节点文献
电弧炉及其电极升降智能控制系统的研究
The Research on Intelligence Control System of the Arc Furnace and Electrode’s Rise and Fall
【作者】 张学辉;
【导师】 李湘闽;
【作者基本信息】 江西理工大学 , 控制理论与控制工程, 2007, 硕士
【摘要】 本文以某钢铁公司的40t电弧炉为研究对象,同时进行50t精练炉控制系统的研究与开发。设计一种综合的、集成的智能控制系统解决电弧炉炼钢复杂的控制问题,以弥补单一控制方法的不足。电弧炉炼钢综合的、集成的智能控制系统对于保证电弧炉优质高产、安全运行、节能降耗具有重要意义和实用价值。采用研华公司的高速数据采集卡设计了一个高速、准确的电参数数据采集系统,使该采集系统可以实时的采集大量的电参数(电弧电压、电弧电流和有效功率等)。使用这些数据在离线的情况下建立一个电弧炉神经网络预估模型。将建立的BP神经网络炉子预估模型在线投入,使用改进的BP算法与遗传算法相结合的方法对神经网络预估模型的结构与权值进行训练仿真,并不断进行在线自调整,使预估模型对实际的电弧炉对象实现最佳逼近。特定的预估补偿程序利用预估结果,对常规控制的电极调节器输出进行优化补偿,用于实时控制,以获得期望的控制效果。同时,为确保智能电极调节器发生故障时冶炼设备仍能正常工作,保证生产的连续进行,采用PID控制算法设计了PLC常规电极调节器。
【Abstract】 The article takes some iron and steel company’s 40t arc furnace as search object, simultaneously carries on the 50t concise stove control system, Designs a kind of comprehensive and intelligent control system to solve arc furnace steel-making complex control question and makes up for the insufficiency of single control method. This intelligent control system has the vital significance and practical value regarding the arc furnace good quality and high output, the safe operation, the energy conservation fall consumes.We use productions of Advances Company to design high speed and accurate electrical parameter data acquisition system, and the system to be allowed gathering massive real-time electrical parameters (arc voltage, arc current and effective power). And use these data to establish a mathematical model with nerve network (NN) in the off-line situation to estimate the electric arc furnace. Then ,let this model to work on-line, at the same tine, use improvement BP and GA algorithm to optimize NN model’s structure and weights, and constantly adjust the model on-line, cause the estimating model to approach electric arc furnace in realization. The specific compensating programs use estimated result to compensate the output of the electrode regulator with conventional control method, obtains the expected control effect. For guaranteed the equipment still to be able to finish the normal work when the intelligent electrode regulator broke down, we use PID algorithm to design the conventional electrode regulator with PLC.
【Key words】 Rise and Fall of Electrode; Neural Networks (NN); Genetic Algorithm (GA); Matlab Simulation; PID Algorithm; PLC;
- 【网络出版投稿人】 江西理工大学 【网络出版年期】2008年 05期
- 【分类号】TM924.4;TP273.5
- 【被引频次】9
- 【下载频次】514