节点文献
基于LM-BPNN方法的爆破震动灾害预测模型
LM-BP Neural Networks of Peak Particle Vibration Velocity Forecast for Blasting and Its Application
【摘要】 为了探索提高控制爆破震动效应的方法,基于Levenberg-Marquardt算法改进的BP神经网络模型,建立以最大段药量、爆心距、高差作为影响爆破振动的主要因素,对爆破震动速度进行预测的模型。用爆破振动观测数据进行训练和预测,预测结果与现场观测结果吻合良好。结果表明:与基于标准BP、Polak_Ribiere共轭梯度、专家经验公式等计算结果比较,LMBPNN算法具有良好的鲁棒性和预测精度,预测效果较优,对爆破震动安全评价及其灾害控制有一定的应用价值。
【Abstract】 Use of robust and learning ability of fuzzy-neural network based on the arithmetic of Levenberg-Marquardt is made of simulate the nonlinearity relation among blasting parameters to build a model to forecasting the peak particle vibration velocity for blasting. The test results have fairly agree with actual projects. Analysis shows that the model has higher theoretical and practical reference to the studies on the vibration effect and the control of blasting vibration damage than other models.
【Key words】 Levenberg-Marquardt BP neural networks blast vibration disaster control;
- 【文献出处】 科学技术与工程 ,Science Technology and Engineering , 编辑部邮箱 ,2014年35期
- 【分类号】TU751.9
- 【被引频次】2
- 【下载频次】47