Fatigue Testing and Analysis: Theory and Practice

Division.

TRANSDUCERS AND DATA ACQUISITION

RICHARD. HATHAWAY

WESTERN MICHIGAN UNIVERSITY

KAH WAH. LONG

DAIMLERCHRYSLER

1.1 INTRODUCTION

This chapter addresses the sensors, sensing methods, measurement systems, data acquisition, and data interpretation used in the experimental work that leads to fatigue life prediction. A large portion of the chapter is focused on the strain gage as a transducer. Accurate measurement of strain, from which the stress can be determined, is one of the most significant predictors of fatigue life. Prediction of fatigue life often requires the experimental measurement of localized loads, the frequency of the load occurrence, the statistical variability of the load, and the number of cycles a part will experience at any given load. A variety of methods may be used to predict the fatigue life by applying either a linear or weighted response to the measured parameters.

Experimental measurements are made to determine the minimum and maximum values of the load over a time period adequate to establish the repetition rate. If the part is of complex shape, such that the strain levels cannot be easily or accurately predicted from the loads, strain gages will need to be applied to the component in critical areas. Measurements for temperature, number of temperature cycles per unit time, and rate of temperature rise may be included. Fatigue life prediction is based on knowledge of both the number of cycles the part will experience at any given stress level during that life cycle and other influential environmental and use factors. Section 1.2 begins with a review of surface strain measurement, which can be used to predict stresses and ultimately lead to accurate fatigue life prediction. One of the most commonly accepted methods of measuring strain is the resistive strain gage.

1.2 STRAIN GAGE FUNDAMENTALS

Modern strain gages are resistive devices that experimentally evaluate the load or the strain an object experiences. In any resistance transducer, the resistance (R) measured in ohms is material and geometry dependent. Resistivity of the material (ρ) is expressed as resistance per unit length × area, with cross-sectional area (A) along the length of the material (L) making up the geometry. Resistance increases with length and decreases with cross-sectional area for a material of constant resistivity. Some sample resistivities (μohms-cm²/cm) at 20°C are as follows:

Aluminum: 2.828

Copper: 1.724

Constantan: 4.9

In Figure 1.1, a simple wire of a given length (L), resistivity (ρ), and cross-sectional area (A) has a resistance (R) as shown in Equation 1.2.1:

FIGURE 1.1 A simple resistance wire.

(1.2.1)

If the wire experiences a mechanical load (P) along its length, as shown in Figure 1.2, all three parameters (L, ρ, A) change, and, as a result, the end-to-end resistance of the wire changes:

FIGURE 1.2 A resistance wire under mechanical load.

(1.2.2)

The resistance change that occurs in a wire under mechanical load makes it possible to use a wire to measure small dimensional changes that occur because of a change in component loading. The concept of strain (ε), as it relates to the mechanical behavior of loaded components, is the change in length (ΔL) the component experiences divided by the original component length (L), as shown in Figure 1.3:

FIGURE 1.3 A simple wire as a strain sensor.

(1.2.3)

It is possible, with proper bonding of a wire to a structure, to accurately measure the change in length that occurs in the bonded length of the wire. This is the underlying principle of the strain gage. In a strain gage, as shown in Figure 1.4, the gage grid physically changes length when the material to which it is bonded changes length. In a strain gage, the change in resistance occurs when the conductor is stretched or compressed. The change in resistance (ΔR) is due to the change in length of the conductor, the change in cross-sectional area of the conductor, and the change in resistivity (Δρ) due to mechanical strain. If the unstrained resistivity of the material is defined as ρus and the resistivity of the strained material is ρs, then ρus − ρs = Δρ.

FIGURE 1.4 A typical uniaxial strain gage.

(1.2.4)

The resistance strain gage is convenient because the change in resistance that occurs is directly proportional to the change in length per unit length that the transducer undergoes. Two fundamental types of strain gages are available, the wire gage and the etched foil gage, as shown in Figure 1.5. Both gages have similar basic designs; however, the etched foil gage introduces some additional flexibility in the gage design process, providing additional control, such as temperature compensation. The etched foil gage can typically be produced at lower cost.

FIGURE 1.5 Resistance wire and etched resistance foil gages.

The product of gage width and length defines the active gage area, as shown in Figure 1.6. The active gage area characterizes the measurement surface and the power dissipation of the gage. The backing length and width define the required mounting space. The gage backing material is designed such that high transfer efficiency is obtained between the test material and the gage, allowing the gage to accurately indicate the component loading conditions.

FIGURE 1.6 Gage dimensional nomenclature.

1.2.1 GAGE RESISTANCE AND EXCITATION VOLTAGE

Nominal gage resistance is most commonly either 120 or 350 ohms. Higher-resistance gages are available if the application requires either a higher excitation voltage or the material to which it is attached has low heat conductivity. Increasing the gage resistance (R) allows increased excitation levels (V) with an equivalent power dissipation (P) requirement as shown in Equation 1.2.5.

Testing in high electrical noise environments necessitates the need for higher excitation voltages (V). With analog-to-digital (A–D) conversion for processing in computers, a commonly used excitation voltage is 10 volts. At 10 volts of excitation, each gage of the bridge would have a voltage drop of approximately 5 V. The power to be dissipated in a 350-ohm gage is thereby approximately 71mW and that in a 120-ohm gage is approximately 208mW:

(1.2.5)

At a 15-volt excitation with the 350-ohm gage, the power to be dissipated in each arm goes up to 161mW. High excitation voltage leads to higher signal-to-noise ratios and increases the power dissipation requirement. Excessively high excitation voltages, especially on smaller grid sizes, can lead to drift due to grid heating.

1.2.2 GAGE LENGTH

The gage averages the strain field over the length (L) of the grid. If the gage is mounted on a nonuniform stress field the average strain to which the active gage area is exposed is proportional to the resistance change. If a strain field is known to be nonuniform, proper location of the smallest gage is frequently the best option as shown in Figure 1.7.

FIGURE 1.7 Peak and indicated strain comparisons.

1.2.3 GAGE MATERIAL

Gage material from which the grid is made is usually constantan. The material used depends on the application, the material to which it is bonded, and the control required. If the gage material is perfectly matched to the mechanical characteristics of the material to which it is bonded, the gage can have pseudo temperature compensation with the gage dimensional changes offsetting the temperature-related component changes. The gage itself will be temperature compensated if the gage material selected has a thermal coefficient of resistivity of zero over the temperature range anticipated. If the gage has both mechanical and thermal compensation, the SYSTEM will not produce apparent strain as a result of ambient temperature variations in the testing environment. Selection of the proper gage material that has minimal temperature-dependant resistivity and some temperature-dependent mechanical characteristics can result in a gage SYSTEM with minimum sensitivity to temperature changes in the test environment. Strain gage manufacturers broadly group their foil gages based on their application to either aluminum or steel, which then provides acceptable temperature compensation for ambient temperature variations.

The major function of the strain gage is to produce a resistance change proportional to the mechanical strain (μ) the object to which it is mounted experiences. The gage proportionality factor, commonly called the gage factor (GF), which makes the two equations of 1.2.6 equivalent, is defined in Equation 1.2.7. Most common strain gages have a nominal gage factor of 2, although special gages are available with higher gage factors.

(1.2.6)

(1.2.7)

The gage factor results from the mechanical deformation of the gage grid and the change in resistivity of the material (ρ) due to the mechanical strain. Deformation is the change in length of the gage material and the change in cross-sectional area due to Poisson’s ratio. The change in the resistivity, called piezoresistance, occurs at a molecular level and is dependent on gage material.

In fatigue life prediction, cyclic loads may only be a fraction of the loads required to cause yielding. The measured output from the instrumentation will depend on the gage resistance change, which is proportional to the strain. If the loads are relatively low, Equation 1.2.7 indicates the highest output and the highest signal-to-noise ratio is obtained with high-resistance gages and a high gage factor.

Example 1.1.

A 350-ohm gage is to be used in measuring the strain magnitude of an automotive component under load. The strain gage has a gage factor of 2. If the component is subjected to a strain field of 200 microstrain, what is the change in resistance in the gage? If a high gage factor 120-ohm strain gage is used instead of the 350-ohm gage, what is the gage factor if the change in resistance is 0.096 ohms?

Solution.

By using Equation 1.2.7, the change in resistance that occurs with the 350-ohm gage is calculated as

By using Equation 1.2.7, the gage factor of the 120-ohms gage is

1.2.4 STRAIN GAGE ARRANGEMENTS

Strain gages may be purchased in a variety of arrangements to make application easier, measurement more precise, and the information gained more comprehensive. A common arrangement is the 90° rosette, as shown in Figure 1.8. This arrangement is popular if the direction of loading is unknown or varies. This gage arrangement provides all the information required for a Mohr’s circle strain analysis for identification of principle strains. Determination of the principle strains is straightforward when a three-element 90° rosette is used, as shown in Figure 1.9.

FIGURE 1.8 Three-element rectangular and stacked rectangular strain rosettes.

FIGURE 1.9 Rectangular three-element strain rosette.

Mohr’s circle for strain would indicate that with two gages at 90° to each other and the third bisecting the angle at 45°, the principle strains can be identified as given in Equation 1.2.8. The orientation angle (φ) of principle strain (ε1), with respect to the X-axis is as shown in Equation 1.2.9, with the shear strain (γxy) as given in Equation 1.2.10:

(1.2.8)

(1.2.9)

(1.2.10)

The principle strains are then given by Equations 1.2.11 and 1.2.12:

(1.2.11)

(1.2.12)

Correspondingly, the principle angle (φ) is as shown in Equation 1.2.13:

(1.2.13)

With principle strains and principle angles known, principle stresses can be obtained from stress–strain relationships. Linear stress–strain relationships are given in Equations 1.2.14-1.2.25. In high-strain environments, these linear equations may not hold true.

The linear stress–strain relationships in a three-dimensional state of stress are shown in Equations 1.2.14-1.2.16 for the normal stresses. The stresses and strains are related through the elastic modulus (E) and Poisson’s ratio (μ):

(1.2.14)

(1.2.15)

(1.2.16)

The relationship between shear strains and shear stresses are given in Equation 1.2.17. Shear strains and shear stresses are related through the shear modulus (G):

(1.2.17)

Equations 1.2.18-1.2.20 can be used to obtain the normal stresses given the normal strains, with a three-dimensional linear strain field:

(1.2.18)

(1.2.19)

(1.2.20)

Shear stresses are directly obtained from shear strains as shown in Equation 1.2.21:

(1.2.21)

Equations 1.2.22 and 1.2.23 can be used to obtain principle stresses from principle strains:

(1.2.22)

(1.2.23)

Principle stresses for the three-element rectangular rosette can also be obtained directly from the measured strains, as shown in Equations 1.2.24 and 1.2.25:

(1.2.24)

(1.2.25)

1.3 UNDERSTANDING WHEATSTONE BRIDGES

The change in resistance that occurs in a typical strain gage is quite small, as indicated in Example 1.1. Because resistance change is not easily measured, voltage change as a result of resistance change is always preferred. A Wheatstone bridge is used to provide the voltage output due to a resistance change at the gage. The strain gage bridge is simply a Wheatstone bridge with the added requirement that either gages of equal resistance or precision resistors be in each arm of the bridge, as shown in Figure 1.10.

FIGURE 1.10 A Wheatstone bridge circuit.

1.3.1 THE BALANCED BRIDGE

The bridge circuit can be viewed as a voltage divider circuit, as shown in Figure 1.11. As a voltage divider, each leg of the circuit is exposed to the same excitation voltage (Eex). The current that flows through each leg of the circuit is the excitation voltage divided by the sum of the resistances in the leg, as shown in Equation 1.3.1. If the resistance value of all resistors is equal (R1 = R2 = R3 = R4 = R), the current flow from the source is the excitation voltage (Eex) divided by R, as shown in Equation