【摘要】 针对机枪枪管初速衰减的建模及寿命预测问题,运用贝叶斯推断方法对最小二乘支持向量机(LSSVM)的正则化参数、核函数参数进行优化选择,提出一种基于贝叶斯推断LSSVM的机枪枪管初速衰减建模方法,应用机枪枪管初速衰减试验数据,建立了以环境温度、射击间隔时间、累计射弹量为输入,相对初速为输出的贝叶斯LSSVM机枪枪管初速衰减模型,并与交叉验证的LSSVM及BP神经网络模型进行比较。研究结果表明,基于贝叶斯推断的LSSVM建立的预测模型明显优于上述两种方法,验证了基于贝叶斯推断的LSSVM方法对以初速下降量枪管的寿命评价的有效性。更多还原

【Abstract】 In order to solve the problem of modeling for muzzle velocity degradation and life prediction of machine gun,the Bayesian inference method was used to optimize the regularization parameter and kernel parameter of least squares support vector machines(LSSVM) in this paper,with a modeling method for muzzle velocity degradation of machine gun with LSSVM based on Bayesian inference proposed. A model for muzzle velocity degradation of machine gun based on Bayesian inference LSSVM was established by taking environment temperature,firing interval time and cumulative projectile quantity as the input and relative muzzle velocity as output according to the experimental data of muzzle velocity degradation through comparison with the cross validation LSSVM and BP neural network model.The results show that the prediction model with LSSVM based on Bayesian inference is better than the above two methods. The validity of the LSSVM method based on Bayesian inference was verified in the life evaluation of a gun with initial speed drop.更多还原