Study On Crack Forecast System Of Continuous Casting Slab

The detailed introuduce of crack forecast in the continuous casting production at national and abroad has been presented in this paper, and the mechanism and reasons of crack. According the temperature detection given by multi-level thermocouple on the mold walls, the signs of cracks are detected on-line through the temperature waveform of pattern recognition. The crack signs of continuous casting production process is very weak, which Submerged susceptible to interference signal and data volatile issues, a data preprocessing methods based on the subtraction clustering of Mountain function is presented, the fuzzy space and membership function and the number of fuzzy rules of systems are determined by fuzzy clustering technologies, and the best number of cluster centers is obtained. Subtraction clustering and fuzzy inference system integration of system modeling method is presented, and construction of the adaptive fuzzy nervous system prediction model on symptoms crack.The simulation results show the guarantee on the premise of model accuracy,the efficiency and accuracy of fuzzy rules generated and the degree of convergence model can be increased,and also reduction the complexity of modeling.A continuous cast slab crack forecast system design is introuduced in this paper. Application VB6.0, Matlab6.0 simulation system and database background interface technology to realize the integration of the forecast system simulation and database background. The forecast System offers users two different entry interface of crack signs data and the options of added, queried and so on. A complete forecast, query, analysis and data management system is established.