基于粒度分布评估与优化的制粒过程 PSO-BP 控制算法

针对钢铁烧结中混合料粒度分布无法在线测量、难以实现混合制粒过程优化控制的问题,提出基于 粒度分布评估函数(Evaluation model of granularity distribution, EMGD)的混合制粒优化控制算法. 首先,根据烧结生产历史数据和混合料筛分实验数据建立粒度分布BP神经网络(BP neural network, BPNN)评估模型; 然后,以该模型为目标函数,以制粒过程状态参数的边界为约束条件,采用粒子群算法(Particle swarms optimization, PSO)计算粒度分布优化值; 最后建立基于BPNN的制粒水分设定模型,根据粒度分布优化值和当前配重实现水分优化控制. 仿真实验和工业应用表明评估模型真实反映了粒度分布对料层透气性的影响; PSO-BP粒度分布优 化控制算法对改善透气性、减少燃料损耗、稳顺烧结生产具有重要意义.

PSO-BP Control Algorithm of Granulation Process Based on Evaluation and Optimization of Granularity Distribution

Since granularity distribution of mixing materials in iron sintering can hardly be measured and controlled online, an optimal control algorithm based on evaluation model of granularity distribution (EMGD) has been developed. Firstly, EMGD based on BP neural network (BPNN) has been built by studying data from screening tests and historical state parameters in sintering process. Secondly, by using particle swarms optimization (PSO), optimal granularity distribution can be found from the optimization model restricted by the boundaries of state parameters, whose objective function is EMGD. Finally, according to optimal granularity distribution and online solid flow measurement, humidity setting model based on BPNN is studied to keep the granularity distribution stable and reasonable. Simulation and industrial application have proven that the algorithm makes great sense in permeability improvement, reduction of fuel consumption and stability of sintering process.