论文部分内容阅读
为提高化工园区火灾热辐射风险预测精度,预防设备二次损坏事故的发生,通过建立3层反向传播(BP)神经网络模型来预测设备损坏概率,研究二次损坏概率与影响因素参数(设备容积,视角系数,与爆炸点的距离)之间关系。以用多米诺风险理论计算出的设备二次损坏概率为样本集,对建立的BP神经网络模型进行训练、测试和误差分析。结果表明,设备损坏概率与神经网络预测的概率之间最小误差值为9.962 5×10-3。设备二次损坏概率随距离的增大而减小,且随设备容积、视角系数的增大而增大。其中,视角系数对损坏概率的影响最明显。
In order to improve the prediction accuracy of thermal radiation risk in chemical industry parks and prevent the occurrence of secondary damage to equipment, the probability of equipment damage is predicted by establishing 3-layer BP neural network model and the secondary damage probability and influencing factors (equipment Volume, viewing angle coefficient, distance from the explosion point). Taking the probability of secondary damage of equipment calculated by domino risk theory as a sample set, the BP neural network model established is trained, tested and error analyzed. The results show that the minimum error between the equipment damage probability and the neural network prediction probability is 9.962 5 × 10-3. The probability of secondary damage to equipment decreases with increasing distance, and increases with equipment volume and viewing angle coefficient. Among them, the perspective coefficient of damage probability of the most obvious impact.