Studies on traffic disaster in cold regions caused by road freezing and drifting snow

在1993- 1994年和1994- 1995年两个冬季，在北海道39号公路的关北山口和333号公路的天野隧道进行了路面结冰观测。在关北道，在普通道路上的道路、桥上的道路（桥路）和雪棚里的道路（雪棚路）上自动观测地表温度。桥梁路面比平时路面更容易结冰，特别是在冬季初、10月和11月。在冬季，即1、2月份，桥梁路面的表面温度比普通路面低约1 ~ 2个ºC，而在春季，即5月份以后，桥梁路面的表面温度高于普通路面，这意味着桥梁路面比普通路面更快地加热和冷却。在积雪路面上，直到4月份才会出现雪露表面，而其他表面则经常是干燥的，这说明太阳辐射对积雪路面没有加热作用。以大气温度、风速和天空辐射温度作为长波辐射的指标，模拟地表温度。所得温度变化趋势与实测数据吻合较好，平均温度低于1 ~ 2个<s:1>℃。道路采暖的室外实验在Kitami进行。采用降水传感器和冻结传感器，获得了最经济的加热操作。通过计算单个雪粒子的运动，模拟雪在障碍物周围的漂移，评价雪栅的性能。对于气流条件，给出了观测所得的平均流线，并假定湍流分量为正弦变化。在该模型中，对每个雪粒子的运动进行了计算，积累了大量的粒子运动，再现了积雪漂移的形成。

Observations of road surface freezing were carried out at Sekihoku Pass on Road 39 and Tanno Tunnel on Road 333 in Hokkaido during two winter seasons (1993-94 and 1994-95). In Sekihoku Pass, the surface temperature was automatically observed on a road on a usual road, a road on a bridge (bridge road) and a road in snowshed (snowshed road). The bridge road surface was more frequently frozen than usual road, especially in the beginning of winter season, October and November. In the winter season, January and February, the surface temperature of bridge road was lower than that of usual road about 1 or 2 ﾟC,but higher in spring season, after May, which means a bridge road is more quickly heated and cooled than a usual road. On the snowshed road, sherbet surface often appeared until April while other surface was frequently dry, which means the solar radiation does not contribute to heating the surface in a snowshed.The surface temperature was simulated from air temperature, wind speed and radiative temperature of sky as an index of long-wave radiation. The obtained temperature trends were well agreed with the observed data on the three types of road whereas the average was lower about 1-2ﾟC.The outdoor experiments for road heating was carried out at Kitami. The most economical heating operation was obtained using of a precipitation sensor and a freezing sensor.Calculating the movement of individual snow particles, the drifting snow around obstacles was simulated to evaluate the snow fence performance. For the air flow conditions, the average stream line was given from observation and the turbulence components were assumed as sinusoidal variation. In this model the movement of each snow particle was calculated, and accumulating lots of particle movement the formation of snow drift was reappeared.