11基于混沌神经元理论的振动式大包下渣检测系统5-12？.docVIP

基于混沌神经元理论的振动式大包下渣检测系统 田陆1,2，黄郁君1，2，黄琪才1，2 1 衡阳镭目科技有限责任公司，湖南 衡阳 421001 2 镭目公司北京研究所 北京 100080 摘 要：为了提高铸坯质量和连铸作业率, 在钢水浇铸过程中,必须严格控制下渣。在对国内外现有的各种下渣检测方法的广泛调研、资料分析和总结的基础上，本文提出了对水口操作臂振动信号进行混沌神经元分析，以此来识别钢水浇注状态的下渣检测系统。文章对振动信号的采集与处理以及钢渣特征参数的提取与识别方法进行了深入研究，并详细阐述了神经网络技术应用于下渣检测的基本原理，简要介绍了该系统的总体框架与功能设计。系统在湘钢等钢厂的使用结果证明，该系统满足实时性的要求，能够显著提高钢水的洁净度和连铸作业率，为钢厂创造效益。 关键词：混沌理论；神经网络；钢渣检测；小波分析；振动；信号处理 Slag Detection System Based on Vibration using Chaos and nerve cell Theory TIAN Lu1,2，, HUANG Qi-cai1,2 1. Ramon Science and Technology Co.,Ltd HengYang HuNan 421001 2. Beijing Research Institute of Ramon. Beijing 100080 Abstract: At the process of molten steel casting, outflow slag must be detected to improve the quality of steel product and continuous casting efficiency. After lots of investigation of the internal and external slag detection technology, data analysis and data summarization, the dissertation brings forward the slag detection system of the casting state of steel by analyzing the vibration signal of nozzle operation arm using chaos and nerve cell method. The collection and process of molten steel vibration signal and the method of slag characteristic parameter identification has been researched deeply, basic principles of the application of neural network in casting state recognition are expounded. And the frame and function design of slag detection system are all briefly introduced. The results testing in Xiang Steelworks proved the effectiveness of the slag detection system, which can satisfy the demand for real-time, improve pure measure of molten steel and continuous casting efficiency and create benefit for the steelworks evidently. Key Words：Mixed method theory; Neural Network; Slag Detection; Wavelet Analysis; vibration; Signal Processing 1 引言 连铸工艺的生产过程中，钢水从钢包流入中间包，再经滑动水口进入结晶器，冷却后凝固成各种截面的铸坯。渣层对钢包及中包钢水有保温防氧化作用，然而当中包渣层过厚则会导致钢水污染，增加中包耐火材料侵蚀并增加中包残余渣层厚度。因此控制钢包中的钢渣流动对提高铸坯质量、增加连铸批次有着非常重要的意义。 目前国内外已有多种对钢包下渣进行自动检测的方法，如红外线检测法、电磁检测法和超声波检测法等，并均在不同程度上获得成功[1]。但是由于连铸过程的特殊性，再加上恶劣的浇