摘要
本文分析了铝电解过程的大量实验测温数据,用三角形隶属函数来模糊化输入和输出,得到了根据热电偶升温速度判断体系温度的30条模糊推理规则用"Product-Sum-Gravity"模糊推理方法,根据热点偶两个阶段的升温速率得到体系的温度值,从而实现铝电解温度的模糊模式识别,通过验证,总体识别的平均相对误差为0.37%,表明该模型可以较准确地识别出铝电解过程的温度。
In this paper, the fuzzy pattern recognition of aluminum bath temperature was studied based on a lot of temperature measurements in aluminum production. First of all, the input and output of the original data were fuzzified by triangle functions, then 30 fuzzy rules of determining the electrolysis temperature were got. The 'Product-Sum-Gravity' method was used to gain the temperature. Experimental results showed the average relative error was less than 0.37% , which indicated the model could recognize the temperature of electrolyte well.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2001年第4期466-469,共4页
Pattern Recognition and Artificial Intelligence
基金
国家自然科学基金
关键词
模糊模式识别
铝电解
温度测量
热电偶
Fuzzy Pattern Recognition, Aluminum Electrolysis, Temperature Measurement
