Atomic model of liquid pure Fe

Using a θ-θ X-ray diffractometer, the liquid structure of pure Fe was investigated and the diffraction intensity, structure factor, pair distribution function as well as the coordination number and atomic distance were obtained. The experimental results showed that there was also a pre-peak on the curve of the structure factor of liquid pure Fe. The pre-peak is a mark of medium-range order in melts. According to the characteristics of pre-peak, an atomic model of liquid pure Fe is constructed, namely, the structure of liquid pure Fe is a combination of clusters consisting of bcc cells with shared vertexes and other atoms with random dense atom distribution.     

Pre-peak on the structure factor of liquid hypoeutectic Al-Fe alloy

Using a θ-θ X ray diffractometer the structure of liquid hypoeutectic Al Fe alloy at 675℃ and its change with different thermal histories are investigated . Liquid Al is used as a reference system. It is found that a pre peak appears on the small angle part of the structure factor of the nonsuperheated hypoeutectic Al Fe alloy, but disappears after superheating whereas the structure factors of liquid Al hardly change with the different thermal histories, and no pre peak can be observed on them. The appearance of a pre peak is a mark of the intermediate range order (IRO). The pre peak is taken as the correlation between Fe atoms on the IRO length scale. Assume that the crystalline structure is the first order approximation of the liquid structure. A model structure is constructed. The basic unit is a cube formed by 8 Al atoms on its corner and one Fe atom occupying its center. The translation of such a unit along its fourbody diagonals by the length of a diagonal can meet the requirement of the Fe_Fe distance and gives a DO 3 like structure. If the vacancy among the units is filled with fcc like Al cells, the composition of the entity is about Al 7Fe, close to that of the metastable phase Al 6Fe from rapid solidification. It is speculated that there are Al 6Fe like clusters in the liquid hypoeutectic Al Fe alloy.     

基于粒子群支持向量机的矿岩强度指标的超声预测

建立以吸水率、干密度、波阻抗、动泊松比、动弹性模量为输入,抗压强度为输出的支持向量机预测模型.为了提高支持向量机预测精度,引入了粒子群算法对支持向量机的参数进行优化,克服支持向量机参数人工选取的不足.通过对鞍千矿和弓长岭矿的矿岩样本数据分析,该模型的预测误差最大为8.2%,精度明显高于传统神经网络法.结果表明基于超声波预测强度的方法具有很好效果,可望成为一种岩石强度预测的新方法.