基于SVM的永磁直驱风电系统滑模变桨距控制

发布时间：2018-10-20 12:26

【摘要】：为改善风力发电系统在恒功率运行区域内的动态性能,提出了一种基于在线支持向量机的滑模变桨距学习控制方案。学习控制分为2步:第1步采用滑模控制器(SMC)进行控制,支持向量滑模学习控制器(SVM-SMC)通过离线和在线学习得到控制器的结构和初步优化参数。当学习到达一定程度时,变桨距系统切换到SVM-SMC控制;第2步采用探索机制,实际控制量由SVM-SMC控制器输出加上均值为零的正态分布的随机扰动构成,根据预测性能指标得到学习样本,通过在线学习算法对控制参数进行实时优化。构建了永磁直驱风力发电系统变桨距控制模型并在RT-LAB实时仿真系统上进行半实物仿真比较,结果表明该方法不仅具有滑模控制的抗干扰、对变化参数鲁棒性强以及速度加快等优点,在保证功率输出稳定在额定值附近的同时,实现桨距角的平稳调节,减轻了机组疲劳度和组件间的磨损。
[Abstract]:In order to improve the dynamic performance of wind power system in constant power operation region, a sliding mode pitch learning control scheme based on online support vector machine is proposed. The learning control is divided into two steps: in the first step, the sliding mode controller (SMC) is used to control the controller. The support vector sliding mode learning controller (SVM-SMC) obtains the structure and initial optimization parameters of the controller by offline and online learning. When the learning reaches a certain degree, the variable pitch system switches to SVM-SMC control. In the second step, the exploration mechanism is used. The actual control is composed of the output of the SVM-SMC controller and the random disturbance of the normal distribution with zero mean value. According to the prediction performance index, the learning sample is obtained, and the control parameters are optimized in real time by online learning algorithm. The pitch control model of permanent magnet direct drive wind power system is constructed and compared with that of RT-LAB real time simulation system. The results show that this method not only has the anti-interference of sliding mode control. With the advantages of high robustness and high speed to the variable parameters, the steady adjustment of the pitch angle is realized while the power output is stable near the rated value, and the fatigue degree of the unit and the wear between the components are reduced.
【作者单位】：湖南工业大学电气与信息工程学院;
【基金】：国家自然科学基金(61074067,21106063) 湖南省自然科学基金(13JJ3110) 湖南省高校科技创新团队和教育厅优秀青年项目(11B038)资助项目
【分类号】：TM315

【相似文献】