Study On Stability Of Fuel Quantity And Emission Optimization With Multiple Factors Based On Multi-Injection

Currently, the petroleum resource faces drying up and effect of greenhouse is increasingly serious. The characteristic of high thermal efficiency of diesel engine becomes more important. But, the exhaust emissions of NOx and PM of diesel engines are relatively more. Along with the exaltation of the consciousness of human's environmental protection and tighter emission control regulations, the high emissions of diesel engines must be solved. The difficulty to solve the emission problems of diesel engines is the trade-off between NOx and PM. With the efforts for many years, the researchers found that splitting one fuel injection per cycle into two or more can reduce the high emissions of NOx and PM simultaneously. That is multiple injections. But, the performance of multiple injections rely heavily on the selection of parameters of injections. Bad parameters of multiple injections will deteriorate performance of diesel engines.Formerly, most of experimental study only changed one factor to investigate the relation between factor and response. When researchers consider several factors, the experiments that adopt the strategy of one-factor-at-a-time can not nearly find out the optimization point. There are some interactions between factors. When changing the level of other factors, the relationship between the factor and the response may change strikingly. The efficiency of the experiment of one-factor-at-a-time is very low. For many times experiment, little information about system can be gotten.In the actual experiment, the author find that when opening the post injection and fix the total fuel mass per cycle and fuel mass of post injection per cycle in calibration system, the actual total fuel mass per cycle which is measured by fuel consumption meter and engine torque produce large fluctuations with the change of the interval between main injection and post injection. Based on the deep analysis of this phenomenon, it is caused by the shutdown event of injector when the main injection closes. The shutdown event leads to the pressure wave in high-pressure oil pipeline connecting the injector and common rail. This paper will consider impact of five factors for emissions and economic impact of diesel engines. They are the timing of main injection (factor A), the pressure of common rail (factor B), the interval between main injection and post injection (factor C), the fuel mass of post injection per cycle (factor D), and the ratio of EGR (factor E). To obtain more information with less experiments and find out accurately optimization point, the author make use of the knowledge of DOE and statistical analysis to analyze and dispose datas.This paper makes the prediction model of the mass of post injection and the prediction model of emission (All is black box model). These models are made use of to stabilize the fluctuation and solve the optimization problem of emissions and economic under multiple factors. The experimental datas making models come from the engine test bench. The experiments were conducted on a CA6DL2-35E3 high-pressure common-rail turbocharged and intercooler heavy-duty diesel engine. After the experiments and the analysis of experimental datas, main contents and conclusions are summarized as follows:1) With the ANOVA, that whether the changes of some fuel injection parameters can cause changes in the amount of fuel injection or not was investigated. The results of ANOVA shows:1 The change of injection timing makes the fuel mass of single injection have fluctuation (underα= 0.05).2 When open the post injection, the change of energizing time of main injection has significant influence on the fuel mass of post injection per cycle.3 When open the post injection, the timing change of main injection has significant influence on the fuel mass of post injection per cycle.2) With the method of changing the energizing time of post injection, the fluctuation of total fuel mass per cycle can be stabilized by stabilizing the fuel mass of post injection per cycle. The prediction model of the mass of post injection was made. Some parameters of model is:the P value of F test of model<0.0001, R2=0.9994,Ra2=0.9988, R2jaekknif=0.9981, but the P value of lack of fit test=0.0145. The model was verified by select 15 points. The result of verify is:R2Pred=0.9863. Before making use of model to correct the energizing time of post injection, the mean square of the deviation between the actual total fuel mass per cycle and the target total fuel mass per cycle is 39.066. After correcting, the mean square of the deviation is 0.457. Its effect is remarkable.3) The emission models were made. Factors respectively are the timing of main injection (factor A), the pressure of common rail (factor B), the interval between main injection and post injection (factor C), the fuel mass of post injection per cycle (factor D), and the ratio of EGR (factor E).The responses are smoke opacity, brake specific emission of NOx and BSFC. Under 1650r·min-1, with the target mass of fuel injection is 85.34 mg/cyc, the experiments are conducted. The parameters which show the goodness of fit of smoke opacity model are:the P value of F test of model<0.0001, R2=0.9995, Ra2=0.9979,R2 jackknif=0.9888, the P value of lack of fit test=0.0513. The parameters which show the goodness of fit of brake specific emission model of NOx are:the P value of F test of model<0.0001, R2=0.9994. Ra2=0.9988, R2 jackknif=0.9975, the P value of lack of fit test=0.1028. The parameters which show the goodness of fit of BSFC model are:the P value of F test of model<0.0001,R2=0.9953,Ra2=0.9917,R2jackknif=0.9878, the P value of lack of fit test=0.3354. Through satisfaction functions the optimization point can be find out. The standers are that the emissions meet Europe IV regulation and the BSFC is as low as possible. The optimization point is: the timing of main injection= 118.08MPa, the pressure of common rail =BTDC3.26°CA, the interval between main injection and post injection=1197us, the fuel mass of post injection per cycle=30mg/cyc, and the ratio of EGR=10%. The response value of prediction is:smoke opacity= 4.18%, brake specific emission of NOx =3.42g/kw.h, BSFC=214.76g/kw.h. Using the combination of factors to conduct engine test bench, the validation values are:Smoke opacity= 4.08, brake specific emission of NOx=3.43g/kw.h, BSFC=214.32g/kw.h. The agreement between the experimental and predicted is quite well. But, the BSFC rise 7%after optimization. The model show if making the smoke opacity and BSFC equal to the levels of the prototype, the minimum brake specific emission of NOx is 5.43g/kw.h that drops 37%compared to the prototype.