信息融合技术在煤矿安全预测中的应用

摘要

研究煤矿井下安全状况预测问题.针对现有预测模型因采用的信息融合算法单一导致的精度低的缺点,提出一种更适合矿井的基于模糊神经网络的信息融合预测模型;并重点针对现有的模型训练算法速度慢、难以适应井下要求的缺点,提出一种改进的LMBP算法,通过引入判别因子大幅提高模型的训练速度.仿真结果表明,预测模型在训练速度上比传统BP算法和LMBP算法分别提升了6倍和3.2倍,而且预测精度能够满足煤矿实际要求.%Research the prediction issue in coal mine safety, the paper proposed an information fusion prediction model which is more suitable for coal mine and based on improved fuzzy neural network. To solve the problem that BP algorithm has low speed and unsuitable for coal mine, an improved LMBP algorithm was proposed to improve the speed of prediction by introducing judgment factor. Simulation result shows that the prediction model is 6 times faster than BP and 3.2 times faster than LMBP in training speeds. The model can satisfy the actual demand in coal mine.