中文摘要：

盛大铁矿巷道变形破坏事故频发,严重影响矿区安全生产,以神经网络理论为基础,通过对盛大铁矿巷道进行变形监测,得到巷道变形规律,为巷道支护优化及变形预测神经网络模型的实现提供了数据支持。利用MATLAB软件建立神经网络模型,以盛大铁矿46条支护稳定巷道为学习和检验样本,建立巷道支护优选及变形预测神经网络模型,通过测试样本对神经网络模型检测,模型成功实现了支护方式优选及变形预测,为矿山支护设计及变形预测提供了科学的技术指导。

英文摘要：

Frequent roadway deformation and failure in Shengda iron ore seriously affected the safe production of mine. Therefore, based on the neural network theory, through monitoring the roadway deformation in Shengda iron ore, the law of roadway deformation was obtained, which provided data support for the optimization of roadway sup- port and realization of the neural network model. Taking 46 roadways with successful support in Shengda iron ore as the study and test samples, the neural network model of road- way support optimization and deformation prediction was es- tablished based on MATLAB software. Then, the neural network model was detected by testing samples. The results showed that the model successfully realized support optimi- zation and deformation prediction, providing technical guid- ance for support design and deformation prediction in mines.