Soft Measurement And Monitoring System Of Vehicle Fuel Consumption Based On The RBF Neural Network

With the development of automobile industry and sustained high growth of carownership, the energy conservation and environmental protection become animportant subject of auto industry, and people are increasingly concerned with thefuel economy performance of automobile. Fuel consumption is the key indicator ofevaluation of fuel economy. Fast and accurate measure of the car fuel consumptionbecomes the important and effective evaluation means of automobile fuel economyperformance, and continuous improvement of fuel consumption measurement methodshave a very important influence on the evaluation of fuel consumption. In view of theproblem that existing fuel consumption measurement methods have high cost,complexity and difficult implement on-board real-time measurement, this article putforward the car fuel consumption on the basis of RBF neural network softmeasurement method, and designed the real time monitoring system for correspondingfuel consumption. The concrete research content mainly includes three facts asbellow.Firstly, research the current method and principle of common fuel consumptionmeasurement,and analysis and compare the advantages and disadvantages of variousmethods. According to the requirement of real-time measurement of the vehicle fuelconsumption, this paper put forward corresponding RBF neural network softmeasurement method.Secondly, after Analysis and selection the main factors affecting fuelconsumption, fuel consumption soft measurement model was designed on the basis ofRBF neural network; The test plan was designed for measurement of fuel consumptionmeasurement model, and the test data can be used as training samples and test samples;In combination with principal component analysis method, RBF neural networktraining speed was improved. The fuel consumption measurement model wassimulated in Matlab and the simulation results were compared with measurement databy fuel consumption meter. With this method, the effective feature of measuringmethod was evaluated.Thirdly, the ARM control and monitoring system were designed for on-board testpart, and they can be used for analysis, processing, sending and transportation of vehicle fuel consumption data, and real-time measurement and monitoring foraccurate fuel consumption.In this paper, the thought of the soft measurement method was applied to vehiclefuel consumption measurement, and the car fuel consumption soft measurement wasput forward based on RBF neural network (RBFNN). This method took advantage ofengine operating parameters easy to measure to reflect indirectly the fuel consumptioninformation which can’t be measured directly on-board and real-time, using RBFneural network algorithm to the nonlinear approximation of measurement model.Analysis and select main factors affecting fuel consumption (engine speed, throttleopening and engine temperature) as the input parts of the measurement model, and theself-learning fuel consumption soft measurement model was set up. Fuel consumptionreal-time monitoring system can be used in real-time monitoring vehicle fuelconsumption, which will be convenient for city, transport and operating company inreal-time measurement and management of car fuel consumption.