| Gaseous nitrous acid（HONO）is an important atmospheric component that participates in atmospheric oxidation reactions.It can act as a detergent for nitrogen oxides（NOx）and influence ozone and secondary organic aerosols formation via generating hydroxyl radicals（OH）through photolysis in the troposphere.Vehicle exhaust is the main source of HONO emissions in urban areas.However,recent actual measurements and simulation studies have shown that vehicle HONO emissions may be seriously underestimated,restricting the generation and elimination mechanism of OH radicals,air quality simulation and the formulation of prevention and control strategies.Due to the wide range of HONO emission concentrations in vehicle exhaust,large fluctuations,and susceptibility to high temperature,high humidity,and interference components,the current HONO measuring instrument cannot meet the on-road emission HONO measurement of vehicle exhaust,and thus cannot characterize the real-driving conditions.There are only a few tunnel and chassis dynamometer test studies on HONO emissions from vehicles,and there is a severe lack of emission factors and emission models that can be used to quantify HONO emissions from vehicles.Given that,this study developed a vehicle-mounted gaseous nitrous acid measurement system that can meet the on-road emission HONO measurement of vehicle exhaust.A total of 43 vehicles were tested on the chassis dynamometer and on-road to identify the HONO emission characteristics and establish comprehensive HONO emission factors classified according to vehicle types and emission standards.The main factors that affect vehicle HONO emissions were identified,and a vehicle HONO emission model was constructed based on machine learning algorithms.The main results and conclusions of this paper are as follows:（1）The self-developed vehicle-mounted gaseous nitrous acid measurement system has good stability and accuracy,which can meet the requirements of vehicle exhaust HONO emission testing.The main performance of the system are:（1）The exhaust gas dilution technology can dynamically adjust the concentration of pollutants in the range of 10-1000times with good dilution stability（deviation less than 3%）and reduce the temperature and humidity of the exhaust.（2）Dual-range HONO measurement of 0-20 ppbv and 0-200 ppbv can be realized,which solves the problem of over-range caused by large fluctuations in HONO concentration.（3）It can effectively collect parameters such as vehicle operation,road conditions,etc.;（4）Automatic and platform-based data collection and analysis can be realized.（2）Traditionally,the HONO/NOx ratio used to calculate HONO emissions has great uncertainty.The actual measurement results found that the ratio of HONO/NOx between vehicles is in the range of 0.003-4.24%.Therefore,this study established HONO emission factors for different vehicle types and emission standards based on chassis dynamometer and on-road tests,which supplement the lack of emission factors.These factors can be directly used to estimate vehicle HONO emissions.They are providing an important data basis for the advanced calculation of vehicle HONO source emissions.（3）The main factors affecting HONO emissions of motor vehicles are fuel type,driving conditions,emission standards,start-up conditions,etc.The fuel type is the most important factor.The HONO emission concentration of diesel vehicle is about 1,000 times higher than that of gasoline vehicle.Driving conditions on the HONO emissions of vehicles is complicated.In general,as the VSP mode increases,HONO emissions increase.Urban and suburban roads have higher HONO emissions than highways,and on-road driving conditions have higher than chassis dynamometers.Tighter emission standards can reduce HONO emissions,and China III vehicles have obvious high HONO emissions.About the starting conditions,vehicles exhibit higher HONO emissions during cold start conditions.In addition to related pollutant emissions,vehicle HONO emissions are highly correlated with THC and NO2emissions.Therefore,vehicle HONO emissions are related to many factors,and comprehensive parameter analysis methods may quantify the vehicle HONO emission.（4）The vehicle gaseous nitrous acid emission model（HONO-VEM）built based on the machine learning algorithm has a good simulation effect（R2=0.871）in the emission prediction process.The linear correlation performance of the HONO-VEM improved2.2-89.3%than the based models of MIR,RF,XGBoost and Light GB.THC and NO2have the highest scores for the feature importance scores,indicating that they play an important role in HONO production.The deviations of HONO-VEM,VSP method,and HONO/NOx are 2.0%,16.3%,and-23.7%,respectively.HONO-VEM has more minor prediction deviations,so HONO-VEM has important application significance for estimating HONO emissions of vehicles.The vehicle-mounted gaseous nitrous acid measurement system developed in this research can be widely used to measure pollutant HONO emissions.At the same time,the HONO-VEM model can accurately calculate the HONO emissions of the vehicle exhausts.Our work can provide technical support and method references for the quantitative research of HONO. |
