Review and Critical Analysis of the Research Papers published till date on Maximum Power Point Tracking in Wind

Energy Conversion System

Syed Muhammad Raza Kazmi Hai-Jiao Guo*

Member

Hiroki Goto
Member

Osamu Ichinokura
Member

*Department of Electrical & Information Engineering, Tohoku Gakuin University, Japan Department of Electrical & Communication Engineering, Tohoku University, Japan syed.raza.kazmi@gmail.com
Abstract -- Maximum power point tracking (MPPT) is a very important necessity in a system of energy conversion from a renewable energy source. Every year a number of publications appear in various journals and conferences claiming to offer better and faster MPPT techniques for wind energy conversion system (WECS). This research paper provides a concise yet comprehensive critical analysis of these techniques with an indepth review of their strengths and drawbacks. This review is unique as there has been no other research paper so far that offers such a complete and up-to-date investigation of MPPT techniques in WECS. Therefore this research paper can serve as a precise reference for the future research on MPPT for WECS. Index Terms -- review; MPPT; wind energy conversion system.

II.

THE LOOKUP TABLE BASED

I. INTRODUCTION The maximum extractable power from a renewable energy source depends not only on the strength of the source but also on the operating point of the energy conversion system. Therefore the Maximum Power Point Tracking (MPPT) is of the paramount importance in renewable energy conversion systems for not only to maximize the systems efficiency but also to minimize the return period of the installation cost. In a wind energy conversion system (WECS) the concept of MPPT is to optimize the generator speed relative to the wind velocity intercepted by the wind turbine such that the power is maximized. Every year numerous research efforts are attempted to achieve better and faster techniques on MPPT in WECS. Till date the count of such research papers appearing in IEEEXplore database is around an impressive total of 100. However the study in this review paper reveals that these techniques actually fall in either of the five categories as explained in the subsequent sections. This review paper provides a critical insight on the strengths and drawbacks of those five categories and hence it can serve as an authentic reference for the further research in this field.

Perhaps the most widely used MPPT techniques are lookup table based. They require either a pre-programmed 2D lookup table with stored values of optimal generator speed and the corresponding maximum power (maximum torque) at various wind velocities; or a cubic (quadratic) mapping function to provide reference signal for optimal turbine power (torque) at the operating generator speed, as shown in Fig. 1. Hence they require a generator speed sensor or observer. Classically the most popular approach is the power signal feedback (PSF) [1], [2] which uses either a 2D lookup table with the maximum power in the ordinate or a mapping function employing the product of the cube of measured generator speed with the optimum proportionality constant.

Fig. 1.

A slight variant is the optimal torque (OT) technique published in [3] which makes use of the quadratic optimal torque curve. Nonetheless there is no difference between PSF and OT in performance and complexity of implementation. Another commercially employed lookup table MPPT is the TSR control which additionally requires an anemometer for the wind speed measurement and also pre-known value of the optimal tip speed ratio to convert the wind velocity measurement into its corresponding reference for optimal generator speed [1]-[2], as shown in Fig. 1(b).

(a) (b) The two widely used lookup table based MPPT techniques: (a) PSF and (b) TSR control.

978-1-4244-5287-3/10/$26.00 2010 IEEE

4075

Critical Analysis & Drawbacks Owing to the non-negligible inertia of WECS, the generator speed changes a bit lethargically as compared to the change in wind. This means that the reference for PSF and OT techniques cannot be set instantaneously with the change in wind because the sudden change in wind may not cause a sudden change in generator speed. In contrast, the TSR control technique can provide the fastest control action as it directly measures the wind speed and sets the control reference instantaneously; hence it is expected to yield more energy. However the accurate wind measurement is not a trivial task specially in case of large size wind turbines. As explained in [4], anemometer provides limited measurements of wind speed only at the hub height and cannot cover for the whole span of large blades. Moreover, due to the interaction between the rotor and the wind, this usual placement of anemometers on nacelles leads to inaccurate wind speed measurements in both upwind and downwind turbines. In other words the wind speed measured by the anemometer may not be the one intercepted by the wind turbine. Plus the additional requirement of a wind speed sensor makes TSR control more costly than PSF. Almost all the research papers employing these lookup table based MPPT regard the stored optimal values to be constants. But as the matter of fact, it is not true. There are a number of factors that can cause drift in these values and therefore the stored curve or mapping function may not remain valid for the optimal power extraction. A brief account of these factors causing the drift in the optimal characteristics is given as follows: A. Change in the air density The power extracted from wind energy is a function of the air density which may vary considerably over various seasons. As a result, the stored optimal curve on the basis of some nominal air-density value may not result in optimal tracking of the peak power point in all seasons [5], [6]. B. The aging factor Mechanical, electrical and thermal stresses in the system over a prolonged use can cause the drift in systems parameters. Also the debris buildup or the blade erosion changes the power coefficient Cp curve [7]. C. Non constant efficiencies of generator and converter subsystems Even if the above two factors may be considered negligible still the non constant efficiency of the system is the major source for the non-uniqueness of optimal characteristics. A unique cubic optimal curve is actually the attribute of the mechanical power Pm of the wind turbine. Whereas the actual interest is the output electrical power Po supplied to the load. The transfer function from Pm to Po is dictated by the generator efficiency g and converter efficiency c.

Po = g c Pm

(1)

The generator efficiency generally gets worse on the increase in phase current due to the copper losses; whereas the increase in rotational speed boosts the eddy current losses in the generator core. Ref. [8] has a graph that shows such varying efficiency of a switched reluctance generator with respect to its torque and speed. Similarly on the converter side the copper losses in the inductor and the losses in the power electronics devices may vary the converter efficiency with the change in the load current. With these non-constant efficiencies of the system, we can infer two very important conclusions, as experimentally shown in [9]: 1) there does not exist a unique optimal curve for Po. 2) the maximum of Po curve with respect to the generator speed does not coincide with the peak of Pm curve. III. THE STATE SPACE LINEARIZATION AND NONLINEAR STATE SPACE BASED Being a nonlinear control problem, linear and nonlinear state space control theory have found their applications in the control of WECS. Ref. [10] combines feedback linearization with optimal control theory whereas [11] employs feedback linearization theory with sliding mode control. Ref. [12] uses TSR control along with input-output feedback linearization to eliminate the effects of nonlinearity caused by magnetic saturation. Ref. [13] formulates a disturbance observer for estimation of parameter uncertainties via feedback linearization. Several publications can be found on sliding mode or passivity based variable structure control (VSC) of WECS [14]-[18].

Critical Analysis & Drawbacks For a specific well defined wind turbine and generator, these state space techniques may provide robustness against disturbances. But the formulation of these techniques require system modeling which may not be available or known in general; specially in the case of WECS driven by the new kinds of special machine whose mathematical models have not been established yet. Therefore these techniques are difficult to implement and their system specific nature makes them sensitive to the drift or modifications in the system parameters.
IV. THE NEURAL NETWORK-FUZZY LOGIC BASED Researchers have extended the realms of artificial neural network (ANN), fuzzy logic and neuro-fuzzy control towards the maximum power extraction control of wind energy systems [19]-[25].

4076

Critical Analysis & Drawbacks Fuzzy logic MPPT controller [21]-[23] is basically the extended version of the hill climb searching (HCS) or perturb & observe (P&O) control which will be explained in the next section. In HCS the control decision is taken based on only one IF-ELSE statement whereas the so called intelligent fuzzy logic control is governed by a set of rules which chooses different control action according to the state of the system at that instant. Like HCS, fuzzy control does not require to know system parameters or equations. But it may require speed sensors. The challenge however is that how to define an optimal set of rules and corresponding control actions. Fuzzy control requires to define lots of boundaries and gains for which these publications do not provide any general purpose guidelines for selection and optimization. Although, as just mentioned above, fuzzy controller does not require to know about the system but the efficient controller design cannot be done without some study of the system. Hence it is not generic. ANN based control [19], [20] can be quite effective and robust only after it is sufficiently trained for all kinds of operating conditions. This is quite a tough requirement and requires a long offline training. This long offline training makes ANN quite unattractive for the real time practical applications. The ANN for its training requires wind velocity sensor additionally with the generator speed sensor which is again not a good feature.
V. THE HILL CLIMBING BASED The hill shaped power curves of WECS exhibit unique maxima with respect to its control variable (generator speed or converter duty ratio). Hence a simple discrete time hill climb searching (HSC) control can be employed by perturbing the control variable and observing the resulting increase or decrease in power; hence also known as perturb & observe (P&O). If it results in the increase in power then the same perturbation is applied for the next control instance; otherwise the sign of the perturbation is reversed in order to track in the direction of increasing power [1], [2], [5]; as visualized in Fig. 2. Mathematically the HCS control law can be expressed as:

Critical Analysis & Drawbacks HCS or P&O is the simplest MPPT algorithm that does not require any prior knowledge of the system or any additional sensor except the measurement of the power which is subjected to maximization. Therefore HCS can be applied to any renewable energy conversion system that exhibits a unique power maximum. Although these features should make HCS the top choice for MPPT in any renewable energy conversion system but in reality it is only feasible in the slow varying systems. For instance it is quite feasible for the PV energy systems where the suns irradiance changes over the period of several minutes but not for the WECS where the wind may change quite fast in the matter of seconds. There are two serious problems with HCS which significantly deteriorate its performance under rapidly changing wind conditions. These two problems are [9]: Perturbation Step Size & SpeedEfficiency Tradeoff As shown in Fig. 3, a larger perturbation step size increases the speed of convergence but deteriorates the efficiency mppt of the MPPT by amplifying the oscillations Pmpp around the maximum power point (MPP) Pmpp. It is due to the fact that HCS control does not halt at Pmpp and hence the oscillations are unavoidable attribute of the HCS control. A smaller step size boosts the efficiency but the controller then becomes slower and may become incapable of tracking MPP under rapidly varying wind conditions. Therefore in the conventional HCS control there always exists a tradeoff between the tracking speed and the control efficiency. A.

(a) (b) Fig. 3. The perturbation is larger in (a) than in (b). Hence the tracking speed in (a) is faster than in (b) but with larger oscillations about MPP.

Fig. 2.

Principle of HCS MPPT. Keep perturbing the control variable in the same direction until the power is decreased.

Perturbation Step Size & SpeedEfficiency Tradeoff In the normal HCS, the direction i.e. the sign of the next perturbation is decided by the increase or decrease in power due to the previous perturbation. Being blind to the wind change, this rule can be misleading as the sign might get dictated by the change in wind rather than the applied perturbation. This wrong decision leads to the failure in keeping track of MPP and the HCS control moves downhill as shown in Fig. 4. Of these two problems, the latter one (B.) is of the major concern as it significantly deteriorates the energy extraction and hence greatly effects the overall input-output efficiency of the system. Therefore mostly the WECS designers prefer the lookup table based MPPT due to its rapid tracking despite the cost of the additional mechanical sensor.

B.

4077

Fig. 4. The HCS control losing its trackability under changing wind conditions and traveling downhill instead of the uphill climb.

Fig. 5.

(a) (b) (a) The principle of variable step size MPPT. (b) The problem with variable step size MPPT under changing wind.

VI. THE MODIFIED OR HYBRID HCS/LOOKUP TABLE TECHNIQUES As evident from the comparison of different approaches discussed in the previous sections, the lookup table based MPPT and HCS are the most feasible of MPPT techniques which can be more promising if their mentioned drawbacks can be removed through some smart upgradation. Therefore the majority of the research papers on MPPT are written on the modified or hybrid versions of HCS and PSF. Here a critical analysis of those research paper is presented.

In [28] the calculation of requires the knowledge of turbines mechanical power which practically is not available in most of the systems. Even if the mechanical power is assumed equal to the electrical power measured at the output, still the methodology of [28] is most likely to fail in fluctuating wind condition because the signs of dP/d and d/d cannot indicate the true operating sector (defined in [28]) under changing wind.

B.

Variable Step Size HCS [26]-[28] As explained above, a fixed step size HCS always suffers from tracking speed versus control efficiency tradeoff. Ref. [26], [27] replace the constant step size by the scaled measure of the slope of power with respect to the perturbed generator speed P/ (can also be with respect to the converters duty ratio). This idea is actually not new as quite a similar one was published in [29] long time back for the HCS in PV systems. Ref. [28] defines four operating regions and their corresponding control actions according to the slopes of power P and generator speed with respect to a variable . Critical Analysis & Drawbacks The variable step is expected to adapt a larger step size when the operating point is away from the peak due to the larger magnitude of P slope and as the peak gets nearer, the step size should automatically approach to zero. Apparently the idea seems quite workable as can be seen from the slopes of a typical power curve in Fig. 5(a). But in reality it is only workable for a constant or very slowly changing wind. Otherwise P/ will not give a correct measure of the distance from the maximum as the operating point shifts from one power curve to another for different wind velocities. Therefore the algorithm is most likely to result in unnecessary large or small step size during changing wind, resulting into big deviation from the maximum. This is explained in Fig. 5(b) where the transition from point 1 to 2 due to the increase in wind speed results in a large P/ whereas the transition from point 3 to 4 due to the decrease in wind gives a small P/ . In both the cases the resulting magnitude of P/ is absolutely inappropriate rather misleading for the next perturbation decision and therefore the so called variable step size algorithm fails to track the maxima under changing wind.

A.

Dual Step Size HCS [30] An alternative for the variable step size HCS is proposed in [30] where only two discrete values are used in a way that until the maximum is achieved or if the wind change is detected then the algorithm uses a larger step size, otherwise it is switched to a smaller one.

Critical Analysis & Drawbacks As explained above in the critical analysis of the variable step HCS, if there is no or very slow change in wind then the variable step HCS can prove to be efficient. But realistically it is very rare to happen and therefore the dual step size HCS is a bit better alternative as it does not suffer much from the unnecessary fluctuations in the command signal with the variations in wind. However the implementation of the algorithm requires to sense the change in wind. One of the novelty of this algorithm is that for the true peak detection it does not merely use P<0 but also the checks on the current and the previous samples of the change in wind speed. Search-Remember-Reuse HCS [31] Ref. [31 turns HCS into a lookup table based technique by online training of a memory table. It starts off with the conventional HCS which [31] rephrases as maximum power error driven (MPED) control. Then a memory is maintained in the form of {Maximum Power, Rectified Generator Voltage Vdc, Current Control Value} triplet. This memory is trained every time by comparing the current output power Po with the recorded maximum power at the nearest Vdc in the memory table and updated if the present power peak is larger than the previously recorded one. The updated memory table then serves as a direct reference for the current command. Another important feature of [31] is that it carries out MPPT for the turbine mechanical power Pm instead of the output electrical power Po claiming that this can cope with the large inertia of WECS. C.

4078

Critical Analysis & Drawbacks There are a few critical problems in the algorithm of [31] explained as follows: 1. The rule laid out for the memory buildup simply compares the output power at the current Vdc with the one stored in the memory at the nearest Vdc. Merely this rule cannot construct a lookup table of maximum power points (MPP). Instead this rule will result in the recording of the Po Vdc at highest occurring wind velocity. Hence the algorithm does not have a proper peak detection mechanism. 2. The paper claims to cope with the fast wind variations and the large inertia of wind turbines by performing the maximization of Pm instead of Po. For this purpose an expression for Pm is derived in terms of Po and Vdc. The sign of Pm is detected by the sign of Po and the sign of (Vdc dVdc/dt). A conflict appears when both signs are opposite to each other and therefore a decision about the sign of Pm cannot be made. At that instance the algorithm proposes to seize the perturbation which is not a good idea as it can adversely effect the trackability of the algorithm under rapidly changing wind. 3. Like almost all other related publications, [31] also assumes a constant optimal characteristic of the system which can become inaccurate as already explained above in the critical analysis of the lookup table based MPPT. D. Modified HCS to avoid Generator Stall [32] Ref. [32] adds a faster control index into the HCS driven control value. The fast index is proposed as an exponential function of the differential generator speed and therefore it causes sharp increase or decrease of generator current command on sudden increase or decrease of wind speed respectively. This augmented fast control is termed in [32] as maximum power differential speed (MPDS).

Similarly during the increase in wind speed the proposed MPDS increases the turbine torque by reducing the generator speed. This will be detrimental for MPPT if the system is now operating in left half plane. Also it may cause unnecessary overshoots due to the exponential function employed in [32].

E.

Modified PSF to avoid Generator Stall [33] Ref. [31] proposes to offset the measured power by a constant value to avoid stall. Moreover it feedbacks the derivative of the stator frequency to increase the sensitivity of the system against the change in wind.

Critical Analysis & Drawbacks The added constant value to avoid stall will make the algorithm unable to achieve MPP as there always be an offset in the TSR with respect to its optimal value. It is argued that the derivative of stator frequency will become zero at MPP. In reality this is only true when wind velocity is constant or changing very slowly, otherwise the derivative gain will never become zero. Also it may cause undesirable overshoots in the control variable. F. Modified OT for Fast Tracking [34]-[37] An interesting modification of the optimal torque control law is proposed in [34]-[36] which is also employed in [37]. Instead of the conventional OT command T = k2, a net torque term responsible for the acceleration or deceleration of generator speed is also incorporated as given below:
T = k 2 G Tm k 2

where Tm is the mechanical torque captured by the wind turbine and G is a positive definite gain to adjust the contribution of the net torque Tmk2 during the transience.

(2)

Critical Analysis & Drawbacks Conventional HCS hangs in the case of continuously decreasing wind because the control law becomes (k + 1) = (k ) . Therefore the controller becomes dysfunctional and this may cause WECS to stall because of the dropping shaft speed with the decreasing wind velocity. The MPDS in [32] sharply increases the generator speed to avoid stall. However this control will lead farther away from MPP if the system is operating in the right hand region with respect to the cubic optimal power curve shown in Fig. 6.

Fig. 6.

The left and right half planes on P curve bisected by the optimal power curve.

Critical Analysis & Drawbacks The motivation behind this modification is to cope with the large inertia of turbine and generator which prohibits the system to instantly follow the torque command under changing wind. For a WECS of large inertia, even an accurately recorded OT curve cannot drive the system to maxima under rapidly changing wind because the generator speed needs longer time to adjust to the new command. This algorithm aids the tracking speed by producing a larger difference between the new available torque Tm (due to change in wind) and the torque command it decreases the torque command with the increase in aerodynamic torque and vice versa. Hence, as can be seen by the Newtonian equation the modified OT control of a one-mass system: Tm T = J generates a larger acceleration/deceleration enabling the wind turbine to follow the command signal rather rapidly. Note that in the steady state, i.e. when the wind is almost constant and the generator acceleration is almost zero, the modified OT law reduces back to the conventional one, implying that in steady state the algorithm eventually drives the system to the actual maxima.

4079

The drawback of this modification however is that now it is required to have a measure of turbine mechanical torque which is either calculated via Newtonian equation as in [35], [36] or estimated via torque observer as in [37]. In both cases we require to know the system parameters such as the turbine inertia J, which may change over time due to aging.

Adaptive TSR Control [38], [39] The algorithm proposed in [38], [39] initiates the TSR control with an approximated optimal TSR value. Later when the measured wind velocity is found to be stable enough, then the algorithm switches to HCS to search for the real optimal point. When true peak is reached, then: 1. a memory table of the optimum generator speed with the corresponding wind velocity is updated and 2. the TSR is corrected. So in the changing wind, the rotor speed reference is either directly applied if a recorded data at current wind speed is present in the memory or it is calculated through the stored TSR. Critical Analysis & Drawbacks In terms of fast and accurate tracking without the need of any pre-programmed system characteristics, this algorithm is so far the best one. It upgrades the TSR control, which is the fastest of all the lookup table based methods as explained earlier, such that now the optimal TSR value is not regarded as a constant but tuned online as the algorithm proceeds in time. This self-tuning also provides the algorithm immunity towards the non-unique optimal curves of output electrical power. The reasons for the occurrence of more than a single unique optimal curve have been already explained above in the critical analysis of lookup table based techniques. This algorithm however suffers from the same issues as of TSR control regarding the wind speed measurement. H. Adaptive OT Control [40] Ref. [40] embeds HCS in the optimal torque (OT) control by perturbing the proportionality gain of the optimal torque equation and observing the change in the estimated measure of power coefficient Cp to decide for the next perturbation. The proposed estimate of Cp is computed by taking the ratio of the average mechanical power over the average wind power. This averaging is performed over a very long time period to average out the wind variation effects.

G.

i.e. n of [40] is taken as 1, then the algorithm will behave like a conventional HCS which is most likely to go astray under fluctuating wind. The long slow adaption of the optimal torque gain has been targeted to cope only with the very slow changes in the system parameters occurring for instance due to aging or due to the seasonal changes in air density. Therefore the slow adaption is justified. However the algorithm needs to know the inertia for its computation which may change due to aging and therefore the adapted gain may still be suboptimal. Also it additionally requires an anemometer.

I.

Limit Cycle Based HCS [41] An interesting alternative of conventional HCS is proposed in [41]. The MPPT is performed via an integrator ramping up or down the current command signal of the load side inverter according to the error in the dc link voltage regulated by the generator side chopper. The current command ramps up till the maximum power is achieved and if it is increased further then the dc link voltage cannot be kept at constant because the power equilibrium cannot be maintained. Therefore the dc link voltage begins to decrease and if it goes down a certain limit, then the integrator gain is shifted to ve value ramping down the current command. In this way the MPPT exhibits non linear oscillations about MPP called limit cycle.

Critical Analysis & Drawbacks The merit of the proposed method over the conventional HCS is its hardware simplicity as it does not require a power calculator and the whole algorithm can be implemented via inexpensive analog elements. However it does suffer from all those failures of conventional HCS. On the top of this, here the control designer needs to adjust various control parameters like integrator time constants and PI control gains etc. The improper tuning of these parameters can effect the MPPT performance. Disturbance Injection Based HCS [42], [43] As the name implies, this algorithm proposes to inject a sinusoidal perturbation instead of the constant step. The output power is then sampled at /2 and 3/2 of each cycle, the difference of which decides about the next perturbation. Critical Analysis & Drawbacks This algorithm does not provide any added advantage over conventional HCS. Instead it adds up the following drawbacks: 1. The conventional HCS requires just a single previous sample but this algorithm requires to measure samples at two time instances to decide about the next perturbation. Therefore this algorithm is more slower than HCS which is totally undesirable for the tracking of a large inertia WECS under rapidly varying wind. J.

Critical Analysis & Drawbacks Normally it is not a good idea to drive the OT control via HCS because this will induce in OT control all those shortcomings which are in the nature of HCS and therefore the MPPT will fail under fast changing wind. But actually here in [40], the perturbation of the OT gain is not performed continuously but periodically after a very long time period. The decision for the next perturbation depends on the change in the Cp estimate. It is in fact quite essential to perform averaging over long time periods before invoking HCS so that the changing wind should not mislead the HCS in the wrong direction. For instance if no averaging is performed,

4080

2.

The algorithm will only work for constant or very slow changing wind, otherwise the samples at /2 and 3/2 will mislead the tracking.

K.

Self Tuning Sensorless Adaptive HCS [44] Authors have proposed a novel self tuning adaptive step size HCS algorithm in [44]. It starts off with an upgraded HCS equipped with a robust peak detection capability which is not present in any of the research papers before. Moreover it uses a very smart and simple scheme for the estimation of generator speed by considering the dependence of the frequency of the phase current on generator speed and rotor poles. Once a peak is detected then it extracts the optimal curve constant kopt through measured output electrical power and estimated generator speed. When the wind changes, the system switches to the proposed adaptive HCS whose principle is shown in Fig. 7. Adaptive HCS decides its next perturbation size as well as direction by measuring the distance vector of current
operating point from the optimal curve defined by k opt 3 .

VII. CONCLUSIONS A thorough review and critical analysis of the techniques published for MPPT in WECS has been presented in this research paper. This review paper spans all the publications till date. The critical analysis concludes the two best techniques to be [38] and [44]. Both algorithms have adaptive tracking with self-tuning capability. However [38] requires both wind speed and generator speed sensors. Whereas [44] has a clear edge over [38] and all the rest for being absolutely mechanical sensorless. Also, in contrast to [38], the algorithm of [44] does not build a lookup table and therefore is easy on the memory requirements. REFERENCES
[1] I. K. Buehring and L. L. Freris, Control policies for wind energy conversion systems, IEE Proceedings, Part C - Generation, Transmission and Distribution, vol. 128, pp. 253-261, Sept. 1981. Eftichios Koutroulis and Kostas Kalaitzakis, Design of a maximum power tracking system for wind-energy-conversion applications, IEEE Transactions on Industrial Electronics, vol. 53, no. 2, April 2006. Morimoto S., Nakayama H., Sanada M., Takeda Y., Sensorless output maximization control for variable-speed wind generation system using IPMSG, IEEE Transactions on Industry Applications, vol. 41, no. 1, 2005. Kathryn E. Johnson and Lucy Y. Pao, A tutorial on the dynamics and control of wind turbines and wind farms, in Proc. American Control Conference, 2009. R. Datta, V. T. Ranganathan, A method of tracking the peak power points for a variable speed wind energy conversion system, IEEE Transactions on Energy Conversion, vol. 18, no. 1, 2003. Joanne Hui and Alireza Bakhshai, Adaptive algorithm for fast maximum power point tracking in wind energy systems, in Proc. IECON 2008. Johnson, K., Fingersh, L., Balas, M., and Pao, L. Methods for increasing region 2 power capture on a variable speed wind turbine, Journal of Solar Energy Engineering, vol. 126, Nov. 2004. David A. Torrey, Switched reluctance generators and their control, IEEE Transactions on Industrial Electronics, vol. 49, no. 1, Feb. 2002. Kazmi Syed Muhammad Raza, Hiroki Goto, Hai-Jiao Guo, and Osamu Ichinokura, A novel algorithm for fast and efficient maximum power point tracking of wind energy conversion systems, in Proc. ICEM, Sept. 2008. K. Uhlen, B. A. Foss, and O. B. Gjosaeter, Robust control and analysis of a wind-diesel hybrid power plant, IEEE Transactions on Energy Conversion, vol. 9, no. 4, Dec. 1994. Jos Matas, Miguel Castilla, Josep M. Guerrero, Luis Garca de Vicua, and Jaume Miret, Feedback linearization of direct-drive synchronous wind-turbines via a sliding mode approach, IEEE Transactions on Power Electronics, vol. 23, no. 3, May 2008. Wei Qiao, Liyan Qu, and Ronald G. Harley, Control of IPM synchronous generator for maximum wind power generation considering magnetic saturation, IEEE Transactions on Industry Applications, Vol. 45, Issue 3, May/June 2009. Juan Manuel Mauricio, Andres E. Leon, Antonio Gomez-Exposito, and Jorge A. Solsona, An adaptive nonlinear controller for DFIMbased wind energy conversion systems, IEEE Transactions on Energy Conversion, vol. 23, no. 4, Dec. 2008. Hernn De Battista, Ricardo J. Mantz, and Carlos F. Christiansen Dynamical sliding mode power control of wind driven induction generators, IEEE Transactions on Energy Conversion, vol. 15, no. 4, Dec. 2000.

[2]

[3]

The distance vector ( *) , as shown in Fig. 7, not only provides the right estimate of the next step size required to drive this distance to zero but also the correct sense of perturbation direction despite the wind change. Since the kopt of electrical power is not valid for a different wind, as discussed in Section II, this technique performs self tuning by returning back to normal HCS once the transience is over. The system seizes the perturbation at maxima to obtain 100% control efficiency.

[4]

[5]

[6]

[7]

[8] [9]

Fig. 7.

Principle of the Adaptive HCS of [44].

[10]

Critical Analysis & Drawbacks From the review presented in this paper it can be witnessed that the algorithm of [44] is the best on of all. It possesses all the virtues of HCS, i.e. mechanical sensorless and absolutely generic, but none of the vices of the reviewed techniques. It can effectively cope with the rapidly changing wind as well as the inconsistent efficiency of the system. However there are two limitations of this technique: 1. Large inertia of WECS and fast rate of change of wind may cause longer time to find a kopt. 2. The limit employed to detect the wind change requires a test run for its tuning.

[11]

[12]

[13]

[14]

4081

[15] Fernando Valenciaga, Pablo F. Puleston, and Pedro E. Battaiotto Power control of a solar/wind generation system without wind measurement: A passivity/sliding mode approach, IEEE Transactions on Energy Conversion, vol. 18, no. 4, Dec. 2003. [16] Fernando Valenciaga and Paul F. Puleston, Variable structure control of a wind energy conversion system based on a brushless doubly fed reluctance generator, IEEE Transactions on Energy Conversion, vol. 22, no. 2, June 2007. [17] Brice Beltran, Tarek Ahmed-Ali, and Mohamed El Hachemi Benbouzid, Sliding mode power control of variable-speed wind energy conversion systems, IEEE Transactions on Energy Conversion, vol. 23, no. 2, June 2008. [18] Iulian Munteanu, Seddik Bacha, Antoneta Iuliana Bratcu, Joel Guiraud, and Daniel Roye, Energy-reliability optimization of wind energy conversion systems by sliding mode control, IEEE Transactions on Energy Conversion, vol. 23, no. 3, Sept. 2008. [19] Hui Li, K. L. Shi, and P. G.McLaren, Neural-network-based sensorless maximum wind energy capture with compensated power coefficient, IEEE Transactions on Industry Applications, vol. 41, no. 6, Nov./Dec. 2005. [20] J. S. Thongam, P. Bouchard, H. Ezzaidi, and M. Ouhrouche, Artificial nural network-based maximum power point tracking control for variable speed wind energy conversion systems, in Proc. 18th IEEE International Conference on Control Applications, July 2009. [21] Rohin M. Hilloowala, and Adel M. Sharaf, A rule-based fuzzy logic controller for a PWM inverter in a stand alone wind energy conversion scheme, IEEE Transactions on Industry Applications, vol. 32, no. 1, Jan./Feb. 1996. [22] Marcelo Godoy Simoes, Bimal K. Bose, and Ronald J. Spiegel, Design and performance evaluation of a fuzzy-logic-based variablespeed wind generation system, IEEE Transactions on Industry Applications, vol. 33, no. 4, July/August 1997. [23] Ghanim Putrus, Mahinsasa Narayana, Milutin Jovanovic, and Pak Sing Leung, Maximum power point tracking for variable-speed fixed-pitch small wind turbines, in Proc. 20th International Conference on Electricity Distribution, June 2009. [24] Vincenzo Galdi, Antonio Piccolo, and Pierluigi Siano, Designing an adaptive fuzzy controller for maximum wind energy extraction, IEEE Transactions on Energy Conversion, vol. 23, no. 2, June 2008. [25] Riad B. Chedid, Sami H. Karaki, and Chadi El-Chamali, Adaptive fuzzy control for wind-diesel weak power systems, IEEE Transactions on Energy Conversion, vol. 15, no. 1, March 2000. [26] Jia Yaoqin, Yang Zhongqing, Cao Binggang, A new maximum power point tracking control scheme for wind generation, in Proc. PowerCon 2002. [27] B. Neammanee, S. Sirisumranukul, S. Chatratana, Control performance analysis of feedforward and maximum peak power tracking for small-and medium-sized fixed pitch wind turbines, in Proc. ICARCV 2006. [28] Vivek Agarwal, Rakesh K. Aggarwal, Pravin Patidar, and Chetan Patki, A Novel Scheme for Rapid Tracking of Maximum Power Point in Wind Energy Generation Systems, IEEE Transactions on Energy Conversion, (In Press).

[29] P. Huynh and B. H. Cho, Design and analysis of a microprocessorcontrolled peak-power-tracking system, IEEE Transactions on Aerospace and Electronic Systems, vol. 32, no. 1, Jan. 1996. [30] Kazmi Syed Muhammad Raza, Hiroki Goto, Hai-Jiao Guo, and Osamu Ichinokura, Maximum power point tracking control and voltage regulation of a DC grid-tied wind energy conversion system based on a novel permanent magnet reluctance generator, in Proc. ICEMS, 2007. [31] Quincy Wang and Liuchen Chang, An intelligent maximum power extraction algorithm for inverter-based variable speed wind turbine systems, IEEE Transactions on Power Electronics, vol. 19, no. 5, Sept. 2004. [32] R. J. Wai, C.Y. Lin and Y.R. Chang, Novel maximum-powerextraction algorithm for PMSG wind generation system, IET Electric Power Applications, vol. 1, no. 2, March 2007. [33] Kelvin Tan and Syed Islam, Optimum control strategies in energy conversion of PMSG wind turbine system without mechanical sensors, IEEE Transactions on Energy Conversion, vol. 19, no. 2, June 2004. [34] Tony Burton, David Sharpe, Nick Jenkins, Ervin Bossanyi, Wind Energy Handbook, Jhon Wiley & Sons, 2001. [35] Lee Jay Fingersh and Palmer W. Carlin, Results from the NREL variable-speed test bed, in Proc. 17th ASME Wind Energy Symposium, 1998. [36] L. J. Fingersh and K.E. Johnson, Baseline results and future plans for the NREL controls advanced research turbine, in Proc. 23rd ASME Wind Energy Symposium, 2004. [37] Ching-Tsai Pan and Yu-Ling Juan, A novel sensorless MPPT controller for a high-efficiency microscale wind power generation system, IEEE Transactions on Energy Conversion, (In Press). [38] Joanne Hui and Alireza Bakhshai, A new adaptive control algorithm for maximum power point tracking for wind energy conversion systems, in Proc. PESC, 2008. [39] ___, Adaptive algorithm for fast maximum power point tracking in wind energy systems, in Proc. IECON, 2008. [40] Kathryn E. Johnson, Lucy Y. Pao, Mark J. Balas, and Lee J. Fingerish, Control of variable-speed wind turbines: standard and adaptive techniques for maximizing energy capture, IEEE Control Systems Magazine, vol. 26, no. 3, June 2006. [41] Mikihiko Matsui, Dehong Xu, Longyun Kang, and Zongqing Yang, Limit cycle based simple MPPT control scheme for a small sized wind turbine generator system-principle and experimental verification, in Proc. IPEMC, 2004. [42] Y. Higuchi, N. Yamamura, and M Ishida, An improvement of performance for small-scaled wind power generating system with permanent magnet type synchronous generator, in Proc. IECON, 2000. [43] Shengtie Wang, Zhiyuan Qi, and Tore Undeland, State space averaging modeling and analysis of disturbance injection method of MPPT for small wind turbine generating systems, in Proc. APPEEC, 2009. [44] Syed Muhammad Raza Kazmi, Hiroki Goto, Hai-Jiao Guo, and Osamu Ichinokura, A novel algorithm for fast and efficient speed-sensorless maximum power point tracking in wind energy conversion systems, IEEE Transactions on Industrial Electronics, (In Press).

4082