基于大数据分析的风机叶片结冰故障诊断

Fault diagnosis of wind turbine blade ice based on large data analysis

为了实现风力发电机叶片结冰故障诊断,及时进行风机叶片除冰,消除隐患。提出了基于大数据分析的人工智能算法识别风机叶片结冰的方法。首先,用结冰机理研究和数据探索的方法对风机运行数据进行分析,初步提取了24个特征量;然后,采用遗传算法对24个特征量、滑动窗口宽度和支持向量机参数进行联合优化,并据此建立叶片结冰故障诊断模型。诊断结果表明,用该模型诊断叶片结冰故障的准确率为86.2%,比采用SCADA采集所有数据或初步提取的24个特征量作为模型输入的准确率有大幅度的提高;并且,将该模型用于另一个#2风机时,故障诊断准确率也达到了78.5%,证明了该方法的有效性,并具有较好的泛化能力,为识别风机叶片结冰故障提供了新思路。

In order to realize the fault diagnosis of wind turbine blade icing,deicing the blade in time and eliminating hidden dangers,an artificial intelligence algorithm based on big data analysis is proposed to judge the wind turbine blade icing.Firstly,the features of fan operation data are extracted by means of icing mechanism research and data exploration,and 36 features are obtained.Then,the genetic algorithm is used to optimize 36 feature parameters and support vector machine parameters,and the fault diagnosis model of blade icing is established.The results demonstrate that the accuracy rate of diagnosing blade icing fault by using the model is 86.2%,which is greatly improved the accuracy rate compared with using SCADA data and initial extracted 36 features as model inputs.Moreover,when the model is applied to the fault diagnosis of#2 fan,the accuracy rate also reaches 78.5%.This proves that the method is effective and has good generalization ability,which provides a new idea for identifying icing faults of fan blades.