Applied Acoustics最新文献_第8页

Research progress in bird sounds recognition based on acoustic monitoring technology: A systematic review 基于声学监测技术的鸟类声音识别研究进展：系统回顾

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-10-03 DOI: 10.1016/j.apacoust.2024.110285

Daidai Liu, Hanguang Xiao, Kai Chen

Bird sound contains rich ecological information, and its related research results can be applied to animal behavior analysis, natural information collection and ecological environment monitoring. Since the early manual monitoring methods, many researchers have continuously innovated and improved the bird sounds recognition technology to overcome the long-standing drawbacks of long cycle time, high cost and poor effectiveness. These developments make bird sounds recognition a highly interesting, but also highly challenging research topic. Acoustic monitoring technology plays a vital role in the automatic recognition of bird sounds. With the popularization of acoustic monitoring technology, the technical routes based on traditional recognition models and neural networks have increased sharply, which has greatly promoted the development of bird sounds recognition. In view of these main technical routes, this paper summarized the research status of bird sounds recognition, provided a summary table of a variety of bird sound sample datasets, introduced the application evolution of various recognition technologies, and analyzed its open challenges. Meanwhile, this paper also cited the published experimental exploration on the improvement of deep learning networks. In general, this paper gives a comprehensive overview of the research process of bird sounds recognition based on acoustic monitoring technology, which has important theoretical and practical value to promote the development of bird sounds recognition technology, and provides a valuable reference for future related research.

鸟声蕴含着丰富的生态信息，其相关研究成果可应用于动物行为分析、自然信息采集和生态环境监测等领域。从早期的人工监测方法开始，许多研究人员不断创新和改进鸟声识别技术，以克服长期以来存在的周期长、成本高、效果差等弊端。这些发展使鸟声识别成为一个非常有趣但也极具挑战性的研究课题。声学监测技术在鸟声自动识别中起着至关重要的作用。随着声学监测技术的普及，基于传统识别模型和神经网络的技术路线急剧增加，极大地推动了鸟声识别的发展。针对这些主要技术路线，本文总结了鸟声识别的研究现状，提供了多种鸟声样本数据集汇总表，介绍了各种识别技术的应用演进，并分析了其面临的挑战。同时，本文还引用了已发表的关于改进深度学习网络的实验探索。总的来说，本文全面概述了基于声学监测技术的鸟声识别研究过程，对推动鸟声识别技术的发展具有重要的理论和实践价值，并为今后的相关研究提供了有价值的参考。

{"title":"Research progress in bird sounds recognition based on acoustic monitoring technology: A systematic review","authors":"Daidai Liu, Hanguang Xiao, Kai Chen","doi":"10.1016/j.apacoust.2024.110285","DOIUrl":"10.1016/j.apacoust.2024.110285","url":null,"abstract":"<div><div>Bird sound contains rich ecological information, and its related research results can be applied to animal behavior analysis, natural information collection and ecological environment monitoring. Since the early manual monitoring methods, many researchers have continuously innovated and improved the bird sounds recognition technology to overcome the long-standing drawbacks of long cycle time, high cost and poor effectiveness. These developments make bird sounds recognition a highly interesting, but also highly challenging research topic. Acoustic monitoring technology plays a vital role in the automatic recognition of bird sounds. With the popularization of acoustic monitoring technology, the technical routes based on traditional recognition models and neural networks have increased sharply, which has greatly promoted the development of bird sounds recognition. In view of these main technical routes, this paper summarized the research status of bird sounds recognition, provided a summary table of a variety of bird sound sample datasets, introduced the application evolution of various recognition technologies, and analyzed its open challenges. Meanwhile, this paper also cited the published experimental exploration on the improvement of deep learning networks. In general, this paper gives a comprehensive overview of the research process of bird sounds recognition based on acoustic monitoring technology, which has important theoretical and practical value to promote the development of bird sounds recognition technology, and provides a valuable reference for future related research.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110285"},"PeriodicalIF":3.4,"publicationDate":"2024-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142419065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multimodal fusion fault diagnosis method under noise interference 噪声干扰下的多模态融合故障诊断方法

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-10-02 DOI: 10.1016/j.apacoust.2024.110301

Zhi Qiu , Shanfei Fan , Haibo Liang, Jincai Liu

In practical industrial production environments, the collection of fault signals is often accompanied by significant background noise. The presence of substantial noise makes feature extraction from fault signals very challenging, thereby reducing fault diagnosis performance. To address this issue, this paper proposes a multimodal fusion fault diagnosis method based on a multiscale stacked denoising autoencoder and dual-branch feature fusion network (MSSDAE-DBFFN). First, the noisy vibration signals are denoised using the MSSDAE. Then, the denoised vibration signals are divided into two branches for feature extraction and fusion. In one branch, the vibration signals are converted into gramian angular summation field (GASF) images using the GASF, and feature extraction is performed with a multiscale convolutional network. In the other branch, the waveforms are subjected to feature extraction using a wavelet scattering network. Finally, the fused features are sent to a classifier to complete the fault diagnosis task. To demonstrate the effectiveness of the proposed method, it is compared with four different denoising methods and five different classification methods across two datasets. The experimental results show that MSSDAE-DBFFN outperforms the other methods in both denoising and classification across five different signal-to-noise ratios (SNR). At an SNR of −10 dB, the SNRs after denoising are 4.582 dB and 5.489 dB, respectively, while the accuracy rates are 89.33 % and 91.67 %, respectively.

在实际的工业生产环境中，故障信号的采集往往伴随着大量的背景噪声。大量噪声的存在使得从故障信号中提取特征非常具有挑战性，从而降低了故障诊断性能。针对这一问题，本文提出了一种基于多尺度堆叠去噪自动编码器和双分支特征融合网络（MSSDAE-DBFFN）的多模态融合故障诊断方法。首先，使用 MSSDAE 对噪声振动信号进行去噪。然后，去噪后的振动信号被分为两个分支，用于特征提取和融合。在一个分支中，使用 GASF 将振动信号转换成格兰角加和场（GASF）图像，并使用多尺度卷积网络进行特征提取。在另一个分支中，使用小波散射网络对波形进行特征提取。最后，将融合后的特征发送给分类器，以完成故障诊断任务。为了证明所提方法的有效性，我们在两个数据集上将其与四种不同的去噪方法和五种不同的分类方法进行了比较。实验结果表明，在五种不同的信噪比（SNR）下，MSSDAE-DBFFN 在去噪和分类方面都优于其他方法。在信噪比为 -10 dB 时，去噪后的信噪比分别为 4.582 dB 和 5.489 dB，准确率分别为 89.33 % 和 91.67 %。

{"title":"Multimodal fusion fault diagnosis method under noise interference","authors":"Zhi Qiu , Shanfei Fan , Haibo Liang, Jincai Liu","doi":"10.1016/j.apacoust.2024.110301","DOIUrl":"10.1016/j.apacoust.2024.110301","url":null,"abstract":"<div><div>In practical industrial production environments, the collection of fault signals is often accompanied by significant background noise. The presence of substantial noise makes feature extraction from fault signals very challenging, thereby reducing fault diagnosis performance. To address this issue, this paper proposes a multimodal fusion fault diagnosis method based on a multiscale stacked denoising autoencoder and dual-branch feature fusion network (MSSDAE-DBFFN). First, the noisy vibration signals are denoised using the MSSDAE. Then, the denoised vibration signals are divided into two branches for feature extraction and fusion. In one branch, the vibration signals are converted into gramian angular summation field (GASF) images using the GASF, and feature extraction is performed with a multiscale convolutional network. In the other branch, the waveforms are subjected to feature extraction using a wavelet scattering network. Finally, the fused features are sent to a classifier to complete the fault diagnosis task. To demonstrate the effectiveness of the proposed method, it is compared with four different denoising methods and five different classification methods across two datasets. The experimental results show that MSSDAE-DBFFN outperforms the other methods in both denoising and classification across five different signal-to-noise ratios (SNR). At an SNR of −10 dB, the SNRs after denoising are 4.582 dB and 5.489 dB, respectively, while the accuracy rates are 89.33 % and 91.67 %, respectively.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110301"},"PeriodicalIF":3.4,"publicationDate":"2024-10-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142419017","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Geometrical acoustic modelling of occupied acoustic conditions in mosques: Application to a case study 清真寺占用声学条件的几何声学建模：应用于案例研究

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-10-01 DOI: 10.1016/j.apacoust.2024.110323

Francesco Martellotta , Mohamed Ladaoui Benferhat , Chiara Rubino , Abdelouahab Bouttout , Samira Debache Benzagouta

Old religious buildings represent an essential cultural heritage whatever the country or the religion they belong to. Thanks to many researches carried out in the last years, their acoustics is now considered part of this heritage. However, for practical reasons, their acoustic characterization is often made under unoccupied conditions, while, given the frequent use of hard reflecting surfaces, the occupied conditions may differ significantly. Geometrical acoustics may represent, if properly used, a valid tool to simulate how sound propagates in an occupied space, allowing to investigate the effect on the full set of acoustic parameters. Occupancy in mosques may be more challenging to simulate than in other spaces because of the different postures of the worshippers and the usually high absorption that they introduce because of high density of occupants. To correctly simulate such effects, a specific modelling approach has been proposed starting from reverberant chamber measurements and validating them against on-site measurements. Using the proposed method, the effect of occupancy in the Jedid Mosque in Algiers, which was built in 1660, in a typical Ottoman style, and later restored in 1855, was studied. The mosque was chosen because it is large and reverberant to allow a better appreciation of the variations due to occupancy. The geometrical acoustic model was first carefully calibrated against measurements in unoccupied conditions, which also pointed out a clearly non-diffuse behaviour in the space, and, finally the occupancy was added. Results showed that due to the strong concentration of absorbing elements on the floor, where carpets already contributed to absorb sound, the occupancy mostly affected reverberation parameters, while clarity for speech remained poor.

无论属于哪个国家或宗教，古老的宗教建筑都是重要的文化遗产。得益于近些年来开展的多项研究，这些建筑的声学特性现已被视为文化遗产的一部分。然而，由于实际原因，其声学特性通常是在无人居住的条件下进行的，而由于经常使用硬反射面，有人居住的条件可能会有很大不同。如果使用得当，几何声学可能是模拟声音如何在有人居住的空间传播的有效工具，可以研究对全套声学参数的影响。与其他空间相比，模拟清真寺内的占用情况可能更具挑战性，因为朝拜者的姿势各不相同，而且由于占用者密度高，通常会产生较高的吸声。为了正确模拟这种影响，我们提出了一种特定的建模方法，该方法从混响室测量开始，并根据现场测量结果进行验证。阿尔及尔的杰迪德清真寺建于 1660 年，具有典型的奥斯曼风格，后于 1855 年进行了修复。之所以选择这座清真寺，是因为它规模大、混响强，可以更好地了解占用造成的变化。首先，对几何声学模型进行了仔细的校准，并与无人居住情况下的测量结果进行了对比，结果表明该空间具有明显的非扩散特性。结果表明，由于吸音元件高度集中在地板上，而地板上的地毯已经起到了吸音作用，因此占用空间主要影响混响参数，而语音清晰度仍然很差。

{"title":"Geometrical acoustic modelling of occupied acoustic conditions in mosques: Application to a case study","authors":"Francesco Martellotta , Mohamed Ladaoui Benferhat , Chiara Rubino , Abdelouahab Bouttout , Samira Debache Benzagouta","doi":"10.1016/j.apacoust.2024.110323","DOIUrl":"10.1016/j.apacoust.2024.110323","url":null,"abstract":"<div><div>Old religious buildings represent an essential cultural heritage whatever the country or the religion they belong to. Thanks to many researches carried out in the last years, their acoustics is now considered part of this heritage. However, for practical reasons, their acoustic characterization is often made under unoccupied conditions, while, given the frequent use of hard reflecting surfaces, the occupied conditions may differ significantly. Geometrical acoustics may represent, if properly used, a valid tool to simulate how sound propagates in an occupied space, allowing to investigate the effect on the full set of acoustic parameters. Occupancy in mosques may be more challenging to simulate than in other spaces because of the different postures of the worshippers and the usually high absorption that they introduce because of high density of occupants. To correctly simulate such effects, a specific modelling approach has been proposed starting from reverberant chamber measurements and validating them against on-site measurements. Using the proposed method, the effect of occupancy in the Jedid Mosque in Algiers, which was built in 1660, in a typical Ottoman style, and later restored in 1855, was studied. The mosque was chosen because it is large and reverberant to allow a better appreciation of the variations due to occupancy. The geometrical acoustic model was first carefully calibrated against measurements in unoccupied conditions, which also pointed out a clearly non-diffuse behaviour in the space, and, finally the occupancy was added. Results showed that due to the strong concentration of absorbing elements on the floor, where carpets already contributed to absorb sound, the occupancy mostly affected reverberation parameters, while clarity for speech remained poor.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110323"},"PeriodicalIF":3.4,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142419016","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Acoustic comfort in primary- and nursery-school canteens: From measurements to recommendations 小学和幼儿园食堂的声学舒适度：从测量到建议

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-10-01 DOI: 10.1016/j.apacoust.2024.110324

Francesco Martellotta , Dario D'Orazio , Deborah De Carolis , Stefania Liuzzi , Chiara Rubino

Canteens usually have critical acoustic conditions resulting from the need to maximize the number of occupants while minimizing volume. Thus, in the absence of specific sound absorbing treatments, very high sound pressure levels are usually observed resulting in significant impairment of communication (with increased vocal effort of speakers and reduced speech intelligibility), and dangerously high exposure levels for workers. The present paper reports acoustic measurements carried out in a nursery school canteen having a volume of 212 m³ and seating about 50 children, and two primary school canteens having volumes of 656 m³ (seating 150 children) and 367 m³ (seating 107 children). Reverberation time was measured in each room as well as sound pressure levels during peak occupation (averaged over 15-minute intervals), resulting in A-weighted sound pressure levels spanning between 81 dB (in the nursery school) and 90 dB in the primary schools. Starting from the observed values, considerations about the group-size of the occupants as a function of age were made, and recommendations were finally given to guide the acoustic correction of similar spaces.

食堂的声学条件通常很苛刻，因为既要尽量增加用餐人数，又要尽量减小音量。因此，如果没有特殊的吸声处理，通常会观察到很高的声压级，导致严重的交流障碍（说话者发声费力，语言清晰度降低），并对工人造成危险的高暴露水平。本文报告了在一个容积为 212 立方米、可容纳约 50 名儿童的幼儿园食堂和两个容积分别为 656 立方米（可容纳 150 名儿童）和 367 立方米（可容纳 107 名儿童）的小学食堂进行的声学测量结果。测量了每个房间的混响时间以及高峰占用期间的声压级（15 分钟间隔的平均值），得出的 A 计权声压级介于 81 分贝（幼儿园）和 90 分贝（小学）之间。从观测值出发，考虑了作为年龄函数的居住者群体规模，并最终提出了指导类似空间声学校正的建议。

{"title":"Acoustic comfort in primary- and nursery-school canteens: From measurements to recommendations","authors":"Francesco Martellotta , Dario D'Orazio , Deborah De Carolis , Stefania Liuzzi , Chiara Rubino","doi":"10.1016/j.apacoust.2024.110324","DOIUrl":"10.1016/j.apacoust.2024.110324","url":null,"abstract":"<div><div>Canteens usually have critical acoustic conditions resulting from the need to maximize the number of occupants while minimizing volume. Thus, in the absence of specific sound absorbing treatments, very high sound pressure levels are usually observed resulting in significant impairment of communication (with increased vocal effort of speakers and reduced speech intelligibility), and dangerously high exposure levels for workers. The present paper reports acoustic measurements carried out in a nursery school canteen having a volume of 212 m<sup>3</sup> and seating about 50 children, and two primary school canteens having volumes of 656 m<sup>3</sup> (seating 150 children) and 367 m<sup>3</sup> (seating 107 children). Reverberation time was measured in each room as well as sound pressure levels during peak occupation (averaged over 15-minute intervals), resulting in A-weighted sound pressure levels spanning between 81 dB (in the nursery school) and 90 dB in the primary schools. Starting from the observed values, considerations about the group-size of the occupants as a function of age were made, and recommendations were finally given to guide the acoustic correction of similar spaces.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110324"},"PeriodicalIF":3.4,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142419066","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Reverberation room design optimisation for low-frequency diffuse sound absorption testing 用于低频扩散吸声测试的混响室优化设计

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-30 DOI: 10.1016/j.apacoust.2024.110311

Paul Didier , Cédric Van hoorickx , Edwin P.B. Reynders

The reproducibility of sound absorption testing with the reverberation room method is a long-standing concern. Absorptive samples induce directionality in the nearfield, while the farfield depends on the room geometry below the Schroeder frequency. Nevertheless, when properly accounting for nearfield effects, the theoretical diffuse absorption coefficient of a sample still represents its average performance across an ensemble of different rooms, even at very low frequencies. Recent research found that particular reverberation room designs allow for an accurate measurement of the diffuse sound absorption coefficient of highly absorptive samples at low frequencies. Pinpointing such designs hence opens up a possibility to sustainably improve the low-frequency reproducibility of sound absorption testing in reverberation rooms. The present paper introduces a numerical optimisation framework that serves this purpose. Specific room shapes are parametrised and the geometrical room parameters are optimised so as to minimise the difference between the measured and the diffuse absorption coefficient under appropriate constraints. The sound absorption testing of a sample in a particular reverberation room is numerically simulated using a method that is both accurate and computationally efficient at low frequencies. The diffuse absorption is computed with a hybrid deterministic-statistical energy analysis approach that accounts for the detailed absorber properties, geometry, and boundary conditions, as well as the nearfield effects. The methodology is applied to both cuboidal and hexahedral room shapes. Certain optimised designs are found not only to provide an excellent match for the absorber that was used during the optimisation, but they also maintain their performance across a range of absorptive samples. Additionally, potential geometrical deviations are found to be well tolerated by these reverberation room designs.

使用混响室方法进行吸声测试的可重复性是一个长期关注的问题。吸声样本在近场具有方向性，而远场则取决于施罗德频率以下的房间几何形状。尽管如此，在适当考虑近场效应的情况下，样品的理论扩散吸收系数仍然代表了不同房间的平均性能，即使在非常低的频率下也是如此。最近的研究发现，特定的混响室设计可以精确测量高吸声样本在低频时的扩散吸声系数。因此，准确定位这些设计为持续改善混响室吸声测试的低频重现性提供了可能。本文引入了一个数值优化框架来实现这一目的。对特定的房间形状进行参数化，并对房间几何参数进行优化，以便在适当的约束条件下最大限度地减小测量值与扩散吸声系数之间的差异。在特定混响室中对样本进行吸声测试时，会使用一种在低频下既精确又高效的计算方法进行数值模拟。采用确定性-统计性混合能量分析方法计算扩散吸声，该方法考虑了详细的吸声特性、几何形状、边界条件以及近场效应。该方法适用于立方体和六面体房间形状。结果发现，某些优化设计不仅能与优化过程中使用的吸收器完美匹配，还能在一系列吸收样本中保持其性能。此外，这些混响室设计还能很好地容忍潜在的几何偏差。

{"title":"Reverberation room design optimisation for low-frequency diffuse sound absorption testing","authors":"Paul Didier , Cédric Van hoorickx , Edwin P.B. Reynders","doi":"10.1016/j.apacoust.2024.110311","DOIUrl":"10.1016/j.apacoust.2024.110311","url":null,"abstract":"<div><div>The reproducibility of sound absorption testing with the reverberation room method is a long-standing concern. Absorptive samples induce directionality in the nearfield, while the farfield depends on the room geometry below the Schroeder frequency. Nevertheless, when properly accounting for nearfield effects, the theoretical diffuse absorption coefficient of a sample still represents its average performance across an ensemble of different rooms, even at very low frequencies. Recent research found that particular reverberation room designs allow for an accurate measurement of the diffuse sound absorption coefficient of highly absorptive samples at low frequencies. Pinpointing such designs hence opens up a possibility to sustainably improve the low-frequency reproducibility of sound absorption testing in reverberation rooms. The present paper introduces a numerical optimisation framework that serves this purpose. Specific room shapes are parametrised and the geometrical room parameters are optimised so as to minimise the difference between the measured and the diffuse absorption coefficient under appropriate constraints. The sound absorption testing of a sample in a particular reverberation room is numerically simulated using a method that is both accurate and computationally efficient at low frequencies. The diffuse absorption is computed with a hybrid deterministic-statistical energy analysis approach that accounts for the detailed absorber properties, geometry, and boundary conditions, as well as the nearfield effects. The methodology is applied to both cuboidal and hexahedral room shapes. Certain optimised designs are found not only to provide an excellent match for the absorber that was used during the optimisation, but they also maintain their performance across a range of absorptive samples. Additionally, potential geometrical deviations are found to be well tolerated by these reverberation room designs.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110311"},"PeriodicalIF":3.4,"publicationDate":"2024-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142359504","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A lightweight framework for unsupervised anomalous sound detection based on selective learning of time-frequency domain features 基于时频域特征选择性学习的无监督异常声音检测轻量级框架

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-30 DOI: 10.1016/j.apacoust.2024.110308

Yawei Wang , Qiaoling Zhang , Weiwei Zhang , Yi Zhang

For industrial anomalous sound detection (ASD), self-supervised methods have achieved significant detection performance in many cases. Nevertheless, these methods typically rely on the availability of external auxiliary information, and they may not work when such information are not feasible. Unsupervised methods do not leverage auxiliary information, whereas they usually obtained lower detection performance compared to self-supervised ones. Though some unsupervised methods have shown potential performance improvements, they are at the cost of complex implementation or large model sizes. As to the issues, this paper presents an unsupervised ASD method based on spectrogram frames selection (SFS) and AutoEncoder for Frequency-feature Selection (AEFS), called SFS-AEFS. First, SFS is developed based upon the temporal characteristics of machine sounds, which aims to select spectrogram frames (SFs) that contains the primary sound information while discarding the portions that are affected by noises or interferences or do not contain the target sound. Next, AEFS is developed by introducing a Scaling Gate (SG) after AE. For the selected SF features, AEFS aims to selectively enhance the mode learning of partial frequency dimensions and weaken the rest ones. Comparative experiments with the current ASD methods were made on the DCASE 2020 Challenge Task2 dataset. The related results demonstrate that our method achieved the best performance among all relevant unsupervised methods and is comparable to the current SOTA self-supervised methods. Moreover, our method is lightweight with model parameters being only 0.08MB.

对于工业异常声音检测（ASD），自监督方法在许多情况下都取得了显著的检测效果。然而，这些方法通常依赖于外部辅助信息的可用性，当这些信息不可行时，这些方法可能会失效。无监督方法不利用辅助信息，但其检测性能通常低于自监督方法。虽然一些无监督方法显示出潜在的性能改进，但它们的代价是复杂的实现或庞大的模型规模。针对这些问题，本文提出了一种基于频谱图帧选择（SFS）和频率特性选择自动编码器（AEFS）的无监督 ASD 方法，称为 SFS-AEFS。首先，SFS 是根据机器声音的时间特性开发的，旨在选择包含主要声音信息的频谱图帧 (SF)，同时剔除受噪声、干扰影响或不包含目标声音的部分。接下来，在 AE 之后引入缩放门（SG），开发出 AEFS。对于选定的 SF 特征，AEFS 的目的是选择性地增强部分频率维度的模式学习，削弱其余维度的模式学习。在 DCASE 2020 Challenge Task2 数据集上进行了与当前 ASD 方法的对比实验。相关结果表明，在所有相关的无监督方法中，我们的方法取得了最佳性能，与当前的 SOTA 自监督方法不相上下。此外，我们的方法非常轻便，模型参数仅为 0.08MB。

{"title":"A lightweight framework for unsupervised anomalous sound detection based on selective learning of time-frequency domain features","authors":"Yawei Wang , Qiaoling Zhang , Weiwei Zhang , Yi Zhang","doi":"10.1016/j.apacoust.2024.110308","DOIUrl":"10.1016/j.apacoust.2024.110308","url":null,"abstract":"<div><div>For industrial anomalous sound detection (ASD), self-supervised methods have achieved significant detection performance in many cases. Nevertheless, these methods typically rely on the availability of external auxiliary information, and they may not work when such information are not feasible. Unsupervised methods do not leverage auxiliary information, whereas they usually obtained lower detection performance compared to self-supervised ones. Though some unsupervised methods have shown potential performance improvements, they are at the cost of complex implementation or large model sizes. As to the issues, this paper presents an unsupervised ASD method based on spectrogram frames selection (SFS) and AutoEncoder for Frequency-feature Selection (AEFS), called SFS-AEFS. First, SFS is developed based upon the temporal characteristics of machine sounds, which aims to select spectrogram frames (SFs) that contains the primary sound information while discarding the portions that are affected by noises or interferences or do not contain the target sound. Next, AEFS is developed by introducing a Scaling Gate (SG) after AE. For the selected SF features, AEFS aims to selectively enhance the mode learning of partial frequency dimensions and weaken the rest ones. Comparative experiments with the current ASD methods were made on the DCASE 2020 Challenge Task2 dataset. The related results demonstrate that our method achieved the best performance among all relevant unsupervised methods and is comparable to the current SOTA self-supervised methods. Moreover, our method is lightweight with model parameters being only 0.08MB.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110308"},"PeriodicalIF":3.4,"publicationDate":"2024-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142359502","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Distributional Green’s functions for the vibrations of multi-cracked Timoshenko beams 多裂缝季莫申科梁振动的分布格林函数

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-30 DOI: 10.1016/j.apacoust.2024.110302

Ilaria Fiore, Francesco Cannizzaro, Salvatore Caddemi, Ivo Caliò

In this study the forced vibrations of cracked beams in presence of damping are analysed. The adopted beam model is in accordance with the Timoshenko beam model and the presence of multiple bending and shear concentrated flexibilities, commonly used to model cracks, is accounted for. The strong discontinuities derived by the localised flexibilities are dealt with by means of a distributional approach avoiding the need of enforcing continuity conditions at the discontinuous sections. First, the exact Green’s functions, that is the steady-state response in the case of concentrated harmonic loads, are obtained via the presented distributional approach. The presented exact solutions are a computationally advantageous evaluation of the steady state response alternative to the direct time integration, as well as to a beam span sub-division. In addition, the presented distributional Green’s functions are employed to evaluate the response of multi-cracked beams subjected to arbitrary loading conditions (i.e. generic spatial distribution and time dependency), via convolution integral equation combined with an appropriate frequency domain analysis.

本研究分析了存在阻尼的裂缝梁的受迫振动。所采用的梁模型与季莫申科梁模型一致，并考虑了通常用于裂缝建模的多重弯曲和剪切集中挠性的存在。局部挠性产生的强不连续性通过分布式方法处理，避免了在不连续性部分强制执行连续性条件的需要。首先，通过提出的分布式方法获得精确的格林函数，即集中谐波载荷情况下的稳态响应。与直接时间积分和梁跨度细分相比，所提出的精确解法在稳态响应评估方面具有计算优势。此外，通过卷积积分方程结合适当的频域分析，所提出的分布式格林函数还可用于评估受到任意加载条件（即通用空间分布和时间相关性）影响的多裂缝梁的响应。

{"title":"Distributional Green’s functions for the vibrations of multi-cracked Timoshenko beams","authors":"Ilaria Fiore, Francesco Cannizzaro, Salvatore Caddemi, Ivo Caliò","doi":"10.1016/j.apacoust.2024.110302","DOIUrl":"10.1016/j.apacoust.2024.110302","url":null,"abstract":"<div><div>In this study the forced vibrations of cracked beams in presence of damping are analysed. The adopted beam model is in accordance with the Timoshenko beam model and the presence of multiple bending and shear concentrated flexibilities, commonly used to model cracks, is accounted for. The strong discontinuities derived by the localised flexibilities are dealt with by means of a distributional approach avoiding the need of enforcing continuity conditions at the discontinuous sections. First, the exact Green’s functions, that is the steady-state response in the case of concentrated harmonic loads, are obtained via the presented distributional approach. The presented exact solutions are a computationally advantageous evaluation of the steady state response alternative to the direct time integration, as well as to a beam span sub-division. In addition, the presented distributional Green’s functions are employed to evaluate the response of multi-cracked beams subjected to arbitrary loading conditions (i.e. generic spatial distribution and time dependency), via convolution integral equation combined with an appropriate frequency domain analysis.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110302"},"PeriodicalIF":3.4,"publicationDate":"2024-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142359596","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Applying Michelson-interferometer based optical fiber sensors for active control of road noise in a car 应用基于迈克尔逊干涉仪的光纤传感器主动控制汽车路噪

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-29 DOI: 10.1016/j.apacoust.2024.110316

Chenguang Wang , Feng Li , Pengju Zhang , Xiaojun Qiu , Weikang Jiang , Sheng Wu , Pinxi Mo

Active noise control (ANC) technology is a promising solution for reducing road noise in cars, where the quality of reference signals is critical for improving the performance. In this paper, an optical fiber sensor based on Michelson interferometer is proposed to obtain the vibration reference signal for active control of road noise in a car. First, the principle of this optical fiber sensor is introduced, then the feasibility of using optical fiber sensor for active control of sound radiation from a panel of an enclosure is verified, where the spectrum, coherence and impulse response of the optical fiber sensor are reported and compared with that of an accelerometer. Finally, the optical fiber sensor is applied in a car for active control of road noise and the simulation results based on the measured data are presented. The research demonstrates the potential of using the proposed optical fiber sensor for active control of road noise in a car by discussing its advantages and weakness.

主动噪声控制（ANC）技术是降低汽车道路噪声的一种有前途的解决方案，而参考信号的质量对于提高性能至关重要。本文提出了一种基于迈克尔逊干涉仪的光纤传感器，用于获取振动参考信号，实现对汽车路噪的主动控制。首先介绍了这种光纤传感器的原理，然后验证了使用光纤传感器对来自外壳面板的声音辐射进行主动控制的可行性，报告了光纤传感器的频谱、相干性和脉冲响应，并与加速度计的频谱、相干性和脉冲响应进行了比较。最后，将光纤传感器应用于汽车，对道路噪声进行主动控制，并展示了基于测量数据的模拟结果。该研究通过讨论所提出的光纤传感器的优点和缺点，展示了将其用于汽车路噪主动控制的潜力。

引用次数: 0

Interior noise characteristics, source identification and its quantification contribution of 400 km/h high-speed train in different operating environments 时速 400 公里高速列车在不同运行环境下的车内噪声特性、声源识别及其量化贡献

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-28 DOI: 10.1016/j.apacoust.2024.110321

Xiaodan Lan , Jian Han , Jie Zhang , Xinbiao Xiao , Laixian Peng , Yuxuan Zhao

Noise level is one of the core technical indicators of high-speed trains, especially for a speed of 400 km/h. Noise source identification and its contribution quantification are key for interior noise control. This article takes a certain high-speed train running up to 400 km/h as the research object. Firstly, based on the interior noise spectrum and experimental results of source identification based on spherical harmonic function, the main source locations and energy distribution inside the train are determined. Secondly, through polynomial fitting, the correlation between vibration and noise in various areas inside and outside the train and speed is determined. Subsequently, by calculating the energy contribution through area integration, the noise contribution of each interior region is determined, and then the variation law of the noise contribution of each region with the speed is determined, and the quantitative contributions of various areas inside the train are given when the high-speed train is running at 400 km/h. Finally, the vibration and noise inside the train, aerodynamic noise on the train body surface, and noise in the bogie area in open lines and tunnel operating environments are compared and analyzed, as well as their variation laws with speed. The main noise sources inside the train under two types of operating environments are identified, and the contribution rates of noise sources in different areas inside the train are analyzed, further studying the interior noise and vibration transmission characteristics. The results show that, when the high-speed train is running at 400 km/h in open line operating environment, the significant frequency range of interior noise is 40 Hz ∼ 2000 Hz, and dominant interior noise sources are located in the roof and floor. In the tunnel operating environment, the significant frequency range of interior noise is 160 Hz ∼ 1000 Hz, and the primary sources of interior noise are predominantly located in the left window and floor. For the sidewall area, the interior noise comes mainly from the vibration of the inner sidewall when in open line operating environment, while in tunnel operating environment, the interior noise comes mainly from the aerodynamic excitation of the body surface.

噪音水平是高速列车的核心技术指标之一，尤其是在时速 400 公里的情况下。噪声源识别及其贡献量化是车内噪声控制的关键。本文以时速 400 公里的某高速列车为研究对象。首先，根据车内噪声频谱和基于球谐函数的声源识别实验结果，确定列车内部的主要声源位置和能量分布。其次，通过多项式拟合，确定列车内外各区域振动和噪声与速度的相关性。随后，通过区域积分计算能量贡献率，确定内部各区域的噪声贡献率，进而确定各区域噪声贡献率随速度的变化规律，给出高速列车以 400 km/h 运行时，列车内部各区域的定量贡献率。最后，对比分析了开放线路和隧道运行环境下列车内部的振动和噪声、车体表面的空气动力噪声以及转向架区域的噪声，以及它们随速度的变化规律。确定了两种运行环境下列车内部的主要噪声源，分析了噪声源在列车内部不同区域的贡献率，进一步研究了列车内部噪声和振动的传播特性。结果表明，高速列车在开放线路运行环境下以 400 km/h 运行时，车内噪声的显著频率范围为 40 Hz ∼ 2000 Hz，主要的车内噪声源位于车顶和地板。在隧道运行环境中，室内噪声的重要频率范围为 160 Hz ∼ 1000 Hz，室内噪声的主要声源主要位于左侧窗户和地板。对于侧壁区域，在开放线路运行环境下，车内噪声主要来自内侧壁的振动，而在隧道运行环境下，车内噪声主要来自车身表面的空气动力激励。

{"title":"Interior noise characteristics, source identification and its quantification contribution of 400 km/h high-speed train in different operating environments","authors":"Xiaodan Lan , Jian Han , Jie Zhang , Xinbiao Xiao , Laixian Peng , Yuxuan Zhao","doi":"10.1016/j.apacoust.2024.110321","DOIUrl":"10.1016/j.apacoust.2024.110321","url":null,"abstract":"<div><div>Noise level is one of the core technical indicators of high-speed trains, especially for a speed of 400 km/h. Noise source identification and its contribution quantification are key for interior noise control. This article takes a certain high-speed train running up to 400 km/h as the research object. Firstly, based on the interior noise spectrum and experimental results of source identification based on spherical harmonic function, the main source locations and energy distribution inside the train are determined. Secondly, through polynomial fitting, the correlation between vibration and noise in various areas inside and outside the train and speed is determined. Subsequently, by calculating the energy contribution through area integration, the noise contribution of each interior region is determined, and then the variation law of the noise contribution of each region with the speed is determined, and the quantitative contributions of various areas inside the train are given when the high-speed train is running at 400 km/h. Finally, the vibration and noise inside the train, aerodynamic noise on the train body surface, and noise in the bogie area in open lines and tunnel operating environments are compared and analyzed, as well as their variation laws with speed. The main noise sources inside the train under two types of operating environments are identified, and the contribution rates of noise sources in different areas inside the train are analyzed, further studying the interior noise and vibration transmission characteristics. The results show that, when the high-speed train is running at 400 km/h in open line operating environment, the significant frequency range of interior noise is 40 Hz ∼ 2000 Hz, and dominant interior noise sources are located in the roof and floor. In the tunnel operating environment, the significant frequency range of interior noise is 160 Hz ∼ 1000 Hz, and the primary sources of interior noise are predominantly located in the left window and floor. For the sidewall area, the interior noise comes mainly from the vibration of the inner sidewall when in open line operating environment, while in tunnel operating environment, the interior noise comes mainly from the aerodynamic excitation of the body surface.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"228 ","pages":"Article 110321"},"PeriodicalIF":3.4,"publicationDate":"2024-09-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142359595","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Motion coprime array-based DOA estimation considering phase disturbance of sensor array 考虑传感器阵列相位干扰的基于运动共轭阵列的 DOA 估算

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2024-09-27 DOI: 10.1016/j.apacoust.2024.110313

Ke Liu , Chengfeng Zou , Junda Ma

To address the phase disturbance issue faced by sensor arrays in practical applications, a cascaded deep convolutional neural network structure is proposed to achieve direction-of-arrival (DOA) estimation for motion coprime arrays. Firstly, the synthesized covariance matrix obtained after motion is inputted into the first-level network for estimating the phase disturbance matrix. Then, we analyze the impact of phase perturbation on the synthesized covariance matrix and utilize the estimated disturbance phase to obtain an undisturbed synthesized covariance matrix. Finally, after phase compensation, the synthesized covariance matrix performs DOA estimation through the second-level network. Furthermore, to acquire three times the virtual unique lags of the coprime array, the synthesis condition about moving distance and unique lags is derived. The proposed method is shown to be effective and superior through the experiment results.

针对传感器阵列在实际应用中面临的相位干扰问题，提出了一种级联深度卷积神经网络结构，以实现运动共轭阵列的到达方向（DOA）估计。首先，将运动后获得的合成协方差矩阵输入一级网络，用于估计相位干扰矩阵。然后，分析相位扰动对合成协方差矩阵的影响，并利用估算出的扰动相位获得不受扰动的合成协方差矩阵。最后，经过相位补偿后，合成协方差矩阵通过二级网络执行 DOA 估计。此外，为了获得三次共轭阵列的虚拟唯一滞后，还推导出了移动距离和唯一滞后的合成条件。实验结果表明，所提出的方法是有效和优越的。

引用次数: 0