液压AGC系统机电液耦合非线性动力学行为的理论及关键技术基础

Hydraulic automatic gauge control (AGC) system is a core control system to guarantee the accuracy of the plate thickness. It possesses some features such as highly nonlinearity, time variation, strong coupling and large lag. Once a fault occurs, there will be a serious and harmful consequence. Moreover, it is difficult to determine the incentive. Taking the characteristics of AGC system into account, the project is aimed at improving the operational stability in rolling process and reducing the fault rate of equipments. Then, the novel multi-level hybrid modeling method is researched and a nonlinear dynamics model of system is established. The bifurcation and chaotic behavior of system under different parametric excitation are explored, moreover, the fault mechanism and incentives are revealed. The new methods pointed at chaos identification and control are studied. The health assessment methods are researched, which are deeply intercrossed and fused with mechanism, data, and knowledge. Finally, a platform with the function of nonlinear dynamics behavior simulation and health assessment prediction is exploited. Based on the above, we try to solve the existing difficult problems, for instance, modeling AGC system, determining fault incentive, and exploring fault mechanism. The research will not only establish the theoretical foundation for nonlinear dynamics performance evaluation and optimization design of AGC system, but also enrich the theoretical system of fault diagnosis and prediction. Furthermore, the achievements will provide the original key generic technologies for the research and technology promotion of rolling gauge control system with autonomous intellectual property, which can solve the urgent problems in rolling process of tandem cold mill.

液压AGC系统是保证板厚精度的核心控制系统，存在高度非线性、时变性、强耦合和大滞后等特性，故障危害严重且诱因不易确定。本项目针对液压AGC系统的特点，以提高轧制过程运行的稳定性、减少设备的故障发生率为目标，研究新颖实用的多级混合建模方法并建立系统的非线性动力学模型；探究不同参数激励下系统的分岔及混沌行为并揭示故障的形成机理和诱因；研究更具针对性的混沌识别与控制新方法；研究机理、数据和知识深度交叉融合的健康状态评估方法；开发非线性动力学行为及健康状态评估预报仿真验证平台，以解决制约液压AGC系统难以精确建模、故障诱因难以确定、故障机理难以探究等难题。研究成果不仅可以为液压AGC系统机电液耦合非线性动力学性能评价及系统优化设计奠定理论基础，而且可以丰富故障诊断与预报的理论体系，为国产自主知识产权的轧制厚控系统的研发和技术推广提供原创性关键共性技术，解决冷连轧制生产过程的急需。

液压AGC 系统是保证板厚精度的核心控制系统，它存在高度非线性、时变性、强耦合和大滞后等特性，发生故障危害严重且诱因不易确定。本项目在密切跟踪国内外最新研究动态基础上，并在深入调研液压AGC系统在实际应用存在问题的前提下，从多级混合建模理论及模型参数修正算法、非线性动力学行为研究及诱因分析、振动信号预处理方法、故障特征提取算法、故障诊断方法、健康状态预报评估理论、以及故障模拟试验台研发等方面进行了系统全面深入的研究。. 在液压AGC系统多级混合建模理论及模型研究方面，运用机理、数据和知识深度交叉融合的多级混合建模方法，建立了轧机辊系两自由度垂直振动非线性动力学模型；在液压AGC系统非线性动力学行为研究方面，探索了非线性液压弹簧力和非线性摩擦力等非线性因素对系统运动特征的影响规律，揭示了系统内在的分岔现象及典型非线性动力学行为；在液压AGC系统振动信号预处理方法方面，提出了形态学差值算子和差分熵理论相结合、局部均值分解和迭代自适应多尺度形态学分析等特征提取方法；在液压AGC系统健康状态预报评估理论方面，提出基于形态差值算子和形态指数相结合、局部均值分解和广义形态分形维数相结合等健康状态评估方法；并且，针对液压AGC系统状态监测与故障诊断理论方法验证问题，创成了一套能够模拟电液伺服系统多种故障的试验台。. 该项目在液压AGC系统多级混合建模、非线性动力学行为研究、故障机理诱因分析、故障特征提取算法、健康状态评估理论、以及故障模拟实验装置研制等方面均取得了一系列新成果、新技术和新理论，具有重要的工程实用价值。