论文部分内容阅读
为保证具有强非线性、建模困难和具有周期性扰动连续浇注过程中钢水液位满足工艺和生产的要求,从而实现连续浇注过程中钢水液位的高品质控制,提出了一种精细迭代控制策略来进一步削弱连续浇注过程中钢水液位的膨胀扰动带来的不利影响。该策略是一种包含P型学习律的迭代学习算法,这种P型学习律除引进遗忘因子和开关装置外,还特别利用了前两次迭代过程中的误差信息(即控制律为PID+遗忘因子+开关函数+误差精细信息的有机组合),以使误差信息精细化,从而进一步改善迭代学习控制的效果;研究表明在系统同时具有模型不确定性,周期性膨胀扰动,可测噪声干扰与初始状态误差情况下能保证系统的输入信号误差、状态误差和输出误差的最终有界性,实现了钢水液位的高品质控制;计算机仿真进一步验证了所提方案的正确性和可行性。
In order to ensure high nonlinearity, modeling difficulty and cyclical turbulence during the continuous casting process, the molten steel level meets the requirements of technology and production, so as to achieve high quality control of molten steel level in continuous casting process. A fine Iteration control strategy to further weaken the continuous pouring process molten steel level expansion disturbance adverse effects. This strategy is an iterative learning algorithm which contains P-type learning law. In addition to the introduction of forgetting factor and switching device, this P-learning algorithm makes special use of the error information in the first two iterations (ie, the control law is PID + forgetting Factor + switch function + error fine information of the organic combination), so that the error information refinement, thereby further improving the effect of iterative learning control; studies have shown that the system also has the model uncertainty, periodic dilation disturbance, measurable noise interference and Under the condition of initial state error, the final boundedness of system error, state error and output error can be ensured, and the high quality control of molten steel level can be achieved. Computer simulation further verifies the correctness and feasibility of the proposed scheme.