基于重要指示变量和特征的高速公路浓雾短临预测研究

Study on Short-term Prediction of Dense Fog on Expressway Based on Important Indicator Variables and Characteristics

浓雾天气严重影响高速公路车辆运行安全,导致道路交通事故频发,而对高速公路浓雾天气的有效预测是解决交通事故多发问题的有效途径。因此,通过交通气象监测站获取全时段高速公路气象数据,立足浓雾形成机理和车辆运行实际需求,选用"滑动窗+模糊统计和线性插值"方法对其气象数据进行预处理,获取浓雾产生过程的连续气象要素数据,并提取出浓雾产生前3 h的各气象指标数据。利用统计分析方法获取和分析浓雾形成前各气象要素数据变化趋势,进而采用先决条件阈值和滑动窗口算法,确定出浓雾形成的重要指示变量及变量的变化特征。最终,提出了基于"能见度前期振荡"和"大气温度回温波动"特性的浓雾短临预测模型,利用沪宁高速公路沿线江苏境内交通气象监测站2013—2015年浓雾发生前3 h气象数据对短临预测模型参数进行标定,进而对其2016年气象数据进行浓雾短临预测。结果表明:浓雾短临预测模型的预测结果中22组数据与实际结果相符,遗漏数据组7组,其预测准确率达到了75.86%,漏检率仅为4.35%,故该模型可以实现对高速公路沿线浓雾产生前短时间内进行有效预测,能及时警示公路运营管理部门提前做好预防处理准备,降低因突发浓雾而造成的交通事故发生的概率,极大的保障高速公路车辆运行安全。

Heavy fog weather seriously affects the safety of vehicles on expressway,leading to frequent road traffic accidents.The effective prediction of heavy fog weather on expressway is an effective way to solve the problem of frequent traffic accidents.Therefore,the full-time meteorological data on expressway are obtained at traffic weather monitoring station.Based on the formation mechanism of dense fog and the actual needs of vehicle operation,the meteorological data are preprocessed by the method of"sliding window+fuzzy statistics and linear interpolation"to obtain the continuous meteorological element data of the dense fog production process,and the meteorological indicator data of the 3 h before the occurrence of dense fog are extracted.The change trend of the meteorological element data before the formation of dense fog is obtained and analyzed by using the statistical analysis method,and then the important indicator variables and their change characteristics during the formation of dense fog are determined by using the prerequisite threshold and sliding window algorithm.Finally,a model for short-term forecasting dense fog based on the characteristics of"visibility early oscillation"and"atmospheric temperature return fluctuation"is proposed,and the parameters of the proposed model are calibrated using the meteorological data before the occurrence of dense fog of the traffic meteorological monitoring stations along the Shanghai-Nanjing Expressway in Jiangsu Province from 2013 to 2015,and then the meteorological data of 2016 are used to predict the short-term dense fog.The result shows that 22 sets of data in the prediction result using the dense fog short-term prediction model are consistent with the actual result,and 7 sets of data are missed.The prediction accuracy reaches 75.86%,and the missed detection rate is only 4.35%.Therefore,the model can effectively predict the dense fog along the expressway within a short period of time,and can promptly warn the highway operation management department to prepare for preventive treatment in advance,reduce the probability of traffic accidents caused by sudden thick fog,and greatly guarantee the safety of vehicles on running on the expressway.