Characteristics Of Diesel Fuel Injection Rate Under Multi-injection Condition And The Control By ANN

Currently high pressure common rail fuel injection system is the most widely used fuel system for diesel engines due to the fact that it can provide a high and constant injection pressure,precise injection strategy and excellent spray quality.As there exists a pressure fluctuation for sequent injection pulse during multi-injection,characteristics of the fluctuation and its effects on injection performance should be examined.Using an in-house experimental platforms,relevant investigations including pressure fluctuation acquisition experiment and injection rate measurement experiment were carried out.Then the difference of the amount of injection fuel and injection rate and also the pressure fluctuation of the high pressure common rail piezoelectric injector system were analyzed.Results show that in two-stage fuel injection strategy,the pressure fluctuation had a significant impact on the second injection pressure,injection rate and fuel-injection quantity.With regard to the pressure fluctuation,experimental results show the periodic fluctuation is around 10%.This results in the fuel injection quantity has consequent change with injection interval.Under the condition of low fuel injection pulse width,injection rate changes obviously with injection pulse-width,and with injection interval.As injection pressure and injection rate will directly influence atomization quality and combustion,it is critical to study the characteristics and make possible control to injection pressure,injection quantity and injection rate.Using BP algorithm of artificial intelligence,three neural network models for injection pressure prediction,fuel injection quantity prediction and fuel injection rate prediction,respectively,were built and trained for realizing the control of the second injection parameters,including injection pressure,injection rate and injection quantity.The control of three neural networks on fuel injection strategy is a logical relations of coordination controls,which reflect injection interval control by injection pressure,fuel injection pulse width control by injection interval and fuel injection quantity,fuel injection pulse width correction control by injection interval and fuel injection rate.Those control relations are actually acting on the control of second injection parameters by two injection interval and injection pulse width.Analysis results of neural network models show that BP neural network models can be used to predict required injection interval and injection pulse width very well.The error of prediction results is less than 5%.This suggests that the developed neural network models can be used to control the fuel injection strategy and can be applied to optimize design of high pressure common rail system for optimal control function.