Wear Monitoring Platform Development For Petrochemical Key Unit

As the pillar industry of China, the petrochemical industry is playing a very important role in the national economy and development.Its productions contain lubricating oil, diesel, gasoline, chemical raw material, synthetic resins,fertilizers and other 3,000 kinds of petrochemical products, which have a close relationship with people’s food, clothes, etc. On the other hand, this is also a high-risk industry. Once the petrochemical industry has some problem, then it must affect the related downstream industries. With the continuous development of technology, petrochemical machinery is becoming large-scale, complex progressively. How to ensure the equipment run safely, stably, long, excellently is an urgent problem to be solved at present.This paper describes and analyzes the current trend prediction methods such as methods based on artificial neural network, support vector machine theory, time series, etc. On this basis, it proposes a trend analysis method for petrochemical equipment wear condition. Oil monitoring technique mentioned in this paper is that using optics, electricity, and other technological means,and gather the equipment lubricating oil sample for running physics and chemical performance, The device is currently located lubrication, wear condition,qualitative and quantitative description of the device status. Then determine thestate of equipment, and forecast its development trend. Since the oil monitoring technology collects kind of information quantities, so the multivariate information fusion is introduced.The key of the monitoring platform is the data analysis system.Because the number and type of collected information is too large,so we choose multivariate information fusion to solve this problem and simplify data analysis process.Choose three physical and chemical indicators :kinematic viscosity, acid value,moisture content as analysis variables which affect the oil and petrochemical equipment most. Select cluster analysis and discriminant analysis method of multivariate statistical analysis as the major method.The paper uses SPSS data analysis software to structures platform system. Then set up bench in the laboratory, sampled the lubricating oil from fixed gear box on a regular time.There are 150 samples of lubricating oil,test their physical and chemical indicators(kinematic viscosity, acid number, water content) and cluster and discriminant analysis operations, to establish the comprehensive evaluation model. The resulting decision function formula is applied to discriminate the petrochemical equipment’s lubricating oil monitoring data and verify its feasibility in working conditions field.According to the judgment result, it shows that the criterion for determination of petrochemical equipment lubrication state is feasible.