Lidan Zhou

Short-term wind power forecasting based on lifting wavelet transform and SVM

Short-term load forecasting is important for the safety and economic operation of the wind power system. In order to forecast the power load more accurately, the Support Vector Machines (SVM) combined with the lifting wavelet transform is proposed in this paper. The lifting wavelet transform is used to find out the characteristics of original signal while the SVM is utilized to improve the precision of forecasting. Finally, the data in September 2010 from a wind farm in North China are adopted. The result shows that wind power load forecasting based on method above is more effective than that of SVM only, thus proving the validity of the method above for power load forecasting.

Short-term wind power prediction based on wavelet decomposition and extreme learning machine

Wind energy has been widely used as a renewable green energy all over the world. Due to the stochastic character in wind, the uncertainty in wind generation is so large that power grid with safe operation is challenge. So it is very significant to design an algorithm to forecast wind power for grid operator to rapidly adjust management planning. In this paper, based on the strong randomness of wind and the short precision of BP network forecasting, Short-Term Power Prediction of a Wind Farm Based on Wavelet Decomposition and Extreme Learning Machine (WD-ELM) is proposed. Signal was decomposed into several sequences in different band by wavelet decomposition. Decomposed time series were analyzed separately, then building the model for decomposed time series with ELM to predict. Then the predicted results were added. Through a wind-power simulation analysis of a wind farm in Inner Mongolia, the result shows that the method in this paper has higher power prediction precision compared with other methods.

A novel recursive integral PI repetitive control for three-phase three-wire shunt Active Power Filter

Active Power Filter (APF) has been widely used to compensate the harmonic currents which are generated by the nonlinear load. Its compensation performance mainly depends on the design of controller. In this paper, a novel control method based on recursive integral PI algorithm and repetitive control strategy has been proposed. Compared with traditional PI control which can not get zero-steady-state error of output current due to its bandwidth limitation, this approach guarantees the steady-state accuracy with its embedded repetitive control, and improves the dynamic response performance of the APF system by its recursive integral PI algorithm. Moreover, it is more appropriate for digital signal processor implementation. The validity of the proposed method that the filtering performance is improved obviously is demonstrated by simulation analysis.

Reactive power optimization based on particle swarm optimization algorithm in 10kV distribution network

The optimization of reactive power compensation plays an important role in power system planning and designing. A mathematical model in the 10kV distribution network is established in this paper. Its objective function is the cost of investment in equipment of reactive power compensation and active power loss of the system should be the least. The node voltages beyond limited and the generator reactive power output beyond limited will be expressed in the way of penalty function. In this paper, particle swarm optimization will be used. Using PSOs characteristic of high convergence efficiency, the speed of reactive power optimization will be improved. Using the binary PSO, the algorithm can better adapt to solve the problem.

The study of hybrid modulation based on cascade SVG and its DC control method

This paper adopts PAM+PWM hybrid modulation method in a five cascade SVG (Static Var Generator). Four PAM (Pulse Amplitude Modulation) units output a voltage which is of fixed amplitude and in phase with the system voltage; and current direct controlling method was applied in the only PWM (Pulse Width Modulation) unit to adjust the output reactive current. To balance the DC-side voltage, pulses rotated cyclic between PAM units. This paper put forward a new method that rotated the pulses in an asymmetric way, in which the DC-side voltage of the PAM units was elevated and that of the PWM unit was accordingly reduced. Further, this paper successfully balanced the DC-side voltage volatility of each H-bridge by reasonably distributing capacitor parameter. Finally, PSCAD simulation results and the SVG experimental prototype prove the validity of this method.

Experimental research of DFIG based on wind energy conversation system

The paper introduces a full-open experimental platform for doubly-fed induction generator (DFIG), which based on wind energy conversation system. The paper starts from the working principles of doubly-fed induction generator (DFIG), then the controlling strategies of gird-connection and variable-speed-constant-frequency (VSCF) and maximum-power-point-tracking. The effects of the controlling strategies were examined in Matlab/Simulink environment. Finally, a full-open experimental platform was developed based on TMS320F28335 and PLC. Controlling strategies of DFIG were applied on the platform by DSP programming. The experiments results verified the validity and efficiency of the controlling strategies, and showed the stability of the platform.

Reactive Power Optimization for Distribution Network Based on Chaos Guide Particle Swarm Optimization Algorithm with Gold Criterion

Voltage is an important aspect to measure the security of power system and reactive power can relatively exert great influence on the voltage level. So planning for reactive power is an important part of network planning. In this chapter a new algorithm called Gold Criterion Chaos Guide Particle Swarm Optimization (GCCGPSO) is presented in reactive power optimization for distribution. Firstly, a mathematical model of reactive power optimization for distribution network by capacitance is established. And the cost of system active power loss and investment in equipment is treated as the optimization objective. Meanwhile the node voltage and reactive power of generator is dealt with penalty function when they pass over the limitation. Then GCCGPSO is proposed. It adopts not only chaos algorithm with gold criterion to guarantee that the particles are not easy to fall into local optimum and search the same place, but also the Neighbor domain optimal item to promote the ability of choosing path. Finally, the result of the simulation shows that the algorithm is useful and has sound performance.

Reactive Power Optimization for Wind Power System Based on Adaptive Weights Flight Adjustment Particle Swarm Optimization

In recent years, the uncertain output of wind power has had growing effects on the regional power grid. Reasonable reactive power optimization can effectively improve the adverse effects of wind power. In this chapter, an Adaptive Weights Flight Adjustment Particle Swarm Optimization (AWFAPSO) is proposed for the reactive power optimization of wind power system. First, it established a mathematic model in which system active power loss will be treated as objective function, and adopted penalty function to process node voltage cross-border and generator reactive power cross-border. Then AWFAPSO was presented. Using variable inertia factor, it can locally regulate the flight speed of the particle which leads to finding the optimal solution effectively and adopting adaptive flight time to guarantee the flight convergence in general, thus preventing particles from oscillating near optimal solution in the late of conventional particle swarm. Finally, the simulation shows that reactive power optimized by AWFAPSO can effectively reduce the system loss and improve the node voltage level.

Three-Phase Four-Wire STATCOM Control Method Based on Neural Network PI Controller

In order to solve the neutral-point imbalance problem and to improve the control precision of three-phase four-wire STATCOM, this chapter focused on the three-phase four-wire STATCOM control method based on neural network PI controller. First by analysis of the voltage imbalance problem of the split capacitors in three-phase four-wire STATCOM, a neutral-point balance control method based on the zero-sequence current is proposed. Then in order to improve the control precision, the neural network PI controller is introduced into three-phase four-wire STATCOM. Finally, the neutral-point balance control and neural network PI controller are combined together to get the neural network triple close-loop control method. Simulation result illustrates that the proposed control method is capable of neutral-point balancing control in three-phase four-wire STATCOM and the control precision is higher than that of the conventional control method.

Short-Term Wind Power Forecasting Based on Lifting Wavelet, SVM and Error Forecasting

In order to improve the forecasting accuracy, a novel forecasting method using wavelet, support vector machine (SVM), and error forecasting technology is presented in this chapter. Firstly, it utilizes lifting wavelet method to decompose data to extract the data’s main characteristics. And then it establishes the SVM forecasting model and error forecasting model to realize the wind power load forecasting, relative error forecasting, and wind load data correcting. Finally, the actual data is adopted for simulation. The experimental results show that the method based on lifting wavelet transform, SVM, and error forecasting can improve the forecasting accuracy greatly. The test shows that the method used for the wind power load forecast is feasible and effective.