基于改进灰色马尔可夫模型的设备故障率预测

摘要

Based on the characteristics of time variation and randomness of power equipment failure rate , this pa-per proposes a state division method for fitted values error series of GM (1, 1) model by fuzzy C means clustering method based on grey GM (1, 1) model.By computing the state transition probability matrix of error series , a grey Markov prediction model is established .This model analyzes high ability of GM (1, 1) model processing monotone series then explores random fluctuation responses when extracting data through transformation of state transition probability matrix .So the worst prediction results when the maximum probability state is not the actual state can be avoided .The actual case study validates that the grey Markov forecasting model based on fuzzy C means clustering method has better performance than conventional GM (1, 1) model and grey Markov forecasting model based on K means clustering , and has higher forecast precision .%电力设备故障率具有时变性、随机性、回退等特点，预测难度大，因此在灰色GM （1，1）模型的基础上，采用模糊C均值聚类方法对GM （1，1）模型拟合值的误差序列进行状态划分；通过计算误差序列的状态转移概率矩阵，建立了电力设备故障率的灰色马尔可夫预测模型。该模型既考虑了GM （1，1）模型较强的处理单调数列的特性，又计及了通过状态转移概率矩阵的变换提取数据随机波动响应的特点，避免了最大概率状态不为实际状态而出现最差的预测结果现象。通过实例证明，基于模糊C均值聚类的灰色马尔可夫模型预测结果优于传统的GM （1，1）模型和基于K均值聚类的灰色马尔可夫预测模型，具有较高的预测精度。