Study On Assessment Of Reciprocating Compressor Performance Degradation

With the application of new technologies,fault diagnosis technology has been rapid development,rapidly toward predictive maintenance phase.Reciprocating compressor is a major equipment for petrochemical enterprises.Working medium is often harmful,toxic media,in case of failure can easily lead to accidents;With large work load and the harsh operating environment,the structure is very complex and prone to failure.So enterprises are eager to solve the problem that how to evaluate the health of the compressor quickly and effectively.In reality,the process of a device or part from good to failure is a continuous degradation process,during this period it will experience some different working condition.The change in the working state reflects the deterioration of the equipment performance.If the health of the equipment can be quantified by corresponding assessment of performance degradation method and notify the relevant person in a timely,Enterprises can make a reasonable arrangements for production planning and maintenance strategies according to the state of equipment,While reducing business costs and preventing production accidents.Taking Reciprocating compressor as the research object,this paper has researched feature selection and assessment methods for the equipment performance degradation assessment.First of all,on the basis of understanding the working principle of reciprocating compressor,the dynamic analysis of the compressor is carried out,obtaining the main vibration source of the compressor.At the same time,the vibration simulation signal of the crosshead of the compressor is initially established.It points out the characteristics of surface vibration signal:Periodicity;Impulses.Then,according to the characteristics of compressor vibration signal,a method of fault feature extraction based on AR and EMD is proposed.The analysis of measured signal shows that this method is able to extracted pulse ingredients associated with the failure of compressor.Equipment performance degradation assessment emphasized particularly on the performance assessment through the whole lifetime,and the quantization of degradation degree,instead of fault classification.A new requirement has been made for the characteristics used in the performance degradation assessment.This paper constructs the sensitivity index(IS)and trend consistency index(TCI)to select the feature set and obtained a good result.On the basis of above study,considering the disadvantages of the existing performance degradation evaluation model,poor robustness,high data integrity requirements,and too many model parameters.The performance degradation assessment methods respectively based c-means and support vector data description are proposed aiming for the problem of existing methods,through analyzing full life data for the crosshead,it proved the validity of the model.Through comparing these two methods,it proved Based support vector data description performance degradation assessment method better robustness,low data quality requirements more suitable for engineering applications.