Processing Quality Forecast And Diagnosis For Gear-box Shell Production Line

With the continuous progress of science and technology, rapid economic development and an increasingly competitive market, product quality become a major problem in China's manufacturing enterprises. Product manufacturing process is an important part of the quality of products and involves the most extensive phase. It is a key to ensure product quality. Process is the basic unit of the manufacturing, product quality is made by manufacturing process. Process is the basic part of the manufacturing process of product parts and components, but also the basic unit of the production process. The effectiveness of process quality and production quality are manifested by the process. The process management can enable the quality of processes and production functions to play better. Process quality is a degree that process quality factor satisfy product manufacturing quality, is a phrase to describe machining process is good or not. Process control is an activity of analyzing, controlling and managing the elements that affect the level of process quality. In the manufacturing process, the movement of process quality factors and their impact on the role of product quality is an extremely complicated process. Process control is the process of discovery and use of this extremely complex process of internal law so that they will reach to a steady purpose of the production of qualified products.A basic problem of process quality control is prediction of product quality, only forecasting the future product quality could pre-adjust production process before product occur quality problems and reach the purposes of product quality control. And diagnosis of product quality is another basic problem. Effective quality control process begins with effective diagnosis, if there is no "diagnosis", there is no basis of "control". Process Quality Diagnosis is in order to improve response to having an examination of process quality, evaluating the process capability, mastering information of process quality and finding major impacts of process quality. Quality diagnosis as an important part of the process is responsible for the selection of products which will be put into use soon. It is important to utilizing the link quality information and feeding back into the production process. That is providing a basis adjustment to production process. The main purpose of the diagnosis process is finding the causes of bad quality and having an objective analysis of them.On the basis of a full analysis of data mining technology and the characteristics of production process, this paper propose a concept that using data mining technology provides solutions for the quality prediction of manufacturing process and a method of process quality diagnosis. In this paper, there are several works done in the following areas:First of all, introduce the basic concepts of the forecast, analyze the model of exponential smoothing and gray, and apply the theories in the actual production. Analyze the long-term changes of size deviation in the 25th workshop in exponential smoothing model and have a result that the deviation is getting smaller and smaller, the quality of machined products is getting higher and higher. Using gray prediction model analyze short-term fluctuations changes of size deviation, and have a result that the size deviation is in a trend of overall decrease but up and down fluctuations. After integrating the two results, make a conclusion that the future size deviation will be fluctuating up and down constantly, the overall trend is decreasing and the quality of products be improved continuously.Secondly, give a detailed introduction of association rules, discuss several issues in the type of quantitative association rule mining, map the quantitative data base into a Boolean database, and realize the mining of a quantitative type of association rules. Through the association rules analysis of the 24th and 25th workshop station, follow-up the changing trend of product quality between the two workshop stations, and judge the 24th workshop have a good impact on the product quality of 25th workshop. The process between the two workshops is well and should be maintained.Finally, introduce the basic theory of process quality diagnosis, analyze the fluctuations reasons of product quality, and explore the mapping relationship between the flow information of quality data, attributes in the management of the production process, and look for the method of process quality diagnosis in manufacturing enterprises among a large number of historical data recorded in production processing. By analyzing the history production data of a gear box plant, put forward a specific idea to diagnose the process quality of the gear box cases, apply it to the actual production, and realize the process quality diagnosis of the gear box cases. Through analyzing, diagnose that operator and testing methods are the incentives to make machining quality better in 25th workshop in July 29, 2006.In summary, by analysis of time series and analyzing the machining quality of the gear box cases in exponential smoothing model and the gray prediction model, predict that product quality will be improved in future. By mining association rules of the machining processes of the gearbox cases and analyzing the impact on product quality in different workshops, get a conclusion that the capacity of machining processing of the gear box cases will increase. At the same time, by analyzing different factors of the same workshop, diagnose the operator and the test methods are the enabling factor to improve product. Combine both to achieve the diagnosis of process quality.