Research Of The Roller Eccentricity Compensation

For the strip steel, the shape dimensions including thickness, width, board line, board convexity, planform and so on. In all dimensions of precision indicators, the thickness is a measure of controlling sheet and strip the most important indicators of the quality of domestic and international metallurgical industry has become the focus of widespread attention.With the increasing requirements of the strip quality more and more accurate, the effect of roll eccentricity can not be ignored and the problem of roll eccentricity has become a hotspot problem. The domestic and foreign scholars from the 20th century seven, the 80's start to study the roller eccentric question, there are many ways for the roller compensation control. With the intelligent control and some new methods of signal analysis to occur recent years, developed some new methods of roll eccentricity compensation. Although the compensation of roll eccentricity in our country has made a enormous progress recent years, Compared with the developed countries is large gap. And there still are many questions that it is both in practical appliance and theory to be solved.At present the domestic strap production generally uses the AGC system. Most commonly used is thickness counts (GM-AGC) type AGC. GM-AGC on the eccentricity fluctuations in production causes reverse movement of roll gap, leads to the more strip thickness error of increasing exports. Therefore it should compensate the error caused by roll gap fluctuations. This paper made a research into eccentricity compensation which is in the preliminary stage in china.First, based on the of understanding and grasping the principle of gauge control system, this paper introduces some common AGC model and Particularly describes the various parts of GM-AGC system, and gain the transfer function of each part of GM-AGC system.Second, it analyses the work principle of gauge hydragauge control system and the problem of roller, comprehensive conclude the peculiarity of roller eccentricity by the numbers,hold its character, point out its influence for the gap,rolling force and gauge of steels.Third, it describes the definitions and causes of the eccentricity, compared to previous using of fast Fourier transform(FFT) to gain the roll eccentricity signals, in this paper, it describes using the neural network to process the data and gain the mathematical model of the roll eccentricity and design a neural network to get roll eccentricity signal (proves the stability of the neural network). More convenient using of fast Fourier transform to get the first harmonic of the mathematical model, and the error is relatively large. But it can get higher harmonics of the mathematical model by using the neural networks, and it is more accurate to get the mathematic model than using the traditional FFT arithmetic.Finally, because it can not compensate the roll eccentricity caused by the period signal using the PID control system, and due to the role of proportional and integral role in the design restrictions, it has a poor tracking performance, lead to poor steady state. Based on the principle of repetitive control compensation this paper designs the iterative PID control system, which greatly improves the tracking accuracy and improves system quality. Finally, PID parameters are optimized by using the intelligent algorithm, it improves the control precision, promotes the speed of the system response time.