数据预处理条件下的神经网络开采沉陷预计模型研究

开采沉陷预计是矿山开采研究领域的核心和难点之一,它对开采沉陷的理论研究和生产实践都有重要的意义。本文以进行沉陷预计为目的,提出建立基于各影响因素数据序列思想,其主要步骤为首先建立影响因素序列;其次利用灰色理论对建立的数据序列进行预处理;最后构建基于数据预处理条件下的BP神经网络预测模型(AGO-BP模型),以用于矿区沉陷预报。通过对几组实例本文对建立的模型进行了多次反复预测实验,预测结果证实该模型预测精度较高,在工程上具有一定的应用价值。

Research on mine subsided prediction model under data preprocessing based on artificial neural network

Subsided prediction is one of the key and difficult research fields in mine.Based on analyzing of some subsidence datum,the paper presented the idea of setting up sequence taking all influenced factors into account based on new artificial neural network model.Then it preprocessed the sequences on the basis of grey theory and constructed a BP neural network model.Several groups of observation stations data was used as learning and training samples to train and test the model.The results indicated that it is comparatively precise to calculate the subsidence of ground surface and the new model is applicable in deformation's prediction.