小故障样本条件下的风电机组分层贝叶斯可靠性模型

建立风电机组部件的可靠性模型，并准确估计模型参数，有利于提高风电机组全寿命周期的健康管理水平。新投运机组的故障样本较少，大样本条件下的传统可靠性建模及其参数估计方法不再适用。将贝叶斯可靠性理论应用于小故障样本条件下的风电机组部件可靠性建模及模型参数估计中。以其他风电场的故障数据为模型参数的先验信息，建立了分层贝叶斯可靠性模型，通过Gibbs算法求解模型并获得模型参数的后验分布，以标准均方根误差及可靠度函数的95%置信区间的平均宽度作为衡量指标，对比了传统可靠性模型、一般贝叶斯可靠性及分层贝叶斯可靠性模型的建模精度。最后以风电机组发电机碳刷为例，验证了小样本故障条件下分层贝叶斯可靠性模型的优越性。

Hierarchical Bayesian Reliability Model for Wind Turbines with Small Fault Sample Sets

It is beneficial to improve the health management level of the full-life-cycle of wind turbines through establishing the reliability model for wind turbine and estimating its parameter accurately. As the fault sample size of new constructed wind turbines is small, traditional reliability modeling and parameter estimation methods with large sample sets are no longer applicable. This paper applies Bayesian reliability theory to reliability modeling and parameter estimation of wind turbine with small fault sample sets. Based on the fault samples from other wind farms as the priori information of the model parameters. A hierarchical Bayesian reliability model is established. Then the proposed model is solved by the Gibbs algorithm and the posterior distribution of the model parameters is obtained. The normalized root mean square error and the mean width of 95% confidence interval of reliability function are chosen as measurement indices. The comparison of the modeling precision among the traditional reliability model, general Bayesian reliability model and hierarchical Bayesian reliability model are performed. Finally, the generator carbon brushes from various wind farms is taken as an sample to demonstrate the superiority of the hierarchical Bayesian reliability model with small fault sample sets.