Research On Theory And Method Of Knowledge-Driven Product Quality Problem-Solving

As people's living standards and purchasing power continue to increase,consumption upgrades have become an important trend in the consumer market.The shift from price-oriented to value-oriented and the popularity of personalized consumption brought about by the high level of product quality and variety,which challenged the manufacturing system and quality management level of manufacturing enterprises.In addition,"Made in China 2025" clearly requires the manufacturing industry to strengthen the construction of quality brands,including the promotion of advanced quality management techniques and methods,accelerate the improvement of product quality,improve the quality supervision system,consolidate the foundation of quality development and promote the construction of manufactur:ing brands.Therefore,manufacturers need to continuously improve the management of enterprises,especially the quality management level,to improve the quality of products and services to meet the needs of the market.With the advent and continuous development of the knowledge economy,knowledge has gradually become the core competitiveness of enterprises.The importance of knowledge in quality management is also receiving more and more attention.On the one hand,the relevant quality management system puts forward requirements for knowledge management in quality management.On the other hand,scholars have also studied the relationship between quality management and knowledge management from various angles.However,existing researches rarely pay attention to the application of knowledge discovery technology in quality problem-solving.To this end,based on the theory of quality management and knowledge management,this study takes quality problem solving as the research object and uses the techniques and methods in knowledge discovery as a means to study how to mine quality problem knowledge from quality problem-solving data.Then feedback the knowledge to the quality problem-solving process to improve the efficiency and effectiveness of quality problem solving,and thus promote the improvement of product quality.The main contents of this research include:First,a new research model of knowledge-driven quality problem-solving(KDQPS)is proposed.Related concepts and process of quality problem-solving are analyzed and discussed.Then construct the system structural model of quality problem-solving from the perspective of the system,and deeply study the elements and relationships in the model.In combination with the knowledge classification of the Organization for Economic Co-operation and Development(OECD),the relationship between quality problems and problem carriers is expressed as know-what knowledge,causality is expressed as know-why knowledge,and the relationship between problem and solution is expressed as know-how knowledge.On this basis,the theoretical framework and application framework of KDQPS are proposed,and the guarantee conditions for the implementation of the method are studied.Second,the mining of know-what knowledge in quality problem-solving.Based on KDQPS framework,acquisition method of dependency relationship of problem carrier and quality problem,which is know-what knowledge,is analyzed.According to the characteristics of product quality problem solving,the component-failure mode matrix is proposed to represent know-what knowledge.Because the association between components and failure modes is implicit in the quality problem text,and the failure modes are described differently,the standard failure mode set construction method based on Apriori algorithm and WordNet is studied,and the component-failure matrix mining algorithm is designed.Finally,an example study based on car seat module data is carried out to verify the effectiveness of the method.Third,the mining of know-why knowledge in quality problem-solving.Based on the KDQPS framework,know-why knowledge required for the cause analysis in the quality problem solving is analyzed.A method of using data mining technology to mine causality,that is,know-why knowledge,from the problem and cause data,and using the digital fishbone diagram to represent the causal relationship is proposed.The method clusters the original problems and causes separately,and classifies the cause data into the "major factors" of the fishbone diagram,such as "Man","Machine","Material","Method," and "Environment".Based on the clustering results and the classification results,the abstract and detailed digital fishbone diagram of the class problem are obtained.Finally,the effectiveness of the method is verified by an example study on a real dataset of an automobile company.Fourth,the mining of know-how knowledge in quality problem-solving.Based on the KDQPS framework,the author deeply studied the association mining of problem and solution,which is know-how knowledge,from a large number of quality problem-solving texts.At the same time,the method of discovery and recommendation of typical immediate measures and long term solutions is proposed.Taking the know-how knowledge mining of problems and immediate measures as an example,according to the characteristics of the immediate measures text,a two-stage clustering method based on verb clustering and noun clustering is studied to extract typical immediate measures.Then build the problem class-immediate measure class knowledge model to represent the know-how.Finally,a case study is carried out to analyze the reliability and effectiveness of the method,and to demonstrate the process of knowledge recommendation of typical immediate measures.Fifth,the application of knowledge-driven quality problem-solving in automobile manufacturing.On the basis of the above research,the paper studies the practical application of KDQPS and know-what,know-why and know-how knowledge mining methods in the automotive industry.The decision support prototype system in automobile quality management is constructed,which is called knowledge-driven quality problem-solving system(KDQPSS).Finally,the scientific and effective of the above method are validated from the design,implementation,application,and evaluation of the model system.This resea rch has important theoretical significance and application value.The research in this paper is an important embodiment of advanced quality management techniques and methods.From a macro perspective,the research in this paper deepens and advances the process of Industry 4.0 and Made in China 2025.At the micro level,the KDQPS framework provides general ideas and methods for managers to make full use of quality data to generate quality problem-solving knowledge to improve problem-solving efficiency.The proposed method of mining quality problems and problem carrier association knowledge from text data can provide statistical information for quality problems on the one hand,and the corresponding component and failure mode data can be obtained directly from the matrix when creating design failure mode and effect analysis(DFMEA)of new products.A framework and method for mining causal associations and automatically creating fishbone diagrams based on clustering and classification in data mining is proposed,which greatly improves the traditional analysis method based on personal experience.At the same time,the process of creating a traditional fishbone diagram is transformed into an automatically generated process.The proposed method of mining the relationship between the problem and the solution from the quality text and recommending the appropriate solution knowledge to the typical problem can provide a basis for the problem-solving team to find a solution to the problem.