抛开数学建模的模式 ,尝试把神经网络方法引入内燃机的排放性能预测。该方法的预测精度不依赖于研究对象的数学模型 ,不需要复杂的数学演算 ,方法简便。以进气成分变化对柴油机排放性能的影响为例进行了探索 ,经过试验验证 ,证明了该方法的可行性。 Regardless of the mode of mathematical modeling, try to introduce the neural network method into the prediction of the emission performance of the internal combustion engine. The prediction accuracy of the method does not depend on the mathematical model of the research object, does not require complicated mathematical calculations, the method is simple. Taking the influence of the variation of intake composition on the emission performance of diesel engine as an example, the feasibility of this method is proved through the test.