【摘要】 为了提高气化配煤煤灰流动温度预测的精度和稳定性,提出将遗传算法(GA)与误差反向传播神经网络(BP)相结合的预测方法,采用GA优化BP神经网路的权值和阈值,再用BP算法训练网络,结合仿真实验分析比较了GA-BP网络算法与常规BP神经网络方法的精度和稳定性。结果表明:GA-BP网络改善了BP网络容易陷入局部极小值和收敛速度慢的缺点,经GA优化的BP神经网络预测方法的预测精度高于BP网络算法,将其应用于气化配煤灰熔点预测有效可行。更多还原

【Abstract】 In order to improve prediction accuracy and stability of flow temperatures of ashes from blended coals, prediction method combining genetic algorithm(GA) with BP neural network was put forward in this investigation. Genetic algorithm was applied to optimize the weights and threshold of the BP neural network, followed by network training based on the optimization. Simulation and analysis were performed to compare prediction accuracy and stability of GA-BP and conventional BP methods.It is shown that GA-BP network can reduce shortcomings of local minimum value and slow convergence speed of conventional BP network. Prediction accuracy of BP neural network optimized by GA is higher than that without optimization. It is effective and feasible to predict coal ash fusion characteristics with GA-BP method.更多还原