Research On Dynamic Modeling And Adaptive Control Technology In Large Wind Farm

Compared to traditional fossil fuel power generation systems, wind power has its own advantages:first, the environment-friendly characteristics (renewable and non-polluting resource) of wind farms; second, although the initial investment is high (for mechanical inputs, siting and installation), but operation costs of wind farm is small respectively, which compared with fossil fuel power generation systems have a considerable competitive edge. Researches in the field of wind power control focus on maximum wind energy capture and output power regulation and quality compensation, etc. The former is committed to maximizing the available wind energy converted into electricity and the latter studies how to improve the output power quality of wind farm.Firstly in my thesis, a detailed model of the wind turbine and operating characteristics of wind farm are introduced, and also briefly the collection system and overall control framework wind farms.Accurate measurement of the wind speed is hardly available in practice due to its dynamic nature both in space and time domain. In this paper, a novel approach is proposed to estimate the effective wind speed by using high-gain observer and Inexact Newton Backtracking (INB) Method. Subsequently, a nonlinear adaptive tracking controller is designed to perform optimal output power tracking by generating a desired trajectory with the help of the effective wind speed estimate. Furthermore, chattering is eliminated in the torque signal to mitigate mechanical stress. Validation has been carried out on a 1.5-MW three-blade horizontal axis and up wind variable-speed wind turbine.For wind farm power quality problems, this thesis first to enhance the dynamic equivalent modeling accuracy of wind farms (especially transient characteristics), proposes a dynamic modeling method using multi-class SVM (Support Vector Machine) based on binary tree to extract patterns and characteristics of pitch angle movements, operating conditions and status of each wind turbine. Then the compound system structured by multiple wind turbines can be classified and represented as a respectively simple model of 4 paralleling characteristic turbines. Simplification of collection system and feature parameter calculation method are given next. To tackle power quality problems caused by wind speed fluctuations and nonlinear loads in wind farm operation, an improved ADRC (Active Disturbance Rejection Control) current compensation control is proposed on the basis of the mainstream hysteresis current control. When the current distortion is small and high-frequency harmonics exist, ADRC is introduced to implement compensation regarding the weak voltage and load current fluctuations as a disturbance signals and construct ESO and nonlinear control law, for enhancement of output power quality compensation accuracy at high frequency jitter characteristic of wind farms.