The Research Of Power Quality Monitoring And Improvement Of Wind Power Generation Unit

Wind energy as a clean energy, has a wide range of applications. One of the most important applications of wind energy is wind power which plays a huge role in promoting the development of the national economy. At the same time we should noted that wind power generation unit is a highly dynamic system, it's a big challenge to apply eligible grid connected voltage. This paper is just fit for this, it's mainly focused on the power quality research, and taking doubly fed induction machine as the research object.The structure of this paper can be divided into three parts:the study of power quality monitoring,the improvement research of power quality,the design of monitoring system. Power quality monitoring firstly research the principles of the signal de-noising, then choose wavelet transform as the denoise instrument. The wavelet threshold method denoise can reach a good effect. It not only can filter out the clutters, but also can improve the accuracy of harmonic analysis. It takes fourier adding window method to analyse harmonic component. The question is that adding window truncation will cause spectrum leakage, while the lack of sampling frequency will lead to barrier effect. Due to these reasons, the paper study the spectrum characteristics of several common window functions, and gives the selection principle of window functions as well as the inhibition method of barrier effect. For the voltage mutation, it gives the wavelet decomposition and reconstruction high frequency signals method which can accurately detect the signal mutation occurs and stop times.For the improvement of power quality part, it mainly study the power control method. Neural network PID control algorithm with the stator voltage oriented vector control technology can achieve active and reactive power control. It can also guarantee the double-fed asynchronous generator running in the condition of variable speed constant frequency. Wind speed affect the energy that generator absorbed from the wind, when the wind speed below the minimum speed to run or above the speed limit, the generator must to cut out of the power grid. The mainly purpose to track wind speed changes is to get the largest wind energy utilization. The paper studies the neural network algorithm in the prediction of wind speed which can get the real-time trend of wind speed effectively.System Monitoring section designed the major hardware and software modules. For the hardware part, the mainly circuits are as the following:main control circuit,sampling circuit,frequency test circuit,power control circuit,network communication circuit, etc. The control chip using Altera's Cyclonell Series FPGA which is a widely used chip with abundance logical resource that can meet most application require-ments. There are two main advantages of using FPGA:On the one hand, it can easily configure the internal logic resources, system upgrades and maintenance easily; On the other hand, it uses HDL programming which the different processes can be executed in parallel, the speed is more faster than any other chip. The software part mainly achieve the designs of signal de-noising and FFT, each module's software processes are also completed here.The methods and theories involved in the paper have been repeated debug and verify too many times in the simulation environment. The effect shows that it can fit the most wind power systems. But ultimately wind power system is a complex system, it still has a lot of work to do, this paper is only a preliminary study.