Study On Fault Diagnosis Expert System For Hybrid Bulldozer Based On Hybrid CBR

Hybrid bulldozer is an innovation application of hybrid technology in the field of engineering machinery.Due to the Multi-system,complex structure,poor operating conditions,the cause of fault of hybrid bulldozers is more complex.So it is difficult to use traditional modeling methods to diagnose faults.In recent years,the application of intelligent theory has rapidly developed.The historical data based fault diagnosis method play a very important role in the field of fault diagnosis.In the light of this trend,this paper diagnosed the hybrid bulldozer fault based on data.Rule-based reasoning(RBR)has an important application in fault diagnosis.It summarizes the predecessors' experience and simulates the way in which the experts think about to solve the problem.Case-based reasoning(CBR)is also widely used in fault diagnosis.It solves the current problem based on the actual cases collected through historical,monitoring and experimental sources.In this paper,the remote monitoring system collects the vehicle data to set up fault diagnosis database.It is the basis of integrated RBR and CBR.Finally,a fault diagnosis expert system for hybrid bulldozer based on Hybrid CBR is formed.This system can help to find the location of the failure,the cause of the fault and the way to deal with the failure quickly.It is of great significance of research and application value to improve the reliability of the equipment,shorten the maintenance cycle and reduce the maintenance cost.Rule-based reasoning technology includes rule summarization,rule reasoning and rule management.The rule summarization is the key point of RBR.The goal of the rule summary is to summarize the knowledge in the domain to form a rule base that could be used for fault diagnosis.In this paper,the Apriori algorithm in association rules was used to mine the data collected through the remote monitoring system,to find out the relationship between the feature data and the fault,and form arule base with which to introduce the RBR diagnosis.The case search is the most important technology in CBR.The retrieval algorithm was directly related to the efficiency of CBR search module and the whole system.In this paper,K-NN algorithm was used to retrieve the cases.On this basis,AHP was used to weight the characteristic items so that it can be searched more accurately.The CBR module of the system divides the search into two steps,firstly,use SQL language for initial search,and then use the K-NN algorithm for accurate search.It ensures the accuracy of the search while improving efficiency.In this paper,RBR and CBR were connected in series to build a Hybrid CBR fault diagnosis expert system for hybrid bulldozer.This system uses Visual Studio 2010 as the software development environment,using SQL Server 2008 as database management tool.Using C#,R and SQL language as programming language in different modules of the system.