钢水到达时间扰动下连铸机二冷过程设定值的优化方法

Optimization for set-point values in secondary cooling process with disturbance on job arriving time

中文关键词  连铸设定优化  偏微分代数方程  多目标优化  差分方法

英文关键词  operation optimization of caster  PDE-algebraic equations  multi-objective optimization  difference method

中文摘要

      连铸机通过喷水将钢水冷却凝固为钢坯, 其机理可以由非线性热传导偏微分方程描述. 由于炼钢生产过程受到钢水温度、成分、设备状态等多种因素的影响, 造成连铸机下一炉浇铸钢水的到达时间经常出现大的波动, 给连铸机二次冷却过程的控制设定造成很大困难, 直接影响生产效率和板坯质量. 本文针对钢水到达时间扰动情况下的连铸二冷设定优化问题, 建立了基于冶金准则的连铸运行指标模型和连铸运行约束模型; 通过运行指标归一化处理、基于钢水质量守恒的生产连续性约束处理、凝固传热方程中心差分计算等方法, 在遗传算法框架下实现了连铸二冷设定优化模型的求解. 以某厂的实际数据为基础, 对本文提出方法进行了验证测试, 结果表明了本文方法的可行性和有效性.

英文摘要

      Continuous caster is a machine in which the molten steel is solidified to slabs by spraying on it the cooling water. Nonlinear partial differential equations (PDE) of heat conduction cannot accurately describe the physical process, because of the uncertainties in molten steel temperature and ingredient, and the machine status that causes the variation in arriving time of the next job. Therefore, the optimal set-point values of the caster speed and the cooling water flow-rate are difficult to be determined. In practical industrial plants, those values are given by operators based on the look-up-table method which can not ensure the optimal efficiency and quality of slabs. By considering several operation performance functions and operation constraints, we present an optimal operation set-point model and develop an algorithm based on the genetic algorithm framework using weighted performance functions and central difference method. This model and algorithm are validated in experimental tests with real industrial data. In comparison with the look-up-table method, experimental results show that our method provides better operation set-point values which improve the production efficiency and slab quality.