Research Of Assembly Quality Based On Process Capability

Products are mostly composed of various parts.During product design,in order to reduce costs and improve quality,the tolerance of parts will be reasonably defined including dimensional tolerance and geometric tolerance,according to structure,function,appearance and assembly requirements.In manufacturing stage,the enterprise will test the quality error of parts to see whether it is within the tolerance or not and will decide whether the quality is qualified,and then conduct assembly in accordance with the process planning.Since the assembly dimension is difficult to directly measure,the enterprise often focuses on the quality control of parts to meet the tolerance requirements,rather than directly evaluate the product quality after assembly.However,even if the products are assembled with qualified parts,they still have subtle differences in quality and performance.Based on the assembly universality in reality,this paper studies on the evaluation and application methods of assembly quality(quality fit capability),according to the statistical process control theory.The detailed research content and conclusion include:First,the analysis of quality fit capability in normal distribution.First of all,on the basis of previous studies,this paper conducts systematic analysis on quality fit of two quality characteristics from tolerance feature(symmetry/asymmetry)and machine distribution feature(no shift/shift)two aspects.Consequently,it proposes the corresponding calculation method of quality fit capability index(QFCI),analyzes some properties and puts forward advice&suggestions on production practice.And the differences of fit capability and process capability in regular purchasing and urgent purchasing are compared also.Next,regarding to the assembly relation composed of multiple quality characteristics,this paper provides the calculation method of the QFCI of linear and non-linear dimension chains from the perspective of closed ring of dimension chain.Second,the analysis of process capability and quality fit capability in skew-normal(SN)distribution.Influenced by tool wear,processing habits and other factors during the machining process,the machining error of parts always tends to be partial distribution with the asymmetry of left and right sides,rather than strictly obey normal distribution.Since the conventional methods cannot be applied in process control under the non-normal distribution,according to SN distributed functions given by Azzalini,this paper considers the tail behavior of distribution and puts forward the calculation method of process capability index(PCI)based on SN distribution.On this basis,since SN distribution has no linear additive property,this paper studies on the calculation method of QFCI with two quality characteristics satisfying some distribution requirement(SN-normal distribution,or two special SN distributions).For the general situation of both two quality characteristics under SN distribution,according to the convolution formula,this paper gives the probability density function of the assembly relation,and its characteristic function.Based on these functions,this paper puts forward the calculation method of QFCI according to the principle of 99.73%.By comparing the normal distribution and SN distribution,this paper found that,for the same PCI,the quality characteristic with SN distribution allows bigger dispersion degree compared with the one with normal distribution,and the greater the slant parameter is the bigger the dispersion degree is allowed.And for the same mean and variance,the QFCI of the quality characteristic with SN distribution is higher than it with normal distribution.Third,the analysis of process capability and quality fit capability with the requirement of composite tolerance.For the quality characteristic with dimensional tolerance and geometric tolerance requirements,this paper studies on the tolerance compensation,puts forward the calculation method of PCI and makes an example analysis on supplier evaluation,from the point of independent/relevant relations among tolerances(including envelope requirement,maximum material requirement and reversible requirement,least material requirement and reversible requirement).For the quality characteristic with dimensional tolerance and geometric tolerance at the same time,according to the principle of 99.73%,this paper puts forward the QFCI calculation methods of two and multiple quality characteristics.And the probability density functions and characteristic functions of different situations are given,including independent/relevant principles,different distributions(folded normal distribution,Rayleigh distribution or Maxwell distribution),different tolerance requirements(dimensional tolerance or geometric tolerance),two/multiple quality characteristics,in order to calculate the cumulative probability interval of 99.73%.Fourth,the application of QFCI on order allocation and selective assembly.This paper brings the whole assembly quality of dimension chain into the research on order allocation and builds a linear programming model of multi-product&multi-assembly-relationship order allocation,with the optimal assembly quality as objective function and with the quality,price,delivery date and service of each product as constraints.At the same time,the changes of objective function value caused by different selective assembly schemes is also considered in the model.For the selective assembly of parts level,considering the complex calculation of selective assembly scheme and the differences among production batches,the production batch is brought into selective assembly and a two stages model is built in accordance with the differences that may exist in the distribution parameters of different batches.At first stage,according to the principle of maximizing QFCI,the parts of each batch would be matched first to determine the group relations among batches and the number of parts in batches.At the second stage,according to the principle of minimizing quality loss,the quality loss functions of line and non-linear dimension chains are respectively built on the basis of the nominal-the-best type,large-the-better type and smaller-the-better type characteristics,to calculate and evaluate the fitness.In addition,genetic algorithm model is established through applying tournament selection&elitist strategy,linear order crossover and swap mutation.This paper applies literature research method,mathematical reasoning method,mathematical programming method and genetic algorithm in researching on the assembly quality and has achieved three innovations:First,systematic research on the calculation method and property of QFCI in normal distribution and SN distribution.It is conducted the research in circumstances that normal distribution,two quality characteristics with different tolerance and distribution features,as well as the dimension chain composed of multiple characteristics.According to the tail behavior of SN distribution,this paper also puts forward the calculation method of PCI and provides the related critical values and analysis methods from the point of production practice.On this basis,the calculation methods of QFCI of two quality characteristics with different distributions are proposed.Second,systematic research on the methods of process control and assembly quality evaluation.For the quality characteristic with dimensional tolerance and geometric tolerance requirements,this paper puts forward the calculation methods of PCI and QFCI under the principle of independent or relevant.Third,take quality fit capability as the evaluation standard for order allocation and production assembly.This paper brings assembly quality into the enterprise's management decision and puts forward the order allocation model based on the quality fit capability,as well as the two stage model of selective assembly based on maximizing quality fit capability and minimizing quality loss.