农用柴油机排气消声器声学性能优化研究  

Optimization of Acoustic Performance of Exhaust Mufflers for Agricultural Diesel Engines

作　　者：袁文华 李放 伏军 马仪 李光明 黄启科 Yuan Wen-hua;Li fang;Fu Jun;Ma Yi;Li Guang-ming;Huang Qi-ke(ShaoyangUniversity,Shaoyang 422000,China)

机构地区：[1]邵阳学院机械与能源工程学院,湖南邵阳422000

出　　处：《内燃机与配件》2023年第22期71-74,共4页Internal Combustion Engine & Parts

基　　金：邵阳学院研究生科研创新项目(CX2021SY029);国家自然科学基金面上项目(编号:52076141)。

摘　　要：针对农用柴油机在工作过程中产生的排气噪声较大的问题,提出一种直通式并联微穿孔消声器结构,利用近似模型结合多岛遗传算法(multi-island genetic algorithm,MIGA)对消声器进行优化。为了探究结构参数对声学性能的影响规律,得到降噪性能较好的一组消声器结构参数,基于声学有限元方法计算了0-3000Hz的传递损失,采用拉丁超立方抽样对5个设计变量进行采样,建立了径向基神经网络近似模型,并验证其计算精度;最后,结合多岛遗传算法对RBF模型进行优化。结果表明,参数优化设计后的消声器在目标频段内平均传递损失由27.4dB增加到了34.5dB,提高了7.1dB,满足了降噪要求,优化效率得到提高,消声性能得到改善。A direct parallel micro perforated muffler structure is proposed to address the issue of high exhaust noise generated by agricultural diesel engines during operation.The muffler is optimized using an approximate model combined with multi-island genetic algorithm.In order to investigate the influence of structural parameters on acoustic performance and obtain a set of muffler structural parameters with good noise reduction performance,the transfer loss of 0-3000Hz was calculated using acoustic finite element method.Latin hypercube sampling was used to sample 5 design variables,and an RBF neural network approximation model was established and its calculation accuracy was verified;Finally,the RBF model is optimized using a multi-island genetic algorithm.The results show that the average transmission loss of the optimized muffler in the target frequency band has increased from 27.4dB to 34.5dB,an increase of 7.1dB,meeting the noise reduction requirements.The optimization efficiency has been improved,and the noise reduction performance has been improved.

关 键 词：柴油机 消声器 近似模型 传递损失 

分 类 号：TU112.597[建筑科学—建筑理论]