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

在两个冬季（1993-94和1994-95），在北海道的39号公路的关北山口和333号公路的丹野隧道进行了路面结冰的观测。在Sekihoku Pass，在普通道路上的道路、桥上的道路（桥路）和雪棚中的道路（雪棚路）上自动观察表面温度。桥梁路面结冰的频率高于普通路面，尤其是在冬季初、10月和11月。在冬季，1月和2月，桥梁道路的表面温度比普通道路低约1或2摄氏度，但在春季，5月后，桥梁道路的表面温度较高，这意味着桥梁道路比普通道路更快地加热和冷却。在雪棚道路上，雪面经常在4月前出现，而其他地表经常是干燥的，这意味着太阳辐射对雪棚内地表的加热没有贡献。所获得的温度趋势与三种类型道路上的观测数据非常一致，而平均值低约1-2摄氏度。在北见进行了道路加热的室外实验。使用降水传感器和结冰传感器获得最经济的加热操作。计算单个雪颗粒的运动，模拟障碍物周围的积雪，以评估防雪栅栏的性能。对于空气流动条件，平均流线由观测给出，湍流分量假定为正弦变化。在该模型中，计算了雪粒子的运动，并通过大量粒子运动的累积再现了雪堆的形成过程。

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.