Virtual Sensing For NO_x Emission Of Diesel Engine Based On Vibration Signals

The diesel vehicle "emission gate" incident has had a significant impact on the development pattern of the global automotive industry,especially the government management and social supervision models of automotive products and automotive manufacturers.In addition to a solid legal foundation and government supervision,higher requirements are also placed on real-time monitoring technology for vehicle exhaust emissions.At present,the European NEDC cycle is mainly used to monitor automobile exhaust emissions.This method fails to record and monitor the transient emissions of vehicles in real time.Therefore,a real-time driving emission RDE(actual driving pollutant emission)monitoring method was developed later.Because the emission equipment monitored by RDE is large in size and expensive,it can only be used in automobiles for professional purposes such as scientific research and testing,and cannot be extended to mass-produced automobiles.Therefore,it is urgent to find a low-cost,portable and reliable emission monitoring method.The gas pressure in the cylinder is closely related to the NOx emission of the engine.Based on the in-cylinder combustion pressure signal,the indirect detection of the emission state of the internal combustion engine can be achieved.In practical applications,the installation of the cylinder pressure sensor will destroy the cylinder head structure,reduce its reliability and service life.In addition,the cost of the cylinder pressure sensor is high,thus limiting The application in the field of diesel engine emission monitoring.In-cylinder combustion events are an important driving source for engine vibration and emissions generation.Therefore,identifying and reconstructing the in-cylinder combustion state based on vibration signals is expected to become a more cost-effective emission prediction method.Vibration sensors are low in cost,small in size,and easy to install.It is of great significance to realize the real-time monitoring of internal combustion engines to identify the combustion state based on vibration signals and perform non-intrusive prediction of NOx emissions.This paper proposes a method of reconstructing cylinder pressure based on vibration signals and building a predictive model to evaluate emission levels.First,the correlation between the vibration signal and the cylinder pressure signal is obtained through the analysis of the similarity of the time-frequency domain atlas.A two-dimensional filter is constructed so that the vibration signal can reconstruct the second-order pilot signal of the cylinder pressure.Based on the reconstructed cylinder pressure signal,the combustion state signal that is strongly correlated with emissions is extracted.Based on the measured vibration and emission data under steady-state conditions,a PCR(principal component regression)and least squares prediction model is constructed to compare and analyze emissions Forecast effect.Finally,the feasibility of the proposed prediction model is verified by comparing the emission state prediction and actual measurement under transient operating conditions.The results show that there is a significant correlation between vibration and emission,and the response delay based on the prediction result of vibration signal is optimized by 1.2 seconds compared to the measurement result of the emission meter.