A method for mapping maximum instantaneous wind speed for monitoring strong wind along railway lines.

Japanese railway companies monitor strong winds by observing instantaneous wind speed using anemometers, discretely installed along railway lines. When anemometers observe strong winds, train operations are controlled. Anemometers are installed only in empirically known windy sections. On the other hand, stronger typhoons and other storms have increasingly approached and landed in Japan in recent years. Consequently, there is a possibility of strong winds blowing even in sections where anemometers are not installed. So, we have developed a method for mapping maximum instantaneous wind speeds in sections where anemometers are not installed, without installing additional anemometers.

We developed a method for mapping wind speeds equivalent to the maximum instantaneous wind speeds. First, we carried out Computational Fluid Dynamics (CFD) analysis using a Large Eddy Simulation model. CFD domain needs to include actual wind observation points, such as AMeDAS stations installed by the Japan Meteorological Agency and anemometers installed by railway companies. We set CFD domain at 30 km x 30 km in the horizontal direction and 10 km in the vertical direction, including one railway line and one AMeDAS station (hereafter, this AMeDAS station referred to as the reference point). The grid spacing in the CFD domain was set to 100 m in the horizontal direction and unevenly spaced in the vertical direction. Also, we conducted CFD analysis for each of the 16 wind directions.

Next, we calculated two indices UR(x,y) and GR(x,y) at any grid point (x,y) within our CFD domain, using CDF results. UR(x,y) are ratios of the averaged wind speed U_ave(x,y) at grid point (x,y) to the averaged wind speed U_ref at the grid point where the reference point is located (UR(x,y) = U_ave(x,y)/U_ref). GR(x,y) are ratios of the maximum wind speed U_max(x,y) at each grid point (x,y) to the averaged wind speed U_ave(x,y) at each grid point (x,y) (GR(x,y) = U_max(x,y)/U_ave(x,y)). By multiplying observational averaged wind speed at the reference point by UR(x,y) and GR(x,y), we can obtain the spatial distribution of wind speed equivalent to the maximum instantaneous wind speed.

To validate our method, we conducted wind observations at six sites along a railway line for two years and obtained 14 cases of strong winds. Each of these 14 cases included daily maximum instantaneous wind speeds of 20 m/s or higher at one or more of the six sites. Observed 10-minute maximum instantaneous wind speeds for 14 cases were taken as the true values, we evaluated error of the wind speeds estimated by our mapping method using Root Mean Squared Error (RMSE). As a result, RMSE of the estimated 10-minute maximum instantaneous wind speed for the 14 strong wind cases were less than 5 m/s at four of the six sites.