中文摘要：

为了更加准确地预测风机齿轮箱故障,采用人工免疫算法对BP神经网络进行改进。把连续的设备健康状态分为若干个离散的等级,并设定预防性维护阈值。当故障等级达到所设定的阈值时进行预防性维护,用以减小因故障停机带来的经济损失。采取以单位时间内的维护成本率最小为优化目标的维护策略,同时在维护模型中考虑了维护和置换所用时间以及回复改善因子,进行优化求解。最后,通过实例验证了所提出的维护策略的有效性。

英文摘要：

In order to predict the failure of the gearbox of wind turbine more accurately,an artificial immune algorithm is proposed to improve the BP neural network.The continuous device health state is divided to several discrete grades,and preventive maintenance threshold value is set.Preventive maintenance is performed when the fault level reaches the threshold value,which can reduce the economic losses caused by the failure of the machine.Taking the maintenance policy of the minimum maintenance cost rate in the unit time as the optimization objective,at the same time,repair time,replacement time and reflex improving factor are considered in the maintenance model,to carry on optimum solving.Finally,the presented maintenance policy is verified by an example.