论文部分内容阅读
针对BP神经网络对工程爆破振动的预测存在精度不够高的缺点,建立遗传算法优化神经网络的模型,并介绍了它的原理。最后通过爆破振动预测实例的介绍,应用MATLAB编程,将总装药量Q、测点与爆源的高差h、孔间微差时间t、最大药包距离L这4个参数作为模型参数,对爆破振动幅值v、振动主频f和振动持续时间T进行预测,得出基于遗传算法的神经网络预测的结果比BP神经网络更为精确,克服了BP神经网络的缺点。
The prediction of blasting vibration in BP neural network has the shortcoming of not high enough precision in the prediction of blasting vibration. The model of genetic algorithm to optimize neural network is established and the principle of BP neural network is introduced. Finally, through the introduction of blasting vibration prediction example, using MATLAB programming, the four parameters of total charge Q, height difference between measuring point and explosive source h, time difference between holes t, maximum drug packet distance L are taken as model parameters, The blasting vibration amplitude v, the vibration frequency f and the vibration duration T are predicted, and the result of neural network prediction based on genetic algorithm is more accurate than BP neural network, which overcomes the shortcomings of BP neural network.