Research And Application Of Vibration Diagnostic Standard For Propane Compressor In Variable Conditions

The national vibration standard for machinery is applicable to the equipment of the rated working speed and steady-state operating conditions.However,the equipment is often subjected to variable operating conditions such as speed fluctuations and load changes.Direct application of national standards can easily lead to errors in the diagnosis of equipment status.In view of the above problems,this paper presents a fusion variable condition threshold model,and then a vibration standard was established suiting for variable working condition,which is of great significance for improving the accuracy of the diagnosis and ensuring safe and reliable operation of the equipment.The main research contents are as follows:(1)Research on the construction method of variable condition threshold model based on Relevance Vector Machine.The RMS of vibration velocity was selected as the feature index;Relevance Vector Machine was used to fit the statistical mean and standard deviation to working condition parameters to obtain a continuous function.The variable speed threshold model was constructed based on the Chebyshev inequality and the two keys parameters of factor k and kernel function width coefficient ? in the model were optimized and selected;the experiment verifies that the model can achieve the alarm threshold setting when the equipment is running under changing conditions,and has a higher accuracy;through the analysis of the mechanism,the influence of speed and load is discussed.The comprehensive condition parameter Z is proposed and the variable condition alarm threshold model is constructed.(2)Research on the construction of threshold model based on Full Vector Spectrum for information fusion.As for the problem of low recognition rate in low-speed areas for variable speed alarm threshold models built with single sensor information,the Full Vector Spectrum was used to fuse the homologous two-channel signals obtained by orthogonal sampling to obtain a fusion signal;the IFFT is used to obtain the fused time domain waveform,and the feature are extracted.A continuous function is obtained by using Relevance Vector Machine fitting,and then an alarm threshold model is constructed.The experiment verifies that this model solves the problem that the model built with single channel information has low alarm accuracy in low speed zone.The accuracy increased from 79.75% to 98.10%.(3)Establishment and verification of propane compressor vibration standards in oilfield.The propane compressor data of Hade in the Tarim oilfield was statistically analyzed.The normal data was used to establish the alarm curve;the normal and abnormal data were used to establish the shutdown curve.The on-site case verified that the built standard can detect anomalies and failures in time,avoiding missed accidents.(4)Design of vibration standard query and diagnosis system.A standard query and diagnosis system was developed based on C# and SQL Server,which can achieve vibration standard query of variable conditions,status level classification,historical trending query,diagnostic analysis and equipment management and other functions,supporting for oilfield equipment maintenance management and scientific decision-making strongly.