Carbon Deposition Predictions Of Engine Based On Time Series Analysis And Neural Networks

With the auto penetration rate increased year by year,the environmental problems become increasingly serious,automobile exhaust pollution of the atmosphere is more and more serious,the appeal of environmental protection and energy saving and emission reduction is getting higher and higher,One of the most serious impact on automobile exhaust emissions is carbon deposition.Carbon deposition will cause incomplete combustion of gasoline,resulting in excessive exhaust gas,increase atmospheric pollution;will reduce the combustion efficiency,more energy consumption;will accelerate the engine damage,vehicle condition decreased.It has important theoretical and practical significance to study the formation factor of engine carbon deposition and predict the trend of its change for energy saving and environmental protection,and to remind drivers to maintain good car condition.At present,there are three main methods to detect the carbon deposition: disintegration,endoscopy and oxygen sensor feedback voltage method.The first two methods need to be implemented in a car repair shop or with an endoscopic device.The third method is a kind of indirect diagnosis method for the carbon deposition of the engine by the feedback voltage of the oxygen sensor.But none of these methods can predict the formation and trend of carbon deposition,and found no factors of carbon deposition,but also can not give the driver the right advice,and can not achieve online monitoring.In this paper,a data acquisition terminal is designed to obtain the parameters of the vehicle through the OBD,and based on the neural network and time series,the carbon deposition prediction model is established,intuitive analysis of the reasons for the formation of carbon deposition,to help drivers improve driving habits,to achieve the purpose of energy-saving emission reduction and economic savings.The main contents of this paper are summarized as follows:First,the hardware,the development of vehicle data acquisition terminal based on OBD.The terminal uses STM32 as the core chip,including OBD protocol analysis module,SD card reader module and GPRS remote transmission module;can realize real-time acquisition of vehicle data,local storage and upload Ali cloud server.The relevant parameters of the vehicle used in this paper are collected from the terminal.Second,information mining,according to the voltage value of the sensor and the vehicle speed information,the influence of different working conditions,driving habits and different road conditions on the formation of carbon deposition was studied.Through the R language for data analysis and mining,it is concluded that idle speed,stable driving and high-speed road traffic is not easy to form carbon accumulation.Third,the software,the establishment of carbon prediction model to predict the trend of carbon deposition based on time series and neural network.Taking the vehicle data as the research object,four kinds of carbon deposition prediction models were established.Through the comparison of the three sets of experiments,an effective method for predicting carbon deposition was explored,and the external factors affecting the formation of carbon deposition are analyzed.It is concluded that the use of time series and neural network to predict carbon deposition is accurate and effective,and Carbon deposition is affected greatly by vehicle driving conditions.