文献类型：期刊文献

中文题名：基于扰动补偿的结晶器液位模糊控制算法

英文题名：Mold level fuzzy control algorithm based on disturbance compensation

作者：孙丽华[1];齐淏天[1];吕立华[2];姜庆超[1]

机构：[1]华东理工大学能源化工过程智能制造教育部重点实验室,上海200237;[2]宝山钢铁股份有限公司中央研究院智能所,上海201900

年份：2024

卷号：48

期号：6

起止页码：66

中文期刊名：冶金自动化

外文期刊名：Metallurgical Industry Automation

基金：国家自然科学基金优秀青年科学基金项目(62322309)。

语种：中文

中文关键词：结晶器液位;特征选择;因果分析;扰动补偿;模糊控制

外文关键词：mold level;feature selection;causal analysis;disturbance compensation;fuzzy control

摘要：结晶器液位控制系统是连铸自动控制系统的核心组成部分,其液位稳定性直接关系到最终产品的质量。液位波动通常由多种干扰源引起,而干扰的有效补偿问题尚未得到充分解决。为此,本文提出了一种基于扰动补偿的结晶器液位模糊控制算法。首先,采用数据驱动的特征选择方法,结合机理分析,选取与结晶器液位密切相关的重要变量;然后,通过偏交叉映射方法分析液位波动的直接因果关系,以识别主要干扰因素。在此基础上,本文进一步开发了一种融合长短期记忆网络(long short-term memory, LSTM)的模糊控制方法。通过精确估计和补偿干扰,该方法显著提高了系统的控制精度。仿真结果表明,所提方法相较于传统控制方法,超调量由40.2%缩小至15.4%,恢复时间从7.6 s减少到4.2 s,同时能有效抑制环境干扰,保持结晶器液位稳定,提升了系统的稳定性,对于提高产品质量和生产效率具有重要意义。
The mold level control system is a critical component of the continuous casting automation system, with level stability directly impacting the quality of the final product.Fluctuations in the mold level are often induced by multiple sources of disturbance, yet the effective compensation for these disturbances remains inadequately addressed.To address this issue, this paper proposed a disturbance compensation-based fuzzy control algorithm for mold level control.Initially, a data-driven feature selection method is employed, in conjunction with mechanism analysis, to identify key variables closely related to the mold level.Subsequently, the partial cross mapping method is utilized to analyze the direct causal relationship in level fluctuations, pinpointing major sources of disturbance.Based on this analysis, a fuzzy control method integrated with a long short-term memory(LSTM) network is further developed.By accurately estimating and compensating for disturbances, this approach significantly enhances control precision.Simulation results demonstrate that, compared to traditional methods, the proposed method reduces overshoot from 40.2% to 15.4% and decreases recovery time from 7.6 s to 4. 2 s,while effectively suppressing environmental disturbances to maintain stable mold levels. This improvement in system stability contributes to enhanced product quality and production efficiency.