摘要
针对矿井涌水系统的复杂性和随机性,提出采用神经网络修正灰色残差模型对矿井涌水量进行预测,既利用GM(1,1)模型能较好预测涌水量发展趋势的特点,又利用神经网络对于复杂非线性系统的优越性,保证了模型的精度,克服了单个模型所存在的不足。结果表明,该模型方法在矿井涌水量的预测中是可行的。
In view of the complexity and randomicity of mine inflow,a BP neural network-based corrected residual gray model is proposed to predict the inflow rate of mine water.GM(1,1) model can forecast developing trend of water inflow,and neural network has superiority over complex nonlinear system.The corrected residual GM(1,1) model can overcome the shortage of residual GM(1,1) model,so the forecast accuracy of mine water inflow can be assured.The results of a case study show that the corrected residual GM(1,1) model is feasible in the prediction of mine water inflow.
出处
《矿业研究与开发》
CAS
北大核心
2011年第2期73-75,共3页
Mining Research and Development
基金
国家自然科学基金资助项目(50774092)
全国优秀博士学位论文专项资金资助项目(200449)