The Research And Application Of Fault Diagnosis Based On Pattern Recognition Technology

Intense market competition and increasing quality demand has changed business's main attention from production to product quality. Non-price competition has become a major means of international competition. Faced with Intense competition in the quality, To research and implement a comprehensive production process control and diagnosis is the main means of improving product quality, Enhancing competitiveness. Fault diagnosis system is an efficient, reliable operation of the necessary protection of advanced manufacturing systems. On the one hand with the development of manufacturing systems, fault diagnosis system develop passive on the other hand, with the enhancement and improvement of diagnostic techniques and related hardware and software environment, it develops initiative.The basic concepts and theories of fault diagnosis are analyzed and summarized in this thesis. On the basis of the current quality control algorithm, this thesis has analyzed and researched the original process control techniques such as SPC. Proposed fault diagnosis technology based on pattern recognition against insufficiencies of modern automated production processes. And apply the algorithm to the project 'Integrated Quality Control System of 4DA1 diesel motor', achieved the desired results. The main contents and innovations are as follows:Firstly, summarized the concept and connotation of model, pattern recognition; Analysis the general model, classification and main tasks of pattern recognition system; explored several common methods of pattern recognition and analysis the advantages and disadvantages of them.Secondly, for a large number of the original sample data difficult to accurately diagnose, explored the two commonly used methods of data statute, that is feature extraction and feature selection; Using principal component analysis (PCA) to reduce the dimensionality of collected samples data, selected the main characteristics to constitute symptoms vector.Thirdly, for the existence of uncertainty and disorder sample noise in the production process; established fault diagnosis model and designed a fault diagnosis technology based on pattern-matching; identified the root causes and researched the error recognition rate and alarm rate; Using MATLAB to program the relevant procedures to achieve the algorithm, tested and analysis the algorithm, has achieved the desired results.Fourthly, applying the above methods and technologies in the plan of enterprise automatization information integration and the project of 'Integrated Quality Control System of 4DA1 diesel motor', designed pecific applications, realization of pattern-matching technology in the project application, achieved pattern-matching technology applied in the project ; Realized the value of the works.