Research On Fuzzy Neural Networks Fault Analysis Technique For NC Machine Center

1.PurposeBased on the subject of the national 863 topic , put forward the scientific research of this thesis . On the foundation ofthe domestic and international popular Computer Numerical Control system ofextensive research, reliability of the domestic popular CNC system is researched .Under current situation of the development of CNC system , we study fuzzy neuralnetwork for hydrocarbon detect ,FMEA analysis and the reliability in developcourse. Intellectual FMEA of the CNC system research has been tried. It establishthe foundation for the systematic industrialization of popular CNC system , offerthe basis for systematic improvement, develop for follow-up high-grade , the wholefunction type CNC system reliability study.2. The theoretics of fuzzy neural network and synthesis analysisWith its applications in various fields, fuzzy comprehensive evaluation isbecoming more and more significant. However, how to determine the weights andgive the single factor evaluation matrix and avoid unreasonable results are neededmore researches. Though many people have put forward some new methods toimprove comprehensive evaluation model, the results are not always satisfied. Withthe advances in artificial intelligence , people begin to use the neural networktechnology and fuzzy techniques in comprehensive evaluation.3. Failure analysis of MCFailure analysis is for estimating products reliability which are under certainwork condition. It presumes reliability of system by reliability of system'scomponents, parts and sub-system. Because a certain non-linear mappingrelationship exists between reliability index of sub-system and reliability index ofwhole CNC system, it happens that the neural network has inherent non-linearcharacteristics , and the characteristics that can study by training. On the basis ofthese characteristics we attempt to carry on prediction and assessment throughunique prediction and learning function of fuzzy neural network.For X type MC,its tool exchanging failure has high frequency,so we setdown regulation of debugging and examining whole machine and design specialtechnics equipment to adjust cutting tool petiole;for damage of electrical sourcemodule due to fluctuating of power supply, we choice guard device of AmericanVickers and take into account unstable supply voltage in choosing regenerationresistance,we also pay attention to that advanced country's power supply is betterthan ours and their criterions are not always fit to us;for NC system failures, wedesign new hanging damping vibration bracket of three dimension to protect harddisk, we adopt closed temperature control electric cabinet so as to improve itsrunning condition and enhance its working reliability;for severe electric earthingfault, we adopt more place earthing style, make NC system earthing singly in orderto remove various disturbance in MC working time.Through the argumentation above, the assessment and Failure analysis ofreliability based on fuzzy neural network model is feasible4. CNC system FMEAWe have discussed the analysis procedure of FMEA of CNC system and takensome type CNC system as an example and carried on FMEA analysis. Based onFMEA analysis for whole machine and subsystem of MC with extension method,the dissertation has proposed a series ways of reliability improvement fordeficiencies of the two types MCs.We have known the bottleneck of the CNCsystem .By using RNP method and draw the main risk position that influences thesystem. We have set up intellectual FMEA based on neural network in tentativefinally. Establish the foundation of follow-up computer reliazer.5. ConclusionThrough the research of this thesis, develop the working content of reliabilityin the course after substantiating the CNC system. Four following conclusions thatso main as to appear:1.Have set up the systematic reliability failure analysis of CNC system basedon fuzzy neural network and model of the assessment, develop to MC course anddependability intelligence of level that use live evaluate thinking of widenning.Establish the foundation.2.Utilize method , FMEA of lathe , that laboratory summarize for many yearsanalyse concrete FMEA to a serial MC, draw into and give drive unit , main board ,measure unit as the weak link of this serial CNC . The reliability of improvingthese three parts is the key to improving this serial CNC reliability. Is it count withfailure analysis measure assess some type system failure mode , draw componentsand parts damage most heavy harmfulness for this type MC to have priority,secondly it is the overload, the reason is not clear, the operating system does notwork. So, as to this systematic development of serial numerical control, strengthenand select the electronic devices and components for use and control, make thedaily components and parts of numerical control system and select criterion for use.It is the key to improving reliability of developing.3. This thesis sets up supplier's norm to require, in reliability designingtechnique of the products, has introduced more popular reliability designingtechnique.4. Studying to CNC reliability design of numerical control based on fuzzyneural network, develop to book the initiative dependability design research underthe life-span and play a role in inspiring for the dependability appraisal of thefinished product in the past.