节点文献
神经网络与遗传算法优化CO2柱塞泵容积效率
Optimization of Volumetric Efficiency for CO2 Plunger Pumps by Using Neural Networks and Genetic Algorithms
【摘要】 容积效率是表征CO2柱塞泵性能的重要指标,提出了CO2压缩滞留特性会对泵的容积效率造成影响,采用数值模拟和正交试验方法,分析了CO2柱塞泵几何尺寸对容积效率的影响。结果表明:容积效率随柱塞直径、柱塞分布圆直径和斜盘倾角的增大而增大,是影响容积效率高显著因素;随腰形槽宽度和开角的增大而减小,是影响容积效率低显著因素;进出口管道直径对容积效率的影响可忽略不计。运用神经网络构建显著性较高几何因素与容积效率的近似函数,预测结果精度较高,误差不超过3.94%。经过耦合遗传算法全局优化,获得了柱塞泵的优化几何尺寸组合,优化后的几何尺寸组合相较初始尺寸容积效率提高了29.16%。研究成果可为CO2斜盘式轴向柱塞泵容积效率优化提供参考。
【Abstract】 The volumetric efficiency was an important index to characterize the performance of CO2 plunger pumps, and it was proposed the influence of the compression retention characteristics of CO2 on the volumetric efficiency of the pump. The influence of the geometric dimensions of CO2 plunger pumps on the volumetric efficiency was analyzed by using numerical simulation and orthogonal test methods. The results showed that the volumetric efficiency increased with the increase of plunger diameter, plunger distribution diameter and swash plate inclination, which was a significant factor that affecting the volumetric efficiency; It decreased with the increase of the width and opening angle of the waist groove, and it was a significant factor that affecting the low volumetric efficiency. The effect of inlet and outlet pipe diameters on volumetric efficiency was negligible. By using the neural network to construct the approximate function of geometric factors with high significance and volume efficiency, the prediction results had high accuracy and the error was not more than 3.94%. Through the global optimization of coupled genetic algorithm, the optimal geometric size combination of plunger pumps was obtained, and the optimized geometric size combination was 29.16% higher than the initial size and volume efficiency. The research results could provide reference for the volume efficiency optimization of CO2 swash plate axial piston pumps.
【Key words】 CO2 plunger pumps; neural network; genetic algorithm; volumetric efficiency; orthogonal experimental design;
- 【文献出处】 新技术新工艺 ,New Technology & New Process , 编辑部邮箱 ,2026年02期
- 【分类号】TE934;TP18
- 【下载频次】8