Reliability Growth Design Of High-speed Precision Press Based On Reliability Allocation And Prediction

As the basic manufacturing equipment in machinery manufacturing industry, punchingequipment has been applied widely in automobile manufacturing, aerospace, shipbuilding,appliance, and other fields, occupying a pivotal position in the national economy. Technicallevel and quantity of forging equipment is an important symbol of national industrialtechnology. Simultaneously, with the progress of industrialization, accuracy, efficiency,automation, flexibility and reliability level of forging equipment is required higher andhigher. Punching machine is typical representative of forging equipment, being called“printing press” which is playing important role in the industry field.High-speed precision press is typical basic industrial equipment for achievingmodernization of equipment manufacturing, representing the most advanced level ofdomestic punching equipment. Currently, such machine has possessed the characteristic ofhigh precision, high efficiency, automation and flexibility. However, comparing with similarforeign machine tools domestic machine tool is facing with a difficult situation thatreliability of this kind of machines is being at low level. Reliability has become one of themain factors restricting the development of domestic punching machine. How to solvereliability growth problem is an important research direction of reliability engineering.Reliability allocation and prediction are the core problems to achieve reliability growth ofhigh-speed precision press.By researching the following aspects a set of technical process for reliability growthdesign of high-speed precision press is created. The detailed works are as follows:(1) Development and status of forging machine at home and abroad, development ofreliability engineering technology, reliability research status of punching machine andresearch status of reliability allocation and reliability prediction technology. Technologyroadmap of reliability growth design is designed for high-speed precision press whichincludes reliability analysis technology, reliability assessment modeling, reliability allocationtechnology and reliability prediction technology.(2) FMECA(failure mode effect and criticality analysis) is the foundation work formachine reliability design of high-speed precision press. Statistical analysis of fault location,failure mode, failure pieces sources, failure nature is conducted to identify the mainweaknesses of machine for each subsystem of high-speed precision press based on historical failure data. With the advantages of quantitative methods dealing with uncertain informationbased on fuzzy theory, failure hazard analysis is conducted by combining fuzzycomprehensive evaluation method and FMECA method. According to the order of fuzzycomprehensive hazard of punching machine failure mode by fuzzy FMECA technology, aseries of reliability improvement measures is suggested for the main failure modes ofmachine tools which provides important basis for reliability growth design.(3) Reliability assessment model could reflect the general law of machine reliabilitycharacteristic quantity. The main drawbacks are indicated by analyzing common modelingmethods used to assess the reliability of complex systems. Weibull neural network modelwhich has good generalization performance is presented for reflecting the true distribution ofpunching machine reliability to the extent possible based on analyzing characteristics ofreliability distribution model and neural network. Improved EM algorithm based onbootstrap sampling for estimating parameters of Weibull neural network model is designedof which the effectiveness is verified by concrete examples. Comparing with the process ofreliability assessment including point estimation and interval estimation of reliability indexesthe results showed that Weibull neural network model could better reflect the actualdistribution of fault data of high-speed precision press.(4) Reliability allocation is the key technology in reliability growth design process forhigh-speed precision press. For machine products, cost is particularly important, given thereliability factors associated with machine costs including direct investment, technology,maintenance, environmental conditions, repairment, fault hazard loss, which are convertedinto fuzzy comprehensive costs by combining expert fuzzy scoring and AHP methods. Asingle parameter log type of effort function applying to each subsystem of high-speedprecision press is defined based on the correlation between cost and reliability of machineproducts. Taking reliability indexes as the constraint and fuzzy comprehensive cost asobjective, a reliability allocation optimization model is modeled based on the new effortfunction. With the penalty function method to eliminate the model constraints, the model issolved by simulated annealing method for actual reliability allocation case of high-speedprecision press getting the reliability allocation design program.(5) After reliability allocation design reliability prediction is used to predict thereliability indexes for high-speed precision press. lots of reliability factors includingfunctional complexity, structural complexity, repairability, technical level, parts quality grade,work load, maintenance status are compared in the progress of reliability prediction for theold and improved machines. Reliability factors are scored by experts based on engineering experience for each subsystems of the old and improved machines, and the average weightedscores are used for computing reliability growth coefficients of all subsystems of high-speedprecision press. Reliability growth prediction model is modeled based on Monte Carlosimulation and reliability growth coefficients. Reliability indexes of improved machine couldbe predicted by the model and the feasibility of the model is verified by actual case.By studying the contents above，failure analysis methods of high-speed precision pressare further improved; an assessment model are established to better reflect the reliabilitydistribution of high-speed precision press; a minimum effort reliability allocation methodconsidering fuzzy cost is proposed; a reliability prediction model based on fuzzy analyticalhierarchy process for high-speed precision press is established. Finally analysis and prospectfor the future trends of reliability assessment and reliability growth is carried out. Theinnovations of this thesis are mainly as follows:(1) Multilevel fuzzy comprehensive evaluation is applied to FMECA of high-speedprecision press. Quantitative analysis of harmfulness of failure mode of high-speed precisionpress is carried out by fuzzy FMECA method, which makes up for the weaknesses of relyingon qualitative analysis of traditional FMECA method.(2) Weibull neural network model is established for reliability assessment of high-speedprecision press. Improved EM algorithm based on bootstrap sampling is designed fortraining parameters of the network. The network model has good generalization performanceand adaptive characteristics effectively avoiding optimization model selection and goodnessof fit test in reliability modeling process.(3) A single parameter log type of effort function reflecting the relationship betweenfailure rate and cost of machine is defined. Taking the lowest fuzzy comprehensive cost astarget minimum effort reliability allocation model is established based on the effort function,which overcomes the problem of lack of quantitative basis in reliability allocation processfor high-speed precision press.(4) A group of reliability factor weights are confirmed by combining expert fuzzyscoring and AHP method in the reliability allocation and prediction process of high-speedprecision press. A reliability prediction model solved by Monte Carlo simulation based onfuzzy analytical hierarchy process is used for verifing rationality of reliability allocationschemeof high-speed precision press.