Research On Optimization Design Of Adaptive Control Chart And Equipment Maintenance Management

With the rapid development of science and technology and the more quality of life that people go after,the products and services provided by enterprises are difficult to find a foothold in the increasingly competitive global market if they have no excellent quality assurance.Control chart method,as the key of statistical process control,is an important means to control the quality of products.However,the traditional control chart method is not sufficient to monitor small process shifts and also results in a waste of monitoring resources sometimes,which is not suitable for todays lean quality management environ-ment.The design of adaptive control chart provides a good solution to overcome these problems of traditional control chart,and it has attracted the attention of many scholars.With these situations,we aim to consider the following topics.First,the improved design of np control chart and X control chart which are common-ly used in industrial manufacturing are proposed.In the second chapter,we improved the traditional Shewhart control chart(take np control chart as an example)in two aspects through the introduction of DSVSI(Double sampling and variable sampling interval)s-trategy and the MDS(Multiple dependent/deferred state sampling)strategy.One is that the control chart can adjust the sampling strategy(sample size and sampling interval)according to the sampling information obtained,by adopting the adaptive strategy,so as to avoid the waste of monitoring resources.The other one is to overcome the problem in which process monitoring only depends on the current sampling information and ignores historical sampling information,by introducing the MDS strategy,and then make control chart detect the process shift more quickly.Comparisons are made to show the capability of the proposed model in yielding great economic cost savings,in the cost of the proposed model compared to situations in which two stand-alone models are used.The design of the model is also robust to system parameters setting.In the third chapter,an improved synthetic X control chart is proposed based on the work of Wu and Spedding[1].The VSSI(Variable sample size and sampling interval)strategy is introduced to the tradition synthetic X control chart and an adjustment on the rule for signaling is also made.Com-parative study found that the improved synthetic X control chart can detect the slight process shift more quickly than the tradition synthetic X control chart.Second,besides the statistical process control method,maintenance management is also a key tool for the control of the production and service process.In process control,statistical process control and maintenance management supplement and promote each other.Therefore,it is logical and necessary to study the combination of maintenance strategy and control chart design.However,in practical application and literatures,people tend to separate them.In this dissertation,the study on the combination of these two methods is also discussed.In the third chapter,an integrated model which combine the imperfect maintenance and the improved synthetic X chart in manufacturing environment is proposed.The cost function of the integrated model is also constructed.Finally,the optimal design scheme of the control chart and the optimal maintenance strategy are obtained through minimizing the total cost per unit time of the integrated model.Referring to specific data,a comprehensive sensitivity analysis of the integrated model and the improved synthetic X chart is conducted.The results are as follows:1)An economic benefit of 2.64%is obtained by the joint optimization model;2)The proposed model is robust to the parameters setting,which has practical implications for practitioners.Although the control chart tool is developed and used in the manufacturing envi-ronment,it is also applicable in the service industry.The key is to find the appropriate and monitorable quality characteristics.In the fourth chapter,under the environment of service industry,an M/M/1 queueing system with finite buffer is considered.The TBE(Time-between-events)control chart is integrated in the maintenance model for ser-vice facility to monitor the system condition.Both imperfect maintenance and perfect maintenance are considered.Based on the failure interval,the TBE control chart deter-mines whether a maintenance is required when a failure occurs.Whenever the system enters an idle period,it has the opportunity to enter the planned maintenance.The RG-factorization method is applied to derive the steady-state performance of the system.Numerical experiments show that the proposed maintenance model yields significant cost savings comparing to the model without using TBE control chart.The improvement is more significant for a busier service system.In the fifth chapter,firstly,a Bernoulli scan statistic control chart for prospective monitoring in public health applications is intro-duced.And then an integrated model which combine an M/M/1 queueing system with infinite buffer and the specific Bernoulli scan statistic control chart is developed.The model is used to detect a rate increase.The matrix-geometric solution method is applied to derive the steady state performance of the queueing system.An availability function is also introduced to evaluate the model.Finally,the effects of system parameters on both the queueing system and the availability function are analysed.At last,we summarize the main results and innovative points of this dissertation,and suggest some promising research topics yet to be pursued.This dissertation provides some guidance on the application of control chart in manufacturing and service industry,and to some extent expand the research on control chart design.Therefore,this dissertation is significant on theory and utility.