Abstract: Some problems in diagnosis extent,depth and prediction of equipment in western oilfields were analyzed.The group diagnosis and correlative prediction models based on the modern methods were proposed.The qualitative diagnosis model was combined with the quantitative data to build an integrated diagnostic database,and a group diagnostic model was established.The group diagnostic model can be used to process and analyze the data of equipment group to solve the problem of diagnosis extent.The identifying model of time variation was built,on the basis of radial basis function neural network to solve the problem of diagnosis depth.The correlative prediction method was used to predict the maintenance periods based on parts recomposing method.The diagnosis and prediction systems for equipment group in the western oil fields were developed.

Key words: western oilfield of China, equipment diagnosis, group diagnosis model, correlative prediction, time variation model, radial basis function neural network

摘要： 分析了西部油田设备诊断在广度、深度、预测方面须解决的实际问题,结合目前先进、适用的诊断方法与技术,提出了群诊及关联预测模型框架.其核心思想是:结合“定性诊断模型”与“定量数据”构建“综合诊断库”,建立能够对“设备群”进行数据自动处理与故障分析的“群诊模型”,解决西部油田设备诊断的“广度”复杂性问题;利用时变基频的求解方法,建立基于径向基函数网络的时变基频识别模型,解决“深度”复杂性问题;利用基于部件重组的“关联预测”方法,对多台设备同时送修周期进行正确预报,形成西部油田设备群诊和预报体系.

关键词: 群诊模型, 关联预测, 时变模型, 径向基网络, 油田设备