基于稀疏优化的气象雷达风力涡轮机杂波抑制方法研究

As the number and scale of the wind farms are increasing exponentially, the dynamic clutter caused by the rotating blades of the wind turbine has become a main factor that limits the detection performance of the weather radar systems. In the existing clutter suppression methods for weather radars, it is hypothesized that the ground clutter is stationary or slowly varying. However, the dynamic wind turbine clutter cannot be effectively mitigated by these methods. In this program, the sparse optimization techniques are introduced into the mitigation of wind turbine clutter for the weather radar. According to the wind farms scale, different sparse optimization models are built, and the novel adaptive suppression methods of wind turbine clutter are studied. The research content mainly include: (1) Automatic detection algorithm for wind turbine clutter; (2) Sparse recovery of weather data using matrix completion (MC) for small-scale wind farm; (3) Sparse separation of wind turbine clutter using morphological component analysis (MCA) for middle/large-scale wind farm; (4) Knowledge-aided fast iterative solutions for sparse optimization.This study not only meets the urgent requirement of clutter suppression in the weather radar, but also provides a further theoretical improvement of sparse sensing and multidimensional sparse recovery, which is of great significance.

由于风电场数量和规模呈指数式增长，风力涡轮机叶片高速旋转引起的时变运动杂波已成为制约气象雷达探测性能的主要限制因素。已有气象雷达杂波抑制方法均假设杂波静止或缓变，因此无法有效抑制快速时变的风力涡轮机杂波。本项目拟将稀疏优化理论引入气象雷达风力涡轮机杂波抑制，根据风电场规模大小，构建不同稀疏优化模型，研究风力涡轮机杂波自适应抑制的信号处理新方法。项目主要研究内容包括：（1）风力涡轮机杂波自动检测算法；（2）小型风电场基于矩阵补全的气象信号稀疏恢复方法；（3）中、大型风电场基于形态成分分析的风力涡轮机杂波稀疏分离方法；（4）先验知识辅助的稀疏优化快速迭代求解算法。项目研究内容既紧密结合当前气象雷达杂波抑制的迫切需求，也是对压缩感知和多维稀疏恢复理论的发展和完善，具有重要的工程价值和理论意义。

针对风力涡轮机叶片高速旋转引起的时变运动杂波，本项目将稀疏优化理论引入气象雷达风力涡轮机杂波（WTC）抑制，构建了基于稀疏优化的风电场WTC全自适应抑制信号处理方案，重点解决了以下四个关键技术问题：（1）针对气象回波和风力涡轮机杂波在多维域的统计分布特性，研究了WTC自动检测算法；（2）针对小型风电场，提出了基于矩阵补全的气象信号高精度稀疏恢复方法；（3）针对中、大型风电场，提出了基于形态成分分析的WTC高精度稀疏分离方法；（4）针对稀疏约束求解的不适定逆问题，提出了先验知识辅助的稀疏优化快速迭代求解算法。上述研究内容既紧密结合当前气象雷达WTC抑制的迫切需求，也是对压缩感知和多维稀疏恢复理论的发展和完善，具有重要的工程价值和理论意义。.本项目相关研究成果包括学术论文、发明专利和科研获奖。已发表学术论文26篇，其中SCI收录论文7篇，EI收录论文17篇；已申请国家发明专利12项。上述研究成果可推广应用于语音、图像、视频和阵列信号处理等领域。