拉格朗日随机模型模拟屋面积雪分布

Roof snow distribution simulation based on lagrange stochastic model

基于拉格朗日随机模型并结合已有文献中最新改进的k-ω湍流模型,对风致屋面雪粒迁移运动进行了数值模拟,其中根据雪深变化对屋面积雪边界采用了自适应变形调整。首先对程序进行了验证,然后对一典型阶梯形屋面的积雪分布进行了数值模拟。通过求解RANS方程及对湍流脉动速度PDF采样,模拟计算了约3×104个粒子的跃移及悬移运动迹线,统计获得了不同粒子大小、风速以及阈值风速下的屋面积雪效率及雪深分布。结果表明:在分离泡及拐角等部位的积雪效率最高,是积雪较多的地方;直径0.2mm的粒子屋面积雪效率最高;当屋面流动剪切速度大于雪粒阈值剪切速度时,不同风速及阈值风速对屋面的积雪效率影响相对较小;模拟结束时计算的屋面积雪系数分布与观测结果吻合较好。

Based on Lagrange stochastic model and the last improved k-ω turbulent model, a numerical analysis program is developed to simulate wind-induced roof snow distributions. In the method, the computational domain boundaries are changed according to snow-depth by using the adaptive deformation technique. The trajectories of about 3×104 particles are calculated, in which the flow instantaneous velocities are obtained through solving the RANS equations and PDF sampling. Snow trapping efficiency and snow depth distribution are obtained through statistics of the particle traces. Through a typical stepped roof snow distribution simulation, the results show that, the snow trapping efficiency are highest in the parts of roof corner and the separation bubble; the particle size, velocity and threshold velocity have an influence on the roof snow trapping efficiency. For different sizes of particles, the diameter of 0.2mm has a maximum trapping efficiency. When shear velocities on the roof are greater than the threshold shear velocity, the effect of velocity and threshold velocity is comparatively small. The calculated snow-depth coefficients on the roof are well consistent with those of field observations.