| Complex industrial production is an important pillar industry of modern industrialized countries. With the development of modern advances in manufacturing technology and production, petrochemical and other complex industry production processes become more and more large-scale and complex. However, it also will bring the challenge of the complex industry production quality and production safety. Therefore, in order to achieve a stable, efficient and safe production process industry, by getting massive data of industrial production on energy, materials and operations, to monitor and guide the production process has very important significance. Process capability analysis is one of the core methods of modern enterprise quality management in the production process, mainly using statistical theory of the production process to monitor the stability of the data, so as to continuously improve the quality of the product. In this paper, process capability analysis and bayesian probability theory are the theoretical basis. Through the systematic review of the literature and current cutting-edge technology-related research, the complex industrial process capability analysis and diagnosis method are conducted in depth research and application. The main research work includes the following aspects:1. Fuzzy process capability analysis based on kernel function and fuzzy evaluation. Cubic spline interpolation is used to dynamically determine the fuzzy specification limits. Fuzzy process capability is calculated combined with kernel function, which can effectively solve the complex industrial data volatile issues. The fuzzy process capability evaluation is also proposed based on the fuzzy number comparison of satisfaction. And the specific applications demonstrate effectiveness and adaptability of the method in complex industrial process capability analysis.2. Process capability analysis and evaluation of process device online. The methods first determine the main parameters of the device through fuzzy interpretive structure model, and then we use analytic hierarchy process method to make fusion of the main process capability indices of the device parameters, and give the overall process evaluation of the device. Additionally, ethylene cracker verified the applicability of the method.3. The directed graph research and bayesian probability are applied to diagnose of process capability. Through the established directed graph model of the system, combined with bayesian probability and directed graph research, we could get the different sequences of the parameter path which are affected by the insufficient process capability. The experimental results of TE model showed that this method plays an important role in the guide of the complex industrial production.4. Design and develop the software for the analysis and diagnose of the proposed methods.Finally, we make the conclusion and the review of this paper, and take the outlook of the process capability research in the future. |
PDF Full Text Request