中深孔爆破振动加速度峰值的遗传BP网络预测

为准确预测爆破振动加速度峰值,保证爆破安全,相对于考虑因素少的经验公式法以及存在收敛性差、易陷入局部极小和计算复杂等缺陷的BP算法,提出了遗传BP神经网络算法,该算法具有更高的预测精度。以田湾核电站船山二期工程的试验数据为背景,比较分析并选择最大段药量、水平距离、总药量、高程差、爆破台阶高度和段别规模等6个参数作为输入层因子,建立了相应的爆破振动加速度峰值预测模型。结果表明,预测精度达到96.97%,验证了方法的可行性和有效性。

Prediction on peak vibration acceleration value of mediumdeep- hole blasting using genetic BP network

T o accurately predict the peak vibrat io n acceler at ion value o f blast ing and ensure the blast ing security, genet ic BP alg orithm w as proposed. T his alg orithm w as super io r to the empirical formula metho d w hich co nsidered few er factors, and the BP algor ithm w ith slow conv erg eve, local minimum and high computation complex ity . Based on the test data in Tianwan Nuclear Pow er Stat ion, a relevant fo recast ing model with the max imal charg e, hor izontal distance, to tal char ge, heig ht dif ference, blast ing bench height and scale as it s input elements, w as established to predict the peak vibration accelerat io n value. T he predict ion accuracy achieves 96. 97%, w hich verifies its feasibility and validity