概率盒框架下混合认知不确定声场的响应预测
Response prediction of acoustic field involving hybrid epistemic uncertainties in the probability box framework
-
摘要: 针对含概率盒-证据混合认知不确定参数声场的响应预测问题,提出了一种概率盒框架下的改进区间蒙特卡洛方法。该方法首先将混合认知不确定参数转换为纯概率盒形式,然后结合有限元方法推导出混合认知不确定声场的盖根鲍尔多项式代理模型,再采用蒙特卡洛方法求解代理模型得到声压响应。以含概率盒-证据混合认知不确定参数的二维管道声场模型和卡车乘客舱声腔模型为例,计算结果表明混合认知不确定参数影响下的声压响应为概率盒形式,其包括声压响应极值和相应的概率信息,并且所提方法较常规混合离散方法效率更优,较基于一阶摄动法的区间蒙特卡洛方法准确性更高。研究结果表明:所提方法可以有效预测混合认知不确定声场的声压响应,并可进行声学性能的风险和保守估计。Abstract: In the probability box framework,an Improved Interval Monte Carlo Method(IIMCM) is proposed for response prediction of acoustic field involving hybrid epistemic uncertainties.Firstly,p-box variables and evidence variables are firstly transformed into unified p-box form.Then,the Gegenbauer polynomial surrogate model of acoustic filed with hybrid variables is conducted by combing the finite element method,the sound pressure response can be obtained by solving the model through the Monte Carlo method.Numerical examples of a 2 D acoustic tube and an acoustic cavity of a truck passenger compartment involving hybrid p-box and evidence variables are investigated.The calculation results show that the sound pressure is in the form of probability box,which includes the extreme value of sound pressure and the corresponding probability information.Besides,the proposed method has higher efficiency than conventional discretization method and higher accuracy than the interval Monte Carlo method based on first order matrix decomposition perturbation method.The results of research prove that the proposed method can effectively predict the sound pressure of the acoustic field with hybrid epistemic uncertain parameters,and can make the risk and conservative estimation of the acoustic performance.