论文部分内容阅读
建立了以纯金属原子半径、熔点、沸点和原子化焓预测表面张力的人工神经网络模型．训练后的神经网络能较好的拟会实验数据.对40种金属的表面张力进行回想和预测结果与实验值的偏差在可接受范围内，表明人工神经网络在纯金属表面张力预测方面有一定的前景．
An artificial neural network model was established to predict surface tension based on pure metal atom radius, melting point, boiling point and atomic enthalpy. The trained neural network can better simulate experimental data.The recall of the surface tension of 40 kinds of metals and the deviation of the predicted results from the experimental ones are within the acceptable range, which shows that artificial neural network has a certain degree in the pure metal surface tension prediction Prospects.