Study On Intelligent Slag Prediction System For Continuous Casting Ladle

Aimed at the problems of low precision,low automation and time lag in traditional continuous casting.The key to solve slab quality problem is to identify and predict slag time by certain methods,which is also an important prerequisite to realize intelligent development of continuous casting production.Feature analysis of existing continuous casting production data,combined with intelligent algorithms for feature extraction,the main research work completed was as follows:1)Aiming at the problem that the accuracy of steel slag identification is not high according to the observation data directly,the prediction model of continuous casting steel slag was established by using HTM algorithm and the KLIEP algorithm.The HTM algorithm was used to predict the data,and the trend of the data was determined according to the prediction residuals.The KLIEP change point detection algorithm was used to detect the prediction residuals,which will predict the slag in advance and have high accuracy.2)In view of the slow adjustment of specific parameters in the HTM algorithm,and easy to fall into the local optimum,HTM algorithm was optimized.BA is a new metaheuristic optimization algorithm,which is based on the echolocation behavior of bats,combined with the advantages of population and local search algorithm.In order to improve the prediction accuracy and operation speed of HTM,Ba was used to optimize HTM algorithm.3)To solve the problem that KLIEP needs to estimate all samples and the calculation efficiency is not high,an online KLIEP method was proposed.Because the change point detection needs to be calculated in time order,the method was constructed recursively on the basis of batch processing.The new solution can reduce the calculation cost and improve the detection accuracy.The combination of the optimized HTM network and the improved KLIEP algorithm can effectively and accurately extract the slag characteristics,improve the slag prediction accuracy and recognition speed,and provide new ideas and ideas for the future development of continuous casting slag prediction technology.Figure31;Table 6;Reference 47...