Abstract: The large centrifugal pumps are widely used as the waterflooding for oil production in the fields,stabilizing and increasing the production of the crude oil.Their most of working problems of the units are caused by the vibration faults.This paper discusses the vibration-signal-based auto-diagnosis technology and system for the centrifugal pumps on basis of fuzzy mathematics and puts forward the concept and method of Compound Knowledge Base (CKB) according to the structural features of pumps,which effectively overcomes the shortages of the traditional expert system.The detecting methods,characteristic parameters and fuzzy inference of the pump units are confirmed in this paper.And the corresponding system software is also developed.Application of the system to waterflooding pumps has obtained better results in oil field.

Key words: centrifugal waterflooding pump, vibration, auto-diagnosis, Compound Knowledge Base, fuzzy mathematics

摘要： 大型离心式注水泵机组振动故障是影响注水泵正常生产的主要问题.本文以模糊数学为基础,研究了基于振动信号的离心式注水泵故障自动诊断技术及系统,根据注水泵的结构特点提出了复合知识库的概念,有效地克服了传统专家系统的不足,确定了对注水泵进行诊断的检测方案、特征参量和模糊推理机,并编制了相应的诊断系统软件.将该系统应用于现场使用的注水泵机组,取得了良好的效果.

关键词: 离心式注水泵, 振动, 复合知识库, 自动诊断, 模糊数学