基于GMM的流体旋转设备运行可靠性在线评价方法

Online evaluation method of fluid rotating equipment operation reliability based on GMM

针对流体旋转设备运行工况多变且难以区分,导致运行可靠性评价准确率低的问题,提出了一种基于高斯混合模型(GMM)的流体旋转设备运行可靠性在线性评价方法。首先,根据设备历史运行数据,基于快速搜索和发现密度峰值的聚类算法(DPC),进行工况划分,构建不同工况条件下的基于GMM的运行可靠性基准模型;其次,使用XGBoost算法对设备实时运行状态进行工况识别,约减冗余指标,构建设备运行可靠性的评价指标体系;然后,计算度量评价指标与对应工况下基准模型指标的偏离程度,以马氏距离作为度量标准,进一步计算得到设备运行可靠性评价指数;最后,以矿用离心机设备为例,进行了多工况下的运行可靠性实例分析和模型验证。研究结果表明,该方法能够在线实时反应设备当前的运行可靠性水平,当离心机设备运行可靠性低于0.857时,认为设备进入劣化状态,且评价准确率达到98%以上。

Aiming at the problem of low accuracy of operation reliability evaluation due to the variable and difficult to distinguish operation conditions of fluid rotating equipment,a linear evaluation method of fluid rotating equipment operation reliability based on Gaussian mixture model(GMM)is proposed in this paper.Firstly,according to the historical operation data of the equipment,the working conditions are divided based on clustering by fast search and find of density peak(DPC),and the operation reliability benchmark model based on GMM under different working conditions is constructed;Secondly,XGBoost algorithm is used to identify the real-time operation status of the equipment,and redundant indicators were reduced to build an evaluation index system for the reliability of the equipment.Then,the deviation of the evaluation indexes from the benchmark model indexes under the corresponding working conditions was calculated,and the Mahalanobis distance was used as the measurement standard to further calculate the equipment operational reliability evaluation index.Finally,taking the mine centrifuge equipment as an example,the operation reliability case analysis and model verification under multiple working conditions are carried out.The result of research show that the method can reflect the current operational reliability level of the equipment online in real time,when the operation reliability of centrifuge equipment is lower than 0.857,the equipment is considered to be in degraded state,and the evaluation accuracy rate is above 98%.