INTELLIGENT PREDICTIVE MAINTENANCE TECHNOLOGY RESEARCH FOR LARGE OIL-FIELD WATER INJECTION UNITS

摘要

Large-sized water injection units are key equipment in oil-field and usually consist of large motors and water injection pumps in a continuous operation. The traditional maintenance way of the units is the preventive maintenance based on time which is not economical and can not completely avoid vicious accidents. To ensure the normal operation of the units, save maintenance costs and improve equipment utilization, the intelligent predictive maintenance technology based on condition is adopted to develop an intelligent maintenance system which has achieved excellent results in large oil-fields. This paper introduces the differences between the traditional maintenance way and the modern one used by the units, interprets intelligent predictive maintenance methods, analyses sensitive factor for trend prediction, provides new intelligent trend prediction methods based on neural network, establishes the neural network trend prediction model and constructs intelligent maintenance expert system based on knowledge. The developed maintenance system undergoes experimental research and practical tests on industrial scene and the result shows that the proposed new intelligent maintenance technology can better reflect the changing trend of units running condition and provide technical means to achieve condition-based predictive maintenance.