基于局部保持投影的光伏组件热斑故障诊断  

HOT SPOT FAULT OF PHOTOVOLTAIC MODULES DIAGNOSIS BASED ON LOCALITY PRESERVING PROJECTIONS

作　　者：张杰 易辉 王远鸣 顾梦埙 黄阅 Zhang Jie;Yi Hui;Wang Yuanming;Gu Mengxun;Huang Yue(College of Electrical Engineering and Control Science,Nanjing Tech University,Nanjing 211816,Jiangsu,China;Jurong Power Plant,Jiangsu Huadian Energy Co.,Ltd.,Zhenjiang 212400,Jiangsu,China)

机构地区：[1]南京工业大学电气工程与控制科学学院,江苏南京211816 [2]华电江苏能源有限公司句容发电厂,江苏镇江212400

出　　处：《计算机应用与软件》2021年第11期49-55,共7页Computer Applications and Software

基　　金：国家自然科学基金项目(61503181);国家重点研发计划项目(2018YFC0808500)。

摘　　要：为了辨识不同阴影状态下光伏组件的热斑故障问题,提出一种基于局部保持投影空间到空间投影的诊断方法。该方法以光伏组件的8个关键参数作为故障特征,对光伏组件的状态进行监测;通过局部保持投影算法将热斑信息在保持数据局部几何结构的基础上,投影到新的空间;使用跟踪算子确定模型中的空间边界,并直接评估光伏组件热斑故障的严重程度。实验将15组光伏组件做不同区域、不同大小的遮挡以模拟不同程度的热斑故障,该方法均能对其进行有效识别。与其他诊断方法相比,该方法在故障识别率、特征提取速度方面有一定的提高,验证了其有效性和可行性。In order to identify the hot spot fault of photovoltaic modules under different shadow states, a diagnostic method based on locally preserving projection space to spatial projection is proposed. The state of the PV module was monitored by using eight key parameters of the PV module as fault characteristics. The hot spot information was projected onto the new space by maintaining the local geometry structure of data by the local preserving projection algorithm. The tracking operator was used to determine the spatial boundary in the model and directly assess the severity of the hot spot failure of the PV module. In the experiment, 15 groups of PV modules were occluded in different regions and different sizes to simulate different degrees of hot spot faults. This method can effectively identify them. Compared with other diagnostic methods, this method has certain improvements in fault recognition rate and feature extraction speed, which verifies the validity and feasibility of method.

关 键 词：光伏组件 热斑 故障诊断 局部保持投影 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] TP181