GA优化TS-FNN的架空线路荷载风险预测

Risk Forecast of T-S Fuzzy Neural Network by Optimized GA for Overhead Line Loads

极端天气下组合荷载的冲击对架空线的运行可靠性产生严重影响,研究架空线路风险预测与评估在预防线路事故中具有现实意义。架空线路荷载风险预测属于求解强耦合时变非线性系统问题,难以建立精确的数学模型求解。基于线路荷载-强度的随机特性与干涉原理以及模糊预测理论,建立了基于GA优化T-S模糊神经网络的线路风险预测模型,提取极端天气下的气象信息典型特征值风速、覆冰厚度、降雨量、气温作为模型输入量,以线路失效概率划分的时间尺度上线路的荷载风险状态作为预测输出量,并采用遗传算法对模糊神经网络参数进行优化。同采用传统理论计算模型和自适应模糊神经网络模型相比,所建立模型具有计算速度快、预测准确度高的优点。具体应用实例验证了模型的实用性和高效性。

The impact of combined loads under extreme weather conditions adversely affects the operation reliability of overhead line. To study the risk assessment of overhead line in prevention accident has practical significance. Overhead line risk forecast is a time-varying and nonlinear problem with strong-coupling, which is difficult to establish accurate mathematical model. According to the random properties of load-strength, the interference theory of load-strength and fuzzy predication theory, a new risk forecast model based on the T-S fuzzy neural network by genetic algorithm(GA)is established for overhead line loadsrisk predication, which takes the typical meteorological characteristics under extreme weather conditions as inputs,such as wind speed, ice thickness, rain fall and air temperature, and regards time-scale failure probability of overhead line as an output, moreover the parameters of fuzzy neural network is optimized by genetic algorithm. The built model has advantages of faster computation speed and higher forecast accuracy over traditional theory model and adaptive fuzzy neural network model.The case study of an actual overhead line verifies the effectiveness and efficiency of the proposed model.