【摘要】 针对实际加工过程中通过调控HTPB推进剂的氧化剂含量来进行燃速调控，提高产品合格率的工程需求，采用数据驱动的方法，建立了以HTPB推进剂其他原材料、燃速值为输入变量，氧化剂含量为输出的燃速调控模型。采用了随机森林回归、支持向量回归等多种机器学习方法，以最大相对误差、均方根误差以及平均相对误差作为模型的性能评价指标，对模型超参数进行调整，并通过工程经验知识对不同模型进行筛选，利用样本集外的8组数据对模型进行验证。结果表明，模型预测值与实际测试值的平均相对误差为0.75%,且模型通过调整氧化剂含量使得调控燃速更接近目标燃速。基于数据驱动的调控模型，对实际推进剂生产加工过程中配方参数调整具有重要的指导意义。更多还原

【Abstract】 In response to the engineering requirements of burning rate regulation by adjusting the oxidizer content during HTPB propellant processing, adate-driven prediction model with other raw materials of HTPB propellant, burning rate as input variables and oxidant content as output was established to enhance the product qualification. Machine learning methods, such as random forest regression and support vector regression were adopted and the model hyperparameters were adjusted with the maximum relative error, the root-mean-square error and the average relative error as the performance evaluation indexes of the model. Additionally, different models were screened based on prior engineering knowledge. Finally, The validated model, tested using 8 data sets external to the sample set, demonstrates that the average relative error between the model predictions and the actual test values was 0.75%, and the model adjusted the oxidizer content to make the burn rate closer to the targetburn rate. The data-driven control model is of great significance for the adjustment of formulation parameters in the actual propellant production process.更多还原