Fuzzy Clustering And The Application Study In Medium Plate Rolling Performance Forecast

To realize the computer control of the produce process in continuous rolling process depends on following three basic factors: mathematical model, control systems and measure instrument. Mathematical model is the foundation of the calculation that decides the state characteristic of the coming rolling process and make the state characteristic of the rolling process run on the best fixed value to obtain the best control effect. Mathematical model is the key factor in all the elements that influence the computer control effect. Yet it is the important means to improve the performance of the control system. Based on the literature, we establish the forecast mathematical model depend on a set of technical parameters by the fuzzy mathematics correlative knowledge. First of all, the conception of relative membership function, absolute membership function, relative membership and relative state (level) eigenvalue that put forward by the professor of Chen Shou-yu was introduced. Based on these, the Fuzzy C-Means algorithm viz. the iterative self-organizing data used widely now is improved. Relative relation between classified variable eigenvalue and forecast object is received and the fuzzy forecast model is established by apply the model of fuzzy cluster, fuzzy pattern identify, classified (level) variable eigenvalue and probability statistically relative analyse. The above is the foundation of forecast the rolling performance. Secondly, using the above algorithm, we introduce the software' design of 16Mn- steel-medium-board' forecast. The emphasis is the presentation of the blocking programmed idea and realization used in the algorithm.Finally, the analysis of the program result based on the data gathered from 16Mn- steel-medium-board' rolling scene of Chongqing Steel and Iron Group shows the error of the performance forecasting is about five percent.