Study On Fault Prognosis Method Of Hidden Markov Model For Shearer Hydraulic System

China is a big energy consumer,and coal resources occupy an important position in China’s primary energy.As an important equipment of coal mining,the safe and efficient operation is the premise for the smooth development of fully mechanized coal mining.With the increasing demand of coal mining,higher requirements are put forward for the performance of coal shearer.In order to meet the demand of multi-working condition mining,different types of shearer are used.However,with the improvement of the overall performance of the shearer and the increase of the complexity of the equipment,the types of faults gradually increase.In most cases,the failure of the hydraulic system is a progressive process,so the early fault of the fault omen information extraction,analysis,identification,and timely maintenance of the fault location,can reduce the failure to the greatest extent.As a key system of shearer,the potential failure of hydraulic system may lead to chain reaction,which may lead to other equipment failure or even the collapse of the whole mining line.Therefore,it is very important to study the fault prediction of shearer hydraulic system.At present,there are many fault diagnosis methods,but there are some limitations and shortcomings.With the development of science and technology,intelligent diagnostic methods have been put forward.As one of intelligent diagnostic methods,pattern recognition has been applied in many fields.According to the fault characteristics of hydraulic system,the method of pattern recognition can be applied to the fault prediction research of hydraulic system of shearer.The fault diagnosis method based on hidden Markov model is subordinate to pattern recognition method and has good modeling and analysis ability.In addition,based on the real spatial statistical characteristics,the evolution law of faults can be effectively described,which has been applied in the field of mechanical fault diagnosis,but for the shearer hydraulic system fault prediction research is less.Therefore,this paper takes the fault of shearer hydraulic system as the research object,and develops the fault prediction research based on Hidden Markov Model.The main research contents of this paper are as follows.Firstly,considering the randomness of the fault of shearer hydraulic system,it is difficult to study the fault equipment directly.Therefore,based on the working principle of shearer hydraulic system,the simulation model is designed.Based on the fault analysis of hydraulic system,the fault model is established.In this paper,the hydraulic system fault caused by height adjustment pump,directional control valve and adjustment height oil vat is studied and analyzed.Secondly,the time domain analysis is carried out for the collected working condition data.By analyzing the working condition data in time domain,the characteristic quantity which can represent the fault state is obtained.Then,the clustering analysis of time-domain eigenvalues is achieved through programming,the sensitive eigenvalues are screened and some parameters of the prediction model are initialized,which lays a foundation for the establishment of the prediction model.Thirdly,according to the analysis results of working condition data,a fault prediction model is established.As an extended form of the Hidden Markov Model,the Hidden Semi-Markov Model is closer to the evolution law in describing the fault law.Therefore,the study of Hidden Markov Model fault prediction theory and the study of Hidden Semi-Markov Model fault prognosis method are considered.Finally,taking the fault of shearer hydraulic system as the research object,the prediction method is verified.The feasibility of the two prediction methods is proved by analyzing the prediction results of fault samples.By comparing the prediction results of the two methods,it is proved that the fault prediction method based on Hidden Semi-Markov Model theory can improve the fault prediction accuracy of shearer hydraulic system.In addition,at the end of the paper,the work carried out in this paper is summarized,and the development of Hidden Markov Model fault research is prospected.