基于模糊神经网络的一种漏钢预报方法*

Steel Leakage Alarm Method Based on Fuzzy Neural Network

中文关键词  连铸  漏钢预报  模糊神经网络  模式识别

英文关键词  continuous casting  leakage alarm  fuzzy neural network  pattern recognition

中文摘要

      现有拉漏预报方法经常会发生误报，本文尝试了一种基于神经网络的漏钢预报方法，它果用神经网络进行模糊模式识别和预测拉漏事故. 实验表明该方法能比原有方法更快速准确地检测出铸坯粘结和裂缝等缺陷，可有效预防连铸中的漏钢事故.

英文摘要

      False alarms are often found in the mould monitoring system of continuous casting process， thus a steel leakage alarm method is suggested in this paper. Fuzzy neural networks are used to perform. tem-perature pattern recognition and leakage alarm. Experiments show that it can detect sticking or splfis of slat more quickly and is very effective in avoiding leakage of molten steel.