本文采用DNS方法，对惯性颗粒所见各向同性湍流中具有平均标量梯度的被动标量场统计特性进行了研究。结果表明：惯性对颗粒温度脉动强度，两相温度关联，自相关特性以及颗粒热流与两相交叉热流的统计特性具有明显的影响。在PDF方法的框架下，系统地推导了非等温气固两相流的PDF方程，且基于朗之万随机体系对方程进行了封闭，并利用前面的DNS结果对相关标量模型进行了检验，结果表明，在颗粒所见流体温度的模型中，采用颗粒所见流体温度的拉氏时间积分尺度作为模型输入参数较采用流场的耗散时间和Kolmogorov常数作为输入参数更接近DNS的结果。“,”In this paper, statistics of passive scalar along inertial particle trajectory in homogenous isotropic turbulence with a mean scalar gradient are investigated by using of DNS. The results show that, the St number has effect on the scalar variance of particle, the scalar correlation and the cross heat flux. In the framework of PDF method, we derived the PDF equation of non-isothermal gas/particle flow, and closed it based on the random Langevin model system. As the results show that the computed results, based on this integral time scale, are more closed to the DNS than those based on the dissipation time scale of fluid flow and Kolmogorov constant.