Comparison: Controlling Wind Turbine

Transcription

1 e-issn Volume 3 Issue 3, March 2017 pp Scientific Journal Impact Factor : Comparison: Controlling Wind Turbine Mr. R.Sivakumar, M.E.,(AP/EEE) 1, A.Nelsonduraisamy 2, K.Manikandan 3, S.Manivannan 4, N.Udhayakumar 5 1,2,3,4,5 Department of Electrical and Electronics Engineering, V.S.B Engineering College, Karur Abstract: The wind turbine is playing a cardinal role in the entire system as it is responsible for the generation of mechanical power needed to drive the generator. The primary factors on which the wind turbine performance depend are, wind speed, direction of wind, blade size, pitch angle and mechanical gears involved in its design. We have discussed those methods and have selected one of the best methods for controlling wind turbine. Keywords: Wind turbine, Hub, Controllers I. INTRODUCTION Wind energy is currently the fastest-growing source of electricity in the world. Wind power investment worldwide is expected to expand three-fold in the next decade, from about $18 billion in 2006 to $60 billion in 2016 [1]. In the U.S., where wind currently only provides about 1% of the nation's electricity, wind has the potential to provide up to 20% of the nation's electricity without major changes to the nation's electricity distribution system. Most wind turbine manufacturers introduce a wind turbine in terms of annual energy output of the turbine. In many cases, this is not sufficient to decide whether or not select that particular wind turbine for a site, just by having an estimate of its annual energy output. This implies an evaluation criterion for performance(s) of any wind turbine. A new performance index based on the energy pattern factor of the wind is introduced in this paper for controller operation evaluation. This index represents the turbine performance in terms of (i) the energy pattern factor of the wind,(ii) Weibull distribution function of the site And (iii) the controller strategy. Based on this new index, a new method has been developed in this paper for estimating energy output of a wind turbine and its controller performance for any potential site. An interactive computer program is developed for this new proposed method. In this section we characterize a linear zed model of the turbine structural dynamics and view the generator as simply a static (unity) gain that translates commanded torque instantaneously into mechanical torque. There are many studies that focus on generator modeling and control and associated issues [6,1], but that is beyond the scope of this article where our focus is on pitch control and its effect on load mitigation and speed regulation. A wind turbine is inherently nonlinear and time. In general, "Lift" type blades (which are rotated by the aerodynamic lift forces as the wind passes the blades) are much more efficient than "Drag" type blades. However maximum efficiency with lift type blades can only be achieved using a controller which can change the rotational speed of the blades based on the variations in the wind speed [5]. Such variable speed wind turbine control enables the energy extraction from the wind to be maximized by maintaining the tip 2 speed ratio of the turbine near optimum [5, 9]. We demonstrate the potential improvement that may be realized when a feedback system is augmented with disturbance feed forward of wind All rights Reserved 114

2 II. OPERATION A wind turbine is a revolving machine that converts the kinetic energy from the wind into mechanical energy. This mechanical energy is then converted into electricity that is sent to a power grid. The turbine components responsible for these energy conversions are the rotor and the generator. The rotor is the area of the turbine that consists of both the turbine hub and blades. As wind strikes the turbine s blades, the hub rotates due to aerodynamic forces. This rotation is then sent through the transmission system to decrease the revolutions per minute. The transmission system consists of the main bearing, high-speed shaft, gearbox, and low-speed shaft. The ratio of the gearbox determines the rotation division and the rotation speed that the generator sees. For example, if the ratio of the gearbox is N to 1, then the generator sees the rotor speed divided by N. This rotation is finally sent to the generator for mechanical-to-electrical conversion. The major components of a wind turbine: gearbox, generator, hub, rotor, low-speed shaft, high-speed shaft, and the main bearing. The purpose of the hub is to connect the blades servos that adjust the blade direction to the low-speed shaft. The rotor is the area of the turbine that consists of both the hub and blades. The components are all housed together in a structure called the nacelle. III. POWER AND EFFICIENCY This section explains what affects the power extracted from the wind and the efficiency of this process. Consider Figure 3 as a model of the turbine s interaction with the wind. This diagram indicates that wind exists on either side of the turbine, and the proper balance between rotational speed and the velocity of wind are critical to regulate performance. The balance between rotational speed and wind velocity, referred to as the tip speed ratio, is calculated using Equation 1. Where: is the blades frequency of rotation (Hz) is the length of a blade (m) IV. CONTROL STRATEGIES A system -level layout of a wind energy conversion system and the signals used. Notice that control is most effective by adjusting pitch angle and controlling the synchronous speed of the generator. Layout of a Wind Energy System All rights Reserved 115

3 V. PERFORMANCE TESTING OF THE WIND TURBINES An important question would be; how different methods and strategies for the wind turbines and from different manufacturers can be compared with each other? The problem arises from the fact that the methods and strategies cannot be implemented on a single wind turbine at a same time! Or, two wind turbines cannot be installed at a same location! At a same time. This would make the comparisons to be for different concepts at different times when at those different times, the wind speed would also be different. This would make the evaluation process almost impossible. Many papers have studied the evaluation by simulation. They consider a simulated wind turbine together with an artificial wind speed. The artificial wind speed is represented by; a step change [2] to represent abrupt changes in the wind speed, a cosine shape change [4] to model a gustwind, a statistically generated wind speed, or a pre-recorded wind speed [8,2]. The simulation models have to be accurate enough to represent the dynamics of the wind turbine system [7]. VI. CONTROL RESEARCH Much of the advanced control research is logically divided between optimization of power capture in Region 2 and load mitigation in Region 3. In Region 2, research is further divided between investigations that incorporate detailed models of the generator electromechanical system and power electronics and those that view the generator torque in terms of a static gain (as in the previous section) that responds instantly to commanded torque. Where studies involve electromechanical models, advanced research congregates around maximum power point tracking (MPPT) and sliding mode approaches [6-8] or extremism seeking control [9]. Many of the MPPT and extremism seeking approaches can be considered nonlinear, but they may begin design with at least some form of linearization. In contrast, the techniques and references therein, incorporate the non-linear relationship between aero-dynamic torque, pitch, and wind speed. A common theme seems to account for the nonlinear dependence of on its parameters, but not structural nonlinearity. The study in [5] is apparently one of the few that designs for generator /electromechanical dynamics and the linear zed, flexible turbine structure. Additionally, adaptive approaches for maximizing power capture are studied in [1-3]. Controls for structural load mitigation (typically absent in studies on power capture) has also been developed for use in Region 2. In simulation [6], these controllers have been shown to reduce blade loading by 24%. In [30], load mitigating Region 2 controllers were designed and tested on an actual turbine (CART2, a 2- bladed turbine similar to CART3) and shown to significantly reduce tower loads. Time invariant, MIMO methods [3,4,7] tend to be the most prevalent in the research of advanced controls for Region 3, but adaptive [9,10] and novel gain-scheduling approaches are also investigated. Since even under constant wind conditions, the turbine has time-varying, periodic (with each rotation of the rotor) dynamics, work has been done on the use of periodic control [32, 33] and multi blade coordinate (MBC) based control [8,4,5]. Under non-extreme conditions, MBC and periodic control are found to be very comparable and time-invariant control is not far behind. Nearly all of these advanced MIMO methods are state space based rather than transfer-function based. State-space control techniques lend themselves nicely to MIMO systems, extend naturally to time-varying and periodic control applications, and commonly utilize an observer to estimate system states. These techniques also incorporate models of the wind, either by augmenting the observer with a model of a persistent wind disturbance a or indirectly in an output referred sense by augmenting the plant model with dynamics. The PID controller from the previous section can be viewed in terms a (persistent, output referred) step disturbance is modeled by the integrator and the proportional, derivative, and notch functions are implemented in the controller. In both of these diagrams the output y may consist of generator, blade, and tower measurements, while the control may include torque and individual pitch commands. When observer augmentation, often referred to as disturbance accommodating control (DAC), is feasible, there can be asymptotically perfect reconstruction of the modeled disturbance. Then, with an additional matching condition [7], it is possible to cancel out All rights Reserved 116

4 effect of this disturbance on the output perfectly through correct selection of the gain. With plant augmentation, the intuitive and explicit interpretation of the disturbance model as being representative of certain types of wind disturbances acting on the turbine is lost. Instead, undesirable, but expected, behaviors in the measured outputs (offsets, once per revolution (1P) oscillations) determine the disturbance model used. In lieu of an intuitive connection with exogenous disturbances, plant augmentation always provides perfect asymptotic rejection of the disturbance at the measured output. The evaluation method is based on statistical behavior of wind patterns. In this methodology, the entire wind turbine output parameters such as energy are represented with a statistical number associated with it. This is the result of considering statistical distributions of the wind speed. One such distribution is the Waybill distribution. The distribution gives the probability of having the wind speed more than a specified value. One such distribution for a site Northern Iran. It describes how dynamic the energy extraction can be. If a wind turbine cannot optimize its energy extraction dynamically, the EPF value is less than one. THE AVERAGE POWER OF THE WIND EPF= THE POWER OF THE AVERAGE WIND The Energy Pattern Factor can be between zero and four, 0 < EPF < 4. If the Energy Pattern Factor for a wind turbine is more than one, EPF > 1, that turbine can be considered as good. However the better the controller of the wind turbine, the more the Energy Pattern Factor. A computer program is developed which can estimate the energy output, accompanied by a probability number associated with it. The program can calculate the probability of having any given interest rate of the invested money, where the payback of the invested money is concerned. It can also compare two or more wind turbines with each other. It also calls a program to design a wind All rights Reserved 117

5 for any specified site. The program is fully interactive. The program creates an interactive menu for the user. The program would utilize the wind turbine information, the energy prices, the information about the wind turbine prices, and information about the site. The VESTAS V47-660KW wind turbine is chosen as an example to study the program for a site in Northern Iran. Those turbines are currently installed in Northern Iran. Based on the site information and the turbine data, the interest rates for domestic and industrial demands can be plotted as functions of the average annual wind speed and for different EPFs the EPF can result in more energy extraction from the wind One of the important and well-accepted statistical tools to evaluate the annual wind speed variations for wind energy studies is the Weibull distribution function. This function, which usually is plotted as a curve, represents the probability of having the annual mean wind speed being greater than a specified value. From that distribution the probability of having any interest rate can be found. The program performs comparison of two wind turbines and plots the interest rates for both turbines on a same plot. VII. NEW CONTROL STRATEGIES Currently, most control algorithms depend on measurements from turbine structure and drive train for use in the control feedback. Often these turbine measurements are unreliable or exhibit delayed response to disturbances acting on the turbine. This constrains the controls to react to complex atmospheric disturbances after their effects have been felt by the turbine. Thus, there is an inherent lag between the time that a disturbance arrives and the time that the control actuator begins to mitigate resulting loads. A considerable advantage in load mitigating capability can be attained by measuring atmospheric phenomena upwind of the turbine before they impact the turbine rotor. The needed control actuation signals can then be prepared in advance and applied as the inflow to the turbine changes with potentially significant load mitigation improvement. New lieder technologies are capable of measuring velocity upwind of the turbine with sample rates in the 10 s of Hz. With these measurements, it is possible to design preview controllers that can adjust pitch (and/or torque) as necessary before wind disturbances arrive at the turbine. The preview control can be designed in unison with the feedback control. A plant augmented, preview controller is designed with the same techniques as used in the design of the MIMO controller of the previous section. The generalized plant approach extends so that a combined feedback and feed forward (preview) controller can be designed. The response of the resulting preview controller is displayed along with that of the PID/notch and MIMO, feedback controllers the preview controller significantly improves performance without large increases in actuation. Preliminary investigation of preview, feed forward techniques indicate the promise of tantalizingly large improvements in controller performance. Implementation of such methods relies heavily on new measurement technologies that each come with their own, characteristic distortion and noise issues. Even if these prove to be surmountable, there remains the fact that an upstream, wind, velocity profile will hardly be the same when it arrives at the turbine. Modeling the stochastic nature of the change in wind profile as it travels and optimizing feed forward control VIII. CONCLUSION A new set of indices for wind turbine evaluation are proposed in this paper. A new computer program is developed which not only applies the methods and extracts the indices for any given turbine from different manufacturers, but also can design a wind turbine based on the user s desired parameters. The numbers and values, given by the program, are associated with a probability which takes into account the statistical behavior of the wind at a given site. It is recommended that the wind turbine manufacturers should perform such studies and provide the proposed indices in their catalogs. The indices, implicitly, take into account; (i) the site parameters, (ii) the All rights Reserved 118

6 parameters, (iii) the operational parameters, (iv) the turbine parameters, and (v) the energy pattern factor of the wind. In this paper we reviewed regions of wind turbine operation and their associated control objectives. The focus was on methods for speed regulation and structural load mitigation, but the reader interested in other areas will find Preview control uses look-ahead measurements of incoming wind disturbances to generate pre-actuation. (a) Compensation designed in conjunction with feedback control. (b) Stand-alone, feed forward control based on plant inversion. Availability of lidar measurements enables the implementation of disturbance feed forward methods. The references a good starting point. We demonstrated methods for pitch control, both standard and advanced, some of which show great potential for improved performance. In conclusion we have touched on further work which may be pivotal in realizing the improvements that recent research suggests is possible with advanced control methods. REFERENCES [1] McIver A.D., Freere P., Holmes D.G. (1995), "Grid Connection of a Variable Speed Wind Turbine", Wind Energy Workshop, Monash University, [2] Muljadi E., Pierce K. and Migliore P. (1998), "Control Strategy for Variable Speed Stall Regulated Wind Turbines", Americans Control Conference, PA, June. [3] Nanayakkara N., Nakamura M. and Hatazaki H. (1997), "Predictive Control of Wind Turbines in Small Power Systems at High Turbulent Wind Speeds", Control Engineering Practice, v5, n8, p [4] NEG-MICON (1998) "Questions for Wind Turbine Companies", Wind Energy Workshop, Monash University, DEC.. [5] Park J. (1981), "The Wind Power Book", published by Cheshire Books. [6] Prado R.A. (1995), "Reformulation of the Momentum Theory Applied to Wind Turbines", Journal of Wind Engineering and Industrial Aerodynamics, v58, p [7] Riahy G., P. Freere and D.G. Holmes (2001), "Experimental and simulation results for sliding mode dynamic wind turbine control using a dc chopper, 2001 International Conference on Power Electronics, Proceedings of ICPE'01, Seoul, Korea, October [8] Riahy G. (2001),"Technical Specifications of Wind Turbine Systems, Extensive Report Prepared for Hirbodan Co., Sadid Industrial Groupand Tehran, Iran. [9] G. Bir, Multi-Blade Coordinate Transformation and its Application to Wind Turbine Analysis, Proc. AIAA/ASME Wind Energy Symp., Jan [10] J.C. Doyle, K. Glover, P.P. Khargonekar, and B.A. Francis, Statespace Solutions to Standard and Control Problems, IEEE Trans. Automatic Control, 34(8): , Aug [11] M. Harris, M. Hand, and A. Wright. Lidar for turbine control. NREL Technical Report, NREL/TP , [12] B. P. Rigney, L. Y. Pao, and D. A. Lawrence, Nonminimum Phase Dynamic Inversion for Settle Time Applications, IEEE Trans. Ctrl. Sys. Tech., Vol. 17, [13] N. D. Kelley, B. J. Jonkman, G. N. Scott, and Y. L. Pichugina, Comparing Pulsed Doppler LIDAR with SODAR and Direct Measurements for Wind Assessment, NREL Report No. CP , National Renewable Energy Laboratory, July [14] J. H. Laks, L. Y. Pao, and A. Wright, Combined Feedforward/Feedback Control of Wind Turbines to Reduce Blade Flap Bending Moments, Proc. AIAA/ASME Wind Energy Symp., Jan [15] International Energy Agency (IEA), The Application of Smart Structures for Large Wind Turbine Rotor Blades, Proc. IEA Topical Expert Meeting, May [16] M. Lackner and G. van Kuik, A Comparison of Smart Rotor Control Approaches Using Trailing Edge Flaps and Individual Pitch Control, Proc. AIAA/ASME Wind Energy Symp., Jan. All rights Reserved 119