Predictive Control Of Wind Turbines Research Articles

Robust model predictive control of wind turbines based on Bayesian parameter self-optimization

This paper studies the effect of different turbulent wind speeds on the operation of wind turbines. The proportion of wind power in the field of new energy generation has increased massively and has gained wide application and attention. However, the smooth operation and the stability of the output power of the wind power generation system are susceptible to wind speed fluctuations. To tackle this problem, this paper takes a 5 MW horizontal axis wind turbine as the research object that proposes a parameter adaptive robust control method to achieve self-optimization of controller parameters by means of Bayesian optimization. The 5 MW wind turbine model is utilized to verify the feasibility of the algorithm by combining the wind speed types commonly found in a high-altitude region in northwestern. The simulation results validate the effectiveness of the proposed scheme. The outcomes demonstrate that Bayesian optimization can significantly decrease the effects of wind speed instability. The output power increases by 1.9% on average at low wind speed and stabilizes on 5 MW at high wind speed. Therefore, the stable controller for wind power output is the robust model predictive controller with parameter improvement.

Model predictive control of wind turbine with aero-elastically tailored blades

The use of aero-elastically tailored blades (ATB) for large wind turbines has shown the benefit of mitigating blade loads, in a passive adaptive manner, with the design of bend-twist coupling (BTC) along the blades. The BTC design makes the blades torsionally flexible and capable of adapting to different wind speeds. However, such increased flexibility makes the turbine modeling computationally demanding and the real-time controller design more challenging. In this work, to include the ATB effect into the turbine model for control, a twofold modeling for ATB characteristics is proposed. First a static BTC distribution is added to the turbine aerodynamics to account for the blade’s pre-bend-twist design, next a second-order transfer function is introduced to approximate the blade structural dynamic response to wind speed variations. The nonlinear model of the whole ATB wind turbine is built up in Simulink, linearized and discretized into a state-space form. A model predictive controller (MPC) is developed with the actuator constraints considered. Simulation studies are conducted on a 5MW ATB wind turbine at a selected above-rated wind speed. The use of the simplified model for control is assessed and the performance of MPC is compared to the gain-scheduling baseline controller.

A novel switched model predictive control of wind turbines using artificial neural network-Markov chains prediction with load mitigation

The existing model predictive control algorithm based on continuous control using quadratic programming is currently one of the most used modern control strategies applied to wind turbines. However, heavy computational time involved and complexity in implementation are still obstructions in existing model predictive control algorithm. Owing to this, a new switched model predictive control technique is developed for the control of wind turbines with the ability to reduce complexity while maintaining better efficiency. The proposed technique combines model predictive control operating on finite control set and artificial intelligence with reinforcement techniques (Markov Chains, MC) to design a new effective control law which allows to achieve the control objectives in different wind speed zones with minimization of computational complexity. The proposed method is compared with the existing model predictive control algorithm, and it has been found that the proposed algorithm is better in terms of computational time, load mitigation, and dynamic response. The proposed research is a forward step towards refining modern control techniques to achieve optimization in nonlinear process control using novel hybrid structures based on conventional control laws and artificial intelligence.

Health-aware LPV Model Predictive Control of Wind Turbines

Wind turbine components are subject to considerable stress and fatigue due to extreme environmental conditions to which they are exposed to, especially when located offshore. Interest in the integration of control with system health monitoring has increased in recent years. The integration of a health management module with model predictive control (MPC) provides the wind turbine a mechanism to operate safely and optimize the trade-off between components’ life and energy production. In this paper, a health-aware LPV model predictive control approach for wind turbines is proposed. The proposed controller establishes a trade-off between the economic objective based on maximizing the energy production but at the same time taking into account the minimization of accumulated stress on the wind turbine blades. The controller uses an LPV model for dealing with the non-linearity of the wind turbine model and the inclusion of the stress model. The proposed approach is tested on a well-known wind turbine case.

Multivariable model predictive control of wind turbines based on Laguerre functions

Multivariable model predictive control of wind turbine was considered in this work in full-load regime in order to maximize the power generation and reduce the loads on the equipment. The controller has as inputs the pitch angle and the generator torque and accounts for constraints on both the inputs and the outputs. In order to reduce the computation time and improve the conditioning of the constrained optimization problem, the inputs were parameterization using Laguerre functions. The optimization objective was thus transformed from the identification of the input sequence over the control horizon to the identification of the Laguerre function parameters allowing a good parameterization of the inputs over the control horizon. Multiple linearizing models were employed in order to cover different operating points corresponding to variable wind speeds in the full-load regime zone.

Predictive control of wind turbine for load reduction during ramping events

With increasing penetration of wind power, the impact of its intermittence and volatility on power systems becomes more severe. A predictive control strategy for wind turbines (WTs) is proposed to deal with wind power ramping events and reduce WT load on the blades. The blade load model is based on the Blade Element Momentum (BEM) theory. The generator speed and pitch angle are simultaneously regulated to realize the control objectives. A two-stage optimization is designed in order to reduce the computational complexity. The objectives of the first stage are minimizing the ramping rate and maximizing the power generation. A trade-off is made between the two contradictory objectives by setting weight coefficients. The second stage reduces the WT load and meanwhile guarantees the power reference from the first stage is tracked. Feedback is designed based on neural network prediction to compensate the error of the prediction model. Case studies with a 1.5MW WT were conducted to demonstrate the efficacy of the proposed predictive control strategy. Simulation results show that the proposed control can reduce the WT load during ramping events and the risk of ramping events.

Automatic weight determination in nonlinear model predictive control of wind turbines using swarm optimization technique

This paper addresses the problem of automatic tuning of weighting coefficients for the nonlinear model predictive control (NMPC) of wind turbines. The choice of weighting coefficients in NMPC is critical due to their explicit impact on efficiency of the wind turbine control. Classically, these weights are selected based on intuitive understanding of the system dynamics and control objectives. The empirical methods, however, may not yield optimal solutions especially when the number of parameters to be tuned and the nonlinearity of the system increase. In this paper, the problem of determining weighting coefficients for the cost function of the NMPC controller is formulated as a two-level optimization process in which the upper- level PSO-based optimization computes the weighting coefficients for the lower-level NMPC controller which generates control signals for the wind turbine. The proposed method is implemented to tune the weighting coefficients of a NMPC controller which drives the NREL 5-MW wind turbine. The results are compared with similar simulations for a manually tuned NMPC controller. Comparison verify the improved performance of the controller for weights computed with the PSO-based technique.

Importance of Dynamic Inflow in Model Predictive Control of Wind Turbines

The efficiency of including dynamic inflow in the model based design of wind turbine controller has been discussed for many years in the wind energy community with out getting to a safe conclusion. This paper delivers a good argument in favor of including dynamic inflow. The main contributions are the use of a very simple inflow model, developed by one of the authors and the use of Pareto fronts to facilitate a safe conclusion. The approach is to compare MPC controllers designed with respectively without including dynamic inflow for a mean wind speed just above rated where this dynamics are most pronounces. For this the well accepted NREL 5MW reference turbine simulated with FAST is used. The main result is a reduction in tower fatigue load at 22% while power error, rotor speed error, generator torque and pitch rate is improved from 2 to 33%.

Short-Term Wind Speed Hybrid Forecasting Model Based on Bias Correcting Study and Its Application

The accuracy of wind speed forecasting is becoming increasingly important to improve and optimize renewable wind power generation. In particular, reliable short-term wind speed forecasting can enable model predictive control of wind turbines and real-time optimization of wind farm operation. However, due to the strong stochastic nature and dynamic uncertainty of wind speed, the forecasting of wind speed data using different patterns is difficult. This paper proposes a novel combination bias correcting forecasting method, which includes the combination forecasting method and forecasting bias correcting model. The forecasting result shows that the combination bias correcting forecasting method can more accurately forecast the trend of wind speed and has a good robustness.

Optical system design and experimental evaluation of a coherent Doppler wind Lidar system for the predictive control of wind turbine

The control of wind turbine blade pitch systems by Lidar assisted wind speed prediction has been proposed to increase the electric power generation and reduce the mechanical fatigue load on wind turbines. However, the sticking point of such Lidar systems is the price. Hence, our objective is to develop a more cost efficient Lidar system to support the pitch control of horizontal axis wind turbines and therefore to reduce the material requirement, lower the operation and maintenance costs and decrease the cost of wind energy in the long term. Compared to the state of the art Lidar systems, a laser with a shorter coherence length and a corresponding fiber delay line is introduced for reducing the costs. In this paper we present the experimental evaluation of different sending and receiving optics designs for such a system from a free space laboratory setup.

LiDAR-enabled model predictive control of wind turbines with real-time capabilities

We consider LiDAR-enabled model-based collective pitch and torque controllers that can be implemented onboard a wind turbine in a hard real-time environment, in the sense that they can be computed efficiently on standard computer hardware and that require a fixed deterministic number of operations at each call.At first, we show that linear parameter varying wind-scheduled models provide for a reasonable approximation (for control purposes) of the wind turbine response over its entire operating regime. Based on these results, we formulate two model predictive controllers making use of such wind-scheduled linear models and a quadratic cost. The first controller is based on a classical constrained receding horizon approach that leads to the efficient on-line solution of a quadratic problem. The second can be interpreted as its steady-state unconstrained approximation; its implementation is straightforward and leads to the off-line computation of gain matrices that are then wind-scheduled at run time.Both controllers are tested in a high fidelity environment comprising of both a LiDAR and an aeroservoelastic simulator, in deterministic and unfrozen turbulent wind conditions. The numerical experiments show that the receding horizon controller outperforms a standard non-LiDAR-enabled one, as expected and as already reported by other authors. More interestingly, the second simpler controller is shown to provide for an almost similar performance of the more sophisticated one, although at a much lower and trivial computational cost. This behavior is interpreted as being due to the fact that, given the high disturbance level and the frequent solution update, even a rough approximation of the control problem is still capable of capturing the essence of the LiDAR preview information.

Short-term wind speed prediction using an unscented Kalman filter based state-space support vector regression approach

Accurate wind speed forecasting is becoming increasingly important to improve and optimize renewable wind power generation. Particularly, reliable short-term wind speed prediction can enable model predictive control of wind turbines and real-time optimization of wind farm operation. However, this task remains challenging due to the strong stochastic nature and dynamic uncertainty of wind speed. In this study, unscented Kalman filter (UKF) is integrated with support vector regression (SVR) based state-space model in order to precisely update the short-term estimation of wind speed sequence. In the proposed SVR–UKF approach, support vector regression is first employed to formulate a nonlinear state-space model and then unscented Kalman filter is adopted to perform dynamic state estimation recursively on wind sequence with stochastic uncertainty. The novel SVR–UKF method is compared with artificial neural networks (ANNs), SVR, autoregressive (AR) and autoregressive integrated with Kalman filter (AR-Kalman) approaches for predicting short-term wind speed sequences collected from three sites in Massachusetts, USA. The forecasting results indicate that the proposed method has much better performance in both one-step-ahead and multi-step-ahead wind speed predictions than the other approaches across all the locations.

Nonlinear model predictive control of wind turbines using LIDAR

ABSTRACTLIDAR systems are able to provide preview information of wind disturbances at various distances in front of wind turbines. This technology paves the way for new control concepts in wind energy such as feedforward control and model predictive control. This paper compares a nonlinear model predictive controller with a baseline controller, showing the advantages of using the wind predictions in the optimization problem to reduce wind turbine extreme and fatigue loads on tower and blades as well as to limit the pitch rates. The wind information is obtained by a detailed simulation of a LIDAR system. The controller design is evaluated and tested in a simulation environment with coherent gusts and a set of turbulent wind fields using a detailed aeroelastic model of the wind turbine over the full operation region. Results show promising load reduction up to 50% for extreme gusts and 30% for lifetime fatigue loads without negative impact on overall energy production. This controller can be considered as an upper bound for other LIDAR assisted controllers that are more suited for real time applications. Copyright © 2012 John Wiley & Sons, Ltd.

Predictive control of wind turbines in small power systems at high turbulent wind speeds

This paper proposes a pitch-control technique for grid-connected wind turbines in a small power system with low sywtem inertia. Wind turbines in such a power system give rise to poor quality of power and low system reliability. The proposed pitch controller incorporates the predicted effective wind speed for power generation to have minimum effects on the power system while producing oprimum wind-generated power. Gain-scheduling is done to compensate for non-linearities of the wind turbine characteristics, and the control region has been divided into three sectors, taking on the assigned constraints of the maximum allowable system frequency fluctuations and the power limit of wind turbines. Simulation studies to test the control performance are based on the autonomous small wind-diesel power system on Koshikijima island, Japan.