中文摘要：

选取共轨柴油机喷油策略的8个核心参数，每个参数取4个水平值，通过正交试验设计方法获得64组正交组合并确定优化模型的目标函数和最佳主喷角，利用CFD计算每一组合的NO，排放、碳烟排放和燃油消耗率等子目标值。运用径向基（RBF）神经网络方法建立各子目标的近似模型，再利用Pointer优化算法结合序列二次规划法（NLPQL）的优化策略寻求最优的目标函数值。最后通过试验验证优化方案的燃烧及排放情况，同时分析优化方案的边界参数对柴油机排放和燃油经济性的影响规律。结果表明：采用“4段喷射＋15％EGR率＋0．28MPa增压压力”的优化方案能较好地解决排放与燃油经济性的平衡关系，并且仿真和试验结果具有很好的一致性，为共轨柴油机多工况不同喷油策略的确定提供优化途径。另外，NO，排放的降幅随着EGR率的增加而变小，碳烟排放和燃油消耗率稍有上升；EGR率增加时，随增压压力的提高，PM排放和燃油消耗率降幅变大，高的EGR率需要较高的增压压力以保持良好的柴油机排放和燃油经济性。

英文摘要：

Selecting 8 core parameters of injection strategies of common rail diesel engine and taking 4 level values for each parameter, the 64 orthogonal arrays were got by using the Orthogonal Test Design Method and the objective function of optimization model and the optimal main injection timing were determined, then CFD calculation was performed for each array to get sub-target values of NO, soot emission and BSFC. The Radial Basis Function（RBF）neural network was used to establish the approximate models for each sub-target respectively and the Pointer optimization algorithm combined with optimization strategy of Sequential Quadratic Programming Method（NLPQL）to seek the optimal objective function value. Finally,the combustion and emissions of optimized scheme were verified by experiments, the influence of optimized scheme boundary parameters on engine emissions and fuel economy were analyzed. The results show that simulation results are consistent with that of experiments and the optimized scheme of ＂4 pulse injection -t- 15EGR ＋ 0. 28 MPa boost pressure＂ can achieve a trade-off between emissions and fuel economy. Moreover,with EGR rate increasing, NOx reduction gets small, soot and BSFC increase slightly, and PM emission and BSFC reductions became larger as supercharging pressure increasing, high EGR rate needs higher supercharging pressure to maintain the engine emissions and fuel economy at low level. In this study, results of simulation are consistent with that of experiments, this method provides an optimization way for common rail diesel injection strategy under multiple conditions.