推进技术 ›› 2012, Vol. 33 ›› Issue (4): 620-624.

• 材料 推进剂 燃料 • 上一篇    下一篇

基于遗传算法结合支持向量机的Mg/PTFE贫氧推进剂配方优化

范磊,潘功配,欧阳的华,陈昕,逄高峰   

  1. 南京理工大学 化工学院,江苏 南京 210094;南京理工大学 化工学院,江苏 南京 210094;南京理工大学 化工学院,江苏 南京 210094;南京理工大学 化工学院,江苏 南京 210094;南京理工大学 化工学院,江苏 南京 210094
  • 发布日期:2021-08-15
  • 作者简介:范磊(1973—),男,博士生,讲师,研究领域为军事化学与烟火技术,推进剂配方设计。E-mail:oydn2010@yahoo.com.cn
  • 基金资助:
    国家部委科研项目(40406030201)。

Application of Genetic Algorithm-Support Vector Machine in Formula Optimization of Mg/PTFE Fuel Rich Propellant

  1. School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China;School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China;School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China;School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China;School of Chemical Engineering, Nanjing University of Science and Technology, Nanjing 210094, China
  • Published:2021-08-15

摘要: 针对Mg/PTFE贫氧推进剂配方设计的复杂性,采用支持向量机理论建立了相关预测模型,结合遗传算法对模型结果进行多目标寻优,以此获得最佳的配方,最后对所得的最佳配方进行了实验验证。结果表明Mg/PTFE贫氧推进剂的最佳配方为PTFE/Mg=0.49,酚醛树脂含量为12.50%,镁粉粒度为26.90μm,PTFE粒度为111.33μm。遗传算法结合支持向量机的优化方法,适合于推进剂配方的优化,具有一定的实际应用价值。 

关键词: 遗传算法;支持向量机;推进剂;配方优化

Abstract: According to the complexity of formulation design for Mg/PTFE fuel rich propellant, a prediction model of formulation design for Mg/PTFE fuel rich propellant with the support vector machine (SVM) was introduced,and the genetic algorithm (GA) was used for multi-objective optimization to obtain optimal formula composition, which was verified with experiment at last. Results show that the optimum formula composition for Mg/PTFE fuel rich propellant is PTFE/Mg= 0.49, the mass content of phenolic resin is 12.50%, the diameter of Mg is 26.90μm, the diameter of PTFE is 111.33μm. GA-SVM is suitable for formula optimization of Mg/PTFE fuel rich propellant, which has certain practical application value. 

Key words: Genetic algorithm; Support vector machine; Propellant; Formula optimization