Study On Immune Mechanism Based Intelligent Diagnosis Method Of Cement Manufacturing Technology

Cementindustry is the pillar industry in our country, which is a typical modern process industry. New dry process kiln stands for the most advanced level cement production technique in the world, whose production process is a complex physical and chemical reaction procedure. So it's important to guarantee running securely and reliably with implementation of monitoring and fault diagnosis towards cement production process.In order to improve the accuracy, real-time capability, robustness and adaptability of the fault diagnosis, it's imperative to perfect existing diagnostic methods and develop intelligent diagnosis based on biological information processing mechanism. Research on monitor and diagnosis on technique fault during cement production process is carried out, with experience of biological immune system and combination of conventional intelligent diagnostic techniques. It's beneficial to take real-time control and adjustment towards technical parameters during cement production process and guarantee the production process safe, stable, efficient, low cost and high quality. Therefore, this paper takes faults of cement kiln system technique as main object of study, and focuses on the research of intelligent fault diagnosis method based on immune mechanism. Main research work is described as follows:(1) Research contents, background and significance of selecting the topic are firstly introduced. With the analysis of failure characteristics of the cement industry and main limitations of existing intelligent fault diagnosis, conventional intelligent fault diagnosis methods and artificial immune system are elaborated.(2) Comprehensive introduction to the composition of biological immune system, functions and basic features is given; and related immune mechanisms such as immune recognition, clonal selection, immune response and immune memory are discussed. This paper describes the common artificial immune system model:shape space model, the binary model, immune cell model and the basic framework for artificial immune algorithm, as well as artificial immune algorithms like negative selection algorithm, clonal selection algorithm, immune algorithm based on immune network; and analyzes application prospects of immune intelligent fault diagnosis method from the perspective of information processing mechanism.(3) One feature extraction algorithm based on fractal dimension and immune clonal selection is proposed and implemented. Compared with FDR algorithm-the conventional method for feature selection with fractal dimension, which takes higher cost in calculation process, the new algorithm has the classifier with lower computational complexity and higher accuracy. Finally, simulation towards the algorithm is carried out, whose results indicate has good performance in feather selection of high-dimensional data set.(4) On the basis of the analysis towards firing system and parameters related with quality and safety of cement production, fractal dimension and immune clonal selection algorithm are applied to extract the feature towards relevant production process among 46 selected parameters and 6 categories of technical failures, and primary recognition is achieved among the 6 categories of technique failures through immune clonal selection algorithm.(5) Two new hybrid intelligent diagnosis algorithms:particle swarm-clonal selection clustering algorithm and immune particle clustering algorithm based on immune memory and vaccination Information on regulatory mechanisms are proposed, with the analysis on immune clonal clustering algorithm and particle swarm algorithm and combination between the two algorithms. Simulations emulation carried out under various conditions are presented to demonstrate that the new algorithms have faster convergence speed and higher recognition accuracy.