CAJ | 학술논문

为了确定各因素对惯性分离室的性能的影响，以及寻找最优的参数组合，提出了一种新的 BP 神经网络的改进方法，对联合收割机惯性分离室的吸运系统压力损失的实验数据进行了拟合，并与二次回归模型方法进行比较。结果表明，改进的BP神经网络的拟合精度明显优于二次回归的拟合精度；同时，通过 BP 神经网络的优化方法求取了4个参数的最优组合值，为惯性分离室的性能研究提供了一种新的方法。
위료학정각인소대관성분리실적성능적영향，이급심조최우적삼수조합，제출료일충신적 BP 신경망락적개진방법，대연합수할궤관성분리실적흡운계통압력손실적실험수거진행료의합，병여이차회귀모형방법진행비교。결과표명，개진적BP신경망락적의합정도명현우우이차회귀적의합정도；동시，통과 BP 신경망락적우화방법구취료4개삼수적최우조합치，위관성분리실적성능연구제공료일충신적방법。
To determine the effects of various factors on the performance of the inertia separation chamber , and to find the optimal combination of parameters , this paper proposes an improved BP neural network method that had fitted the ex-perimental data of suction system pressure drop in the inertia separation of combine harvester .Compared with the quadrat-ic regression , the results showed that the fitting precision based on improved BP neural network was superior than quad -ratic regression .The value of the optimal combination of the four parameters was calculated by optimizing BP neural net -work.This paper provides a new method for the study of the performance of the inertia separation chamber .