基于自回归求和移动平均的冬季路温短临预测

Pavement Temperature Short-impending Prediction Based on ARIMA in Winter

挖掘冬季路面温度在其他外部变量影响下未来短时间内的波动规律,建立冬季路面温度短临预测模型.基于交通气象监测站的冬季历史监测数据,利用统计学方法确定路面温度的主要影响因素,应用自回归求和移动平均(ARIMA)模型建模分析,对未来短时间内的路面温度进行预测.结果表明:允许误差在±0.5℃和±1.0℃范围内,未来3 h的平均预测准确率分别达到81.25%和99.65%,对应的平均绝对误差为0.21℃和0.26℃;允许误差在±0.5℃范围内,未来第1 h的平均预测准确率最高,平均绝对误差最低,分别达到92.50%和0.15℃.

Under the effect of other external variables, the fluctuation rule of pavement temperature with time was excavated in the future. Based on the historical data of a traffic meteorological monitoring station,the main affecting factors of pavement temperature were determined, and the shortimpending prediction model of pavement temperature time series was established by applying autoregressive integrated moving average（ ARIMA） model in order to predict pavement temperature in a short time. The results show that the predicted average accuracy reaches 81. 25% and 99. 65% respectively within allowable error range of ± 0. 5 ℃ and ± 1. 0 ℃ in the next 3 h. The predicted average accuracy and the mean absolute error are up to 92. 50% and 0. 15 ℃ respectively within allowable error range of ± 0. 5 ℃ in the next 1 h.