基于MOPSO的抽水蓄能机组健康性能区间预测

Health performance interval prediction of pumped storage unit based on MOPSO

为描述抽水蓄能机组健康性能的随机性和不确定性,考虑不同部件健康性能水平的影响,提出一种基于信息融合与多目标智能优化的区间预测方法.首先,为有效描述机组的运行特性,利用运行监测数据构建基于高斯过程回归的机组健康状态模型,引入熵权法对机组不同部件的健康信息进行融合,构建可表征机组整体性能水平的综合健康指标;然后,将区间宽度控制系数纳入待优化参数范围,以区间覆盖率和区间平均宽度为目标约束,提出基于多目标粒子群优化核极限学习机模型参数的全局寻优策略,获取机组设备健康性能未来的波动范围;最后,通过某抽水蓄能电站机组监测数据验证所提方法的有效性.

In order to describe the randomness and uncertainty of the health performance,an interval prediction model based on information fusing and multi-objective optimization was proposed for pumped storage unit(PSU),where the influence of the health level from different components was taken into consideration.Firstly,in order to effectively describe the operational characteristics of the unit,a Gaussian process regression based unit health status model was constructed using operational monitoring data.The entropy weight method was introduced to fuse the health information of different components of the unit,and a comprehensive health indicator that can characterize the overall performance level of the unit was constructed.Then,the interval width control coefficient was included in the range of parameters to be optimized.With the interval coverage and average interval width as the objective constraints,a global optimization strategy based on multi-objective particle swarm optimization kernel limit learning machine model parameters was proposed to obtain the future fluctuation range of unit equipment health performance.In the final,the effectiveness of the proposed model was validated with the monitoring data of PSU.