预测钢坯温度分布的时间顺序粒子群优化算法

Time Series-based Particle Swarm Optimization Algorithm for Predicting Billet Temperature Profiles

以加热炉内钢坯的一维模型为研究对象,采用基于有限差分法的瞬态导热正问题以及基于时间顺序的粒子群优化算法来构建一维瞬态导热反问题的数学模型。首先以钢坯离散点处温度测量值作为已知条件,并将其与导热正问题计算的温度值的均方差作为目标函数,通过粒子群优化算法反演得到边界的热流密度值。其次将反演的热流密度值代入导热正问题中求得离散点处的温度模拟值。最后使用Python语言编写通用算法程序。计算结果表明:离散点处的温度模拟值与测量值具有较好的一致性,说明所构建的导热反问题模型可以较好地反演出钢坯表面的边界条件,而且当存在随机误差时,仍能精确地预测钢坯温度分布。后面可以进一步结合实际工程,将该模型推广到三维钢坯温度预测系统中。

The one-dimensional model of the billet in the heating furnace is taken as the research object.The forward problem of transient heat conduction based on the finite difference method and the particle swarm optimization algorithm based on the time sequence are used to construct the mathematical model of the inverse problem of one-dimensional transient heat conduction.First,the measured temperature at the discrete point of the billet is taken as the given condition,and the mean square deviation between the measured temperature and the temperature calculated by the forward heat conduction problem is taken as the objective function.The heat flux at the boundary is retrieved by particle swarm optimization algorithm,Secondly,the inverse heat flux value is brought into the forward heat conduction problem to obtain the values of simulated temperature at the discrete points.Finally,the general algorithm program is coded by Python.The results show that the simulated temperature at the discrete points is in good agreement with the measured values,which indicates that the inverse heat conduction problem model constructed can better reverse the boundary conditions of the billet surface.The billet temperature profiles can still be predicted actually when random errors exist.Later,the model can be further extended to the three-dimensional billet temperature prediction system in practical projects.