for cylinder m=1 for sphere m=2
! !" ! !" !" ! ! !"

! ! ! !" ! ! !" ..[6]

! !

!"

Where ! ! !! Thermal diffusivity

!

97 To solve the above equation, analysis of transient heat conduction is needed. We need, two boundary conditions in one direction (to ease the solving) and the initial condition and temperature history at internal line of symmetry. The second possible approximate solutions of adequate accuracy for most of engineering problems are the experimental analysis. Biot and Fourior numbers In some transient problems, the internal temperature gradient is very low (high thermal diffusivity), thus the object temperature at any location may change rapidly to reach to steady state (the surrounding temperature). Assume transient heating problem of a cool object being heated, considering a simple case of steady one-dimensional heat transfer as shown in the Fig. 3 below. Heat flux will transfer by conduction through the object body from the surface toward the center. ! ! ! !!"# !
!! !!! !

..[7]

T0 Ts

Ti

X or R

Convection External resistance =1/h

Conduction Internal resistance

Figure 3

98

r R

x Sphere Radius R

Semi-Infinite Slab thickness 2x

Semi-Infinite Cylinder Radius R

Figure 4

And same heat flux will transfer from the surrounding to ward the objects surface area by convection. ! ! ! ! !!"# ! !! ! !! ..[8] Where (2,Rl) is the surface area of the object =As R or X is the material thickness and ! is the average heat transfer coefficient for the convection from entire surface.

99 If the object is not cylinder or have no regular shape, then a characteristic length scale (L) may be used to substitute the material thickness and the surface area of the object. Where (L) may be obtained by dividing the volume of the solid by its surface area. ! ! ! ..[9]
!

By dividing equations 7&8 will get

!! !!! !! !!!

!! !

! !"#$!!"#$%&..[10]

And in transient cooling case equation (10) becomes as

!! !!! !! !!! !! !

! !"#$!!"#$%&..[11]

Biot number is a dimensionless number and it represents the ratio of: ! !"#$%&'()*!!"#$#%&'("! ! !"#$!!"#$%& ! ! !"#$%&'($%!!"#$#%&'("! ! Low Biot number (Bi < 0.1) Biot number less than 0.1indecates that the conduction thermal resistance is practically negligible compared with the convection resistance and the internal temperature gradients are very small. The transient thermal response can be treated by the (Lumped Thermal Capacity) approach. The Lumped thermal capacity assumption suggests that the object has a uniform mass temperature. Energy balance for the heat transfer between the surrounding fluid and the object is:

! !! !!! ! !!!!! !!!

Solid object T(t) #, Cp V

T0

Figure 5

!"#$!!"#$%&!!!!!!"#$%&! !"#$%&'%!!"!!!!!!"#$ ! ! !!!"#$%"&' ! ! !"!#$%!!"#$%&!!"#$!!" !"#$%&!!"#$!!"

100 !!! !"# ! ! !!! !!!! ! !! ..[12]

!" ! !! !!!!

! !! !! !"..[13]
!

!!

Integrate and apply initial condition to equation (13) The appropriate initial condition is: IC: t = 0 T = Ti
! !!! !! !!!

!!

! ! !"# ..[14]

where ! ! !

!
!

The quantity [ ! ! !! ] is called the Time Constant, it has (time unit) (e.g. sec) Substitute 1 in equation (14)
! !!! !! !!!

!! !

! ! !! ..[15]

Rearrange the exponent term in equation (15)

! !

! !!

!!
!!

!! !

! !!! !!

! ..[16]

Knowing that: ! ! ! Thermal diffusivity !!

!

and !! ! !"#$!!"#$%& ! See the second term in the right of equation (16) is Non dimensional time parameter abbreviated (Fo) Fourier modules. ! ! ! ! !" ! !! !! ! ! !! ! ..[17]
!

2Equation (15) becomes

! ! ! !!! ! !!!"!!" ..[18] ! ! Notice that equation (18) solution for transient conduction for Lumped parameter (low Bi number). High Biot number (Bi > 0.1) In this case the conduction resistance could not be neglected, and the temperature distribution in the solid object is a function of time and position within the body.

!!!

101

For a semi-infinite solid object as shown in Fig. 2 the energy balance can be expressed as: At r=R (Second boundary condition)

T ! !!!! ! !!!!! !!!

!" ! ! !!"! !!!! !"

T0 r Figure 6 x=R

!"#$!!"#$%!&%' ! !"#$!!"#$%!&'$ ! ! !!!"#$!!!!!!!"#$%&' !!"#$!!!!!!"#$% !!!!! ! !!!!! ! !!"!!" !!!! ..[19] But when r do not equal R, then !!
!!!! !!! !! !!! !"

..[20]

Boundary condition !" At r=0 T is finite and !" ! ! At t = 0, T=Ti The governing equation (6) can be written in dimensionless form by defining the following variables: Where !! !! ! ! !!! !! ! !! !

! ! !! !! ! !!

102 In dimensionless form equation (6)

!" !"

! !! !" !! !" ..[21]

!"

Where at The initial condition, at - = 0 at 4 = 1 3=1

!" !"

! !!" ! !..[22]

The resulting equation (22) contains only one dimensionless parameter, Bi, the Biot number for heat transfer. Therefore, the solution, in its dimensionless form, must be a function only of this parameter. (The student should note the distinction between the Biot number and the Nusselt number for heat transfer). The transient heat transfer equations (21 or 22) maybe solved by: 1- Graphical method using Heisler Charts for different geometries. 2- Analytical method of separation variables or the Laplace Transform method. 3- Polynomial approximation method for different geometries. Analysis 1Graphical method may be used: This method is based on the solution of Eq. (21). Graphs of this equation are presented in Appendix F of ref. 1. First, use the results obtained for the steel cylinder since its k value is given. Using Heisler chart together with the experimental results in terms of 3 and 5t/L2 find 1/Bi for each time interval. Determine Bi for steel by interpolating a curve through the data points and reading the corresponding Bi from the graph. Knowing k and Bi determine h. In order to determine k for the other cylinder having same radius, prepare a master plot of the theoretical solution of Eq. 21 in terms of 3 versus ht/# Cp L = Bi* - for several Bi value. Re-plot the experimental data for the brass cylinder in this new graph using the value of h obtained for steel cylinder. Obtain the Bi value from the curve that coincides with the experimental data and determine k from Bi and h. Repeat the calculation to find h (heat transfer coefficient for other steel cylinders of different radius. Same method may be repeated for slab and sphere objects. By known h and k, one can find temperature as a function of position by using the provided graphs on Figures 7, 8, and 9. 2Analytical method of separation variables or the Laplace Transform method:

103 Using the method of separation of variables3,4 or the Laplace Transform method,4,3 the solution of Eq. (21) is !!!
!" !! !!! ! ! ! !!! !" ! !! ! ! !! ! !"#!!! ! ! ! ! !! !! ..[23]

where qm are the roots of the transcendental characteristic equation, !! !!! !! ! !" !!! !!! ! ..[24]

where Jo and J1 are the zeroeth and first order Bessel functions, respectively. After a small amount of time has passed (@ ten seconds), the leading term in the infinite series expansion, Eq. (23), dominates the series, i.e.,
! !"# ! !! !! ! ! !! !! !! ! ..[25]

and so all terms associated with the higher order eigenvalues, qm, for which m ' 2, may be ignored and the solution is given approximately by the single term ! ! !!" ! !!! !
!

!!"

!! !!! ! ! !! !!! !

! !"#!!!!!! !! ..[26]

Graph ln 3 versus time (real time t), yield straight line of slope, !"#$% ! ! !!! !! ..[27]
!

!!!

and Intercept, ! ! !"!

! ! !! !" !! !!! !

!!" !! !!! !

! !

..[28]

By solving equation (24) together with the slope and the intercept (eqs 27 and 28), one may evaluate the three unknown parameters, or can be done graphically using the following table.

104

q(r) 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2

J0 (r)/J1(r) 3.181 2.678 2.295 1.989 1.739 1.528 1.347 1.188 1.046 0.9174 0.7791 0.6888 0.5847 0.4846 0.3882

3- Polynomial approximation method for different geometries: Using polynomial approximation to solve equation (21) yield ! ! !"#!!!!"!.(29) Where p is the modified Biot number For Cylinder Slab: ! ! !"!!..[30]
!!"

Sphere

! ! !"!!..[31] ! ! !"!!..[32]
!"!"

!!"

By graphing ln 3 versus t, we may use the slope to evaluate k or h. Note: Please see the lecture for full derivations.

105

USEFUL DATA: Brass Test Shapes Properties Thermal conductivity Specific heat Density Thermal diffusivity k Cp % ( 121 W/mK 385 J/kg 7930 kg/m3 3.7*10-5 m2/s

Stainless steel Test Shapes Properties Thermal conductivity Specific heat Density Thermal diffusivity PUMP: Speed I II III References 1. Welty, J. R., C. E. Wicks and R. E. Wilson, Fundamentals of Momentum, Heat and Mass Transfer, Third Edition, John Wiley and Sons, New York, 1984. 2. Bird, R. B., W. E. Stewart and E. N. Lightfoot, Transport Phenomena, John Wiley and Sons, New York, 1960. 3. Carlslaw, H. S. and J. C. Jeager, Conduction of Heat in Solids, Oxford University Press, Cambridge, England, 1959. 4. Handbook of Mathematical Functions, M. Abramowitz and I. Stegun, Eds. National Bureau of Standards, Washington, D. C., 1962. k Cp % ( 16.3 W/mK 460 J/kg 8500 kg/m3 0.45* 10-5 m2/s 36 Watt 46 Watt 52 Watt

Salma Saddawi January 2014

106

Figures 7, 8, and 9

107

108

109

PI-plus-Feedforward Water Level control For Coupled Water Tanks Chemical Engineers to day need comprehensive knowledge of the basic principles and the advanced technique in process control due to the fact that the modern industries are designed to use controlling system to control the chemical process, such as food processing, nuclear power generation plants, pharmaceutical industries, water purification systems, industrial chemical process, etc.. Where automation and diagnostic process control take place. The typical actuators used in liquid level control systems are level sensors like floating body, pressure sensor, in addition to on/off valves, pumps, etc. In this laboratory, students will exercise to derive a model to achieve and maintain a desired set point water level by implementing a Mathematical models of a Coupled-Tank system were to track the trajectories of the water level in the top and the bottom tanks based upon the voltage applied to the systems pump. The configuration of these models in MATLAB provided the implementation of real-time PI (Proportional-Integral) and Feed-Forward control, along with the ability to analyze and optimize the actual performance of the system. Apparatus The apparatus for this experiment consisted of the Quanser Consulting Water Level Control Two!Tank system, the Quanser VoltPAX!Xq single channel linear voltage amplifier, and a Lenovo computer with MATLAB software. The two!tank system consisted of a pump with a water basin and two 30!mL tanks, each with an area of 15.5179 cm2

In addition, the apparatus contains three different orifices with areas of small, medium and large, respectively.

110 Procedure Because this lab is extremely MATLAB intensive, a series of assignments should be completed before any data can be gathered. Ideally, students will complete the first three assignments for the one tank system, which calculate necessary set points, voltages, and values for the specific controller scheme, and then initialize those values in MATLAB. Before an experiment can be performed, Students should also calibrate the pressure sensors to ensure a proper water level height reading in the MATLAB window. Then students may complete all assignments that found in the attached student user manual for this experiment.

Saddawi July 2014

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142 Appendix A A NOTE ON STATISTICS In analyzing your experimental data you will make use of linear regression. It is important that you also estimate the errors in your fitting parameters, and the errors in your model predictions. I offer here a few simple expressions you can use in this analysis. More detailed information and background can be found in statistics texts, or off the net in the CHEG 258 notes for last spring. The relevant pages are pp 76 - 118 on 2/12/97 to 2/28/97. There are also several matlab example programs which illustrate the statistical analysis procedures. Suppose we have a set of n observations bi which are made at some times or positions ti (this depends on the problem you are looking at, of course). We wish to fit the data with an m parameter linear regression model:
m

bi j= 1

xj !j (ti)

This can be written most easily in matrix form: b)Ax where the matrix A is the n x m matrix of modeling functions. The m x 1 column vector x is the vector of modeling parameters. Note that the modeling functions do not have to be linear functions of t, rather linear regression simply requires that the model be linear in the modeling parameters- x. The unweighted least squares minimization problem is simply the set of parameters x which minimizes the 2-norm:

min || A x - b ||2 x
It can be shown that the vector x is the solution to the m x m set of linear equations: AT A x = AT b The solution to the problem is just: x = [ATA]-1 AT b = K b where the matrix K is defined as K = [ATA]-1 AT. This is a very convenient relationship as it immediately illustrates how the fitting parameters are related to the observed variables b. Specifically, the parameters are linear combinations of these variables, and as a consequence it is easy to calculate the variance of the parameters! If the each of the observations b are normally distributed random variables, then the fitting parameters will also be normally distributed.

143

Recall that the variance of a random variable z with mean z is given by: 6z2 = E{(z - z)2} where E denotes the expectation value over many realizations. If z is a vector instead of a scalar, we can define the matrix of covariance: 7 z2 = E{(z - z) (z - z)T} where the diagonal elements of this square matrix are just the variances of each of the individual elements of z. The off-diagonal terms represent the covariances of each of the elements of z, and will have non-zero values if the variability of, say, z1 and z2 are not independent. If each of the elements are statistically independent then the matrix will be diagonal. Using all this, we can determine the matrix of covariance of the fitting parameters. We get: 7 x2 = E{(x - x) (x - x)T} = K E{(b - b) (b - b)T} KT = K 7 b2 KT where the matrix of covariance of modeling parameters is thus determined in terms of the matrix K calculated from the matrix of modeling functions and the matrix of covariance of the observations. Usually you do not know the error in the individual observations b , so it is useful to calculate these from the scatter in the data itself. If we make the assumptions that the error in each of the observations is independent and that they are equal, the covariance matrix 7 b2 just becomes the product of 6b2 and the identity matrix: 7 b 2 = 6b 2 I Assuming that the deviation of the data from the best fit model is due to random error (not systematic bias) we can calculate the variance 6b2:

!2 b=

1 || A x - b ||2 n-m

where n is the number of observations and m is the number of modeling parameters. With these assumptions the matrix of covariance of the fitting parameters becomes: 7 x2 = 6b2 K KT with the diagonal elements being the elements of particular interest.

144 As a final note, consider the predictions based on the best fit modeling parameters. Often we wish to extrapolate or interpolate our experimental results based on the fitted parameter values. These predictions will also have some error associated with them, and it is useful to calculate this statistical uncertainty. Suppose we wish to predict the value of b at some time t. We thus have: bp = a(t) x = a(t) K b where a(t) is the row vector (1 x m) of modeling functions evaluated at time (or position) t. Again, we find that the prediction is a linear combination of the original observations. The variance in the predicted value bp is thus given by: 6bp2 = a(t) K 7 b2 KT aT(t) For independent, constant variance in the observed values b, this reduces to: 6bp2 = 6b2 a(t) K KT aT(t) The variance in the predicted value for a number of values of t can be determined simultaneously by replacing the array a(t) with a matrix. In this case only the diagonal elements of the resulting matrix 7 bp2 have any importance. The equations presented here are trivially simple to implement using Matlab. examples in the CHEG258 course notes will give you some ideas on how to do this. The

All of the above equations have been developed for unweighted linear regression. If the errors of each of the elements of b are known independently, and if they are not the same, then it is more appropriate to do weighted linear regression. I refer you to the CHEG258 lecture notes for a discussion of this procedure. This discussion of error has been limited to random error only - we have tacitly assumed that the only deviation between the experimental results and the fitted model is due to random measurement error. As can be easily seen from the above formulas, if the number of observations used in the fit goes to infinity, the calculated error in the fitting coefficients due to random fluctuations goes to zero. Of course, as you should all be aware, there is also systematic error present in any experiment, and the assumed model may also not be taking account of all relevant physical phenomena. Thus the statistical error bounds that you can calculate using the above techniques will not necessarily give you a true picture of the accuracy of your results, and really must be regarded as a lower bound on the total error which is comprised of both random and systematic sources. In the limit of large numbers of replications, the statistical error (the only source of error about which you have information from observing random fluctuations) will be negligible relative to the systematic or modeling error. One way of looking for sources of systematic error is to plot your residuals - the deviation between the fitted model and the experimental observations - and determine if the deviation is random or not. If it is non-random, then there is a physical process influencing your

145 data which is not included in the model you are describing it with, and your error bounds on fitting parameters will be too low. Even if the residuals are truly random, it still does not mean that there is no systematic error in the measurements. The only way to properly gauge the magnitude of systematic error is to replicate the measurement using a completely independent technique - often a difficult or impossible thing to do! Thus in papers or publications you will read, the error estimate is usually that given from statistical sources, which may have little bearing on the accuracy of the final result. Note on Accuracy and Reproducibility. Accuracy refers to the measurement error inherent to an instrument/technique, and for most instruments is typically specified by the manufacturer. For example, a pressure gauge may have a range of 0-100 psia with an accuracy of 0.02 psi. The accuracy tells you how well the average value of repeated measurements with the same instrument (the one you bought) compares with values obtained with other types of techniques/instruments. Reproducibility tells you how well an experimental value of a quantity can be reproduced from experiment to experiment. In order to determine the reproducibility you will need to repeat the experiment starting from entirely new conditions. For example, in the Azeotropic Distillation experiment you will need to repeat at least three times the following: 1) prepare the toluene/isopropanol solutions to produce the refractive index calibration curve, 2) prepare the 50/50 toluene/isopropanol solution to be used in the experiment, 3) startup the distillation apparatus and measure the relevant variables (temperature, pressure, azeotropic composition). The reproducibility can be specified (to the least) as a standard deviation from the average (example: azeotropic composition Zaz=713 mol%, where 71 is the average and 3 the standard deviation). By reproducing an experiment several times you will be able to reduce random errors, but not systematic errors.

146 Appendix A: Sample Output. clear echo on format compact %This program shows how you can use Matlab %to do a simple error calculation. The example %is determining the acceleration due to gravity %by measuring the height of a ball as a function %of time which is initially thrown upward with %some velocity. Thus: %First we set up the array of times: t=[0:.125:2.5]'; %Note that we were careful to make this %a column vector! %Next we generate some artificial 'data' %with random noise: b=(-0.5*t.^2+1.25*t)*9.8+randn(size(t)); %This is the matrix of modeling functions: a=[ones(size(t)),t,t.^2]; %And this is the matrix k used in the solution: k=inv(a'*a)*a'; %Here are the best fit modeling parameters: x=k*b %Now we estimate the variance in b: [n m]=size(a); varb=norm(a*x-b)^2/(n-m); %This yields the standard deviation in b: stddev=varb^.5 %which is close to the expected value of unity %for the random routine. %We use the variance in b to calculate the variance %in x: varx=varb*k*k' %We want to generate a smooth curve for our %best fit curve:

147 tp=[0:.1:2.5]'; ap=[ones(size(tp)),tp,tp.^2]; bp=ap*k*b; %And we plot this up with the data: plot(t,b,'o',tp,bp) xlabel('time (sec)') ylabel('height (m)') drawnow hold on %Now we calculate the variance in the model %predictions: varbp=varb*ap*k*k'*ap'; %And we extract the square root of the diagonal %to get the standard deviation: sigbp=diag(varbp).^.5; %We add the 95% confidence interval to the plot: plot(tp,bp+2*sigbp,'w--',tp,bp-2*sigbp,'w--') hold off %And we finish by outputting the best fit parameter %values and their standard deviations: [x,diag(varx).^.5] %The first coefficient is just half the acceleration %due to gravity, thus the estimated gravitational %acceleration, with random error is: gravity = [-x(3),varx(3,3)^.5]*2 %which is reasonably close to the expected value of %9.8 m/s^2.

148 format compact %This program shows how you can use Matlab %to do a simple error calculation. The example %is determining the acceleration due to gravity %by measuring the height of a ball as a function %of time which is initially thrown upward with %some velocity. Thus: %First we set up the array of times: t=[0:.125:2.5]'; %Note that we were careful to make this %a column vector! %Next we generate some artificial 'data' %with random noise: b=(-0.5*t.^2+1.25*t)*9.8+randn(size(t)); %This is the matrix of modeling functions: a=[ones(size(t)),t,t.^2]; %And this is the matrix k used in the solution: k=inv(a'*a)*a'; %Here are the best fit modeling parameters: x=k*b x= 0.8069 11.5519 -4.8007 %Now we estimate the variance in b: [n m]=size(a); varb=norm(a*x-b)^2/(n-m); %This yields the standard deviation in b: stddev=varb^.5 stddev = 0.9368 %which is close to the expected value of unity %for the randn routine. %We use the variance in b to calculate the variance %in x: varx=varb*k*k' varx = 0.3127 -0.4876 0.1586

149 -0.4876 1.0744 -0.4006 0.1586 -0.4006 0.1602 %We want to generate a smooth curve for our %best fit curve: tp=[0:.1:2.5]'; ap=[ones(size(tp)),tp,tp.^2]; bp=ap*k*b; %And we plot this up with the data: plot(t,b,'o',tp,bp) xlabel('time (sec)') ylabel('height (m)') drawnow hold on %Now we calculate the variance in the model %predictions: varbp=varb*ap*k*k'*ap'; %And we extract the square root of the diagonal %to get the standard deviation: sigbp=diag(varbp).^.5; %We add the 95% confidence interval to the plot: plot(tp,bp+2*sigbp,'w--',tp,bp-2*sigbp,'w--') hold off %And we finish by outputting the best fit parameter %values and their standard deviations: [x,diag(varx).^.5] ans = 0.8069 0.5592 11.5519 1.0365 -4.8007 0.4003 %The first coefficient is just half the acceleration %due to gravity, thus the estimated gravitational %acceleration, with random error is: gravity = [-x(3),varx(3,3)^.5]*2 gravity = 9.6015 0.8006 %which is reasonably close to the expected value of %9.8 m/s^2.

150

10

6
height (m)

-2

0.5

time (sec)

1.5

2.5