Structural Noise Research Articles

Information limits and Thouless-Anderson-Palmer equations for spiked matrix models with structured noise

We consider a prototypical problem of Bayesian inference for a structured spiked model: a low-rank signal is corrupted by additive noise. While both information-theoretic and algorithmic limits are well understood when the noise is a Gaussian Wigner matrix, the more realistic case of structured noise still remains challenging. To capture the structure while maintaining mathematical tractability, a line of work has focused on rotationally invariant noise. However, existing studies either provide suboptimal algorithms or are limited to a special class of noise ensembles. In this paper, using tools from statistical physics (replica method) and random matrix theory (generalized spherical integrals) we establish the characterization of the information-theoretic limits for a noise matrix drawn from a general trace ensemble. Remarkably, our analysis unveils the asymptotic equivalence between the rotationally invariant model and a surrogate Gaussian one. Finally, we show how to saturate the predicted statistical limits using an efficient algorithm inspired by the theory of adaptive Thouless-Anderson-Palmer (TAP) equations. Published by the American Physical Society 2025

Reconstruction of vibration noise in plate structures based on Data-Physics driven model and transfer learning.

The identification of vibration and reconstruction of sound fields of plate structures are important for understanding the vibroacoustic characteristics of complex structures. This paper presents a data-physics driven (DPD) model integrated with transfer learning (DPDT) for high-precision identification and reconstruction of vibration and noise radiation of plate structures. The model combines the Kirchhoff-Helmholtz integral equation with convolutional neural networks, leveraging physical information to reduce the need for extensive data. By embedding transfer learning, it enhances generalization across different structures. Two plate models of different sizes and publicly experimental data were used to evaluate the model's performance. Results show that the DPDT model achieves superior prediction accuracy stability, and faster convergence compared to the DPD model, with high R2, normalized cross-correlation, and low normalized mean squared error values, demonstrating its robustness and efficacy in reconstructing sound fields even with limited data points. This approach demonstrates significant potential for practical engineering applications, particularly in bridge vibration and noise control.

A Hybrid Structure Noise Shaping SAR ADC

Hybrid structure noise shaping (NS) SAR ADC combines both SAR ADC and sigma–delta ADC technologies with the advantages of high precision and low power consumption. In NS SAR ADC, mismatch error shaping (MES) method is used to solve the mismatch of low-level capacitors, and Data Weighted Averaging (DWA) method is used to solve the mismatch of high-level capacitors. The single-input comparator is improved into a dual-input comparator. The buffer and integrator circuits are realized by the static op-amp. The noise of op-amp is reduced by chopper technology and the accuracy of ADC is improved. A high-precision NS-SAR ADC with a sampling frequency of 1MSps and bandwidth of 25[Formula: see text]kHz was designed using the GSMC 90[Formula: see text]nm process. When the input signal is at full scale, the ENOB of the NS SAR ADC is higher than 14.73 bits, the SFDR is higher than 98[Formula: see text]dB, the average power consumption is 567.93[Formula: see text][Formula: see text]W and the quality factor is 169.2[Formula: see text]dB.

Adltformer Team-Training with Detr: Enhancing Cattle Detection in Non-Ideal Lighting Conditions Through Adaptive Image Enhancement.

This study proposes an image enhancement detection technique based on Adltformer (Adaptive dynamic learning transformer) team-training with Detr (Detection transformer) to improve model accuracy in suboptimal conditions, addressing the challenge of detecting cattle in real pastures under complex lighting conditions-including backlighting, non-uniform lighting, and low light. This often results in the loss of image details and structural information, color distortion, and noise artifacts, thereby compromising the visual quality of captured images and reducing model accuracy. To train the Adltformer enhancement model, the day-to-night image synthesis (DTN-Synthesis) algorithm generates low-light image pairs that are precisely aligned with normal light images and include controlled noise levels. The Adltformer and Detr team-training (AT-Detr) method is employed to preprocess the low-light cattle dataset for image enhancement, ensuring that the enhanced images are more compatible with the requirements of machine vision systems. The experimental results demonstrate that the AT-Detr algorithm achieves superior detection accuracy, with comparable runtime and model complexity, reaching 97.5% accuracy under challenging illumination conditions, outperforming both Detr alone and sequential image enhancement followed by Detr. This approach provides both theoretical justification and practical applicability for detecting cattle under challenging conditions in real-world farming environments.

ConvMPAnet: A Novel End-to-End Lightweight Damage Localization Framework Under Heavy Noise in Composite Structures

ConvMPAnet: A Novel End-to-End Lightweight Damage Localization Framework Under Heavy Noise in Composite Structures

On the Assessment of Porosity of Metals Obtained by Hot Isostatic Pressing Based on the Analysis of Structural Acoustic Noise

On the Assessment of Porosity of Metals Obtained by Hot Isostatic Pressing Based on the Analysis of Structural Acoustic Noise

Acoustics of rubbing feathers: the velvet of owl feathers reduces frictional noise.

One feather structure associated with an owl's ability to fly quietly is the soft dorsal surface on their flight feathers: the velvet. This velvet is a mat of elongated filamentous pennulums that extend up from feather barbules. The aerodynamic noise hypothesis posits this velvet reduces aerodynamic noise caused by formation of turbulence, while the structural noise hypothesis posits the velvet acts as a dry lubricant, reducing frictional noise produced by feathers sliding past one another. We investigated the structural noise hypotheses by quantifying the length of velvet on 24 locations across the wing of Barred owl (Strix varia) and then qualitatively assessing the presence of the velvet in 24 bird species. We found that the velvet has evolved at least 4 times independently (convergently) in owls, nightbirds, hawks and falcons. Then, we rubbed 96 pairs of feathers together from 17 bird species (including the four clades that have independently evolved the velvet) under three experimental treatments: control, hairspray applied (to impair the velvet), and hairspray removed. The sound of feathers rubbing against each other was broadband, similar to sound of rubbing sandpaper or Velcro. Species with velvet produced rubbing sounds that were 20.9 decibels (dB) quieter than species without the velvet, and velvet-coated feathers became 7.4 dB louder when manipulated with hairspray, while feathers lacking velvet only increased in loudness by 1.7 dB, relative to the control treatments. These results all support the hypothesis that the velvet primarily functions to ameliorate the sounds of feathers rubbing against other feathers.

Temperature-responsive metamaterials made of highly sensitive thermostat metal strips.

Temperature-responsive metamaterials have remarkable shape-morphing ability during thermal energy conversion. However, integrating the thermal shape programmability, wide-working temperature range, fast temperature response, and actuation into metamaterials remains challenging. Here, we introduce using thermostat metal strips to assemble metamaterials with desirable and balanced temperature-responsive properties, and we systematically investigate the thermal deformation performance. Achieving 70 to 80% of the designed strain requires only 5 seconds of heating. A thermal strain of around 30% is achieved for the assembled metamaterials, surpassing other bimetallic metamaterials by a magnitude of 100 to 200. The actuation capacity of thermostat metal strips exceeds 26 times their weight. Further, by leveraging the highly programmable thermal deformation, the tuneable bandgap range is 3847 to 40,000 hertz. These fully integrated mechanical performances in the multiphysics have great application potential, for example, as soft actuators and soft robots in intelligent structure systems, vibration isolation and noise reduction in hypersonic vehicles, and unique thermal deformation in precision instruments.

A foundation model for enhancing magnetic resonance images and downstream segmentation, registration and diagnostic tasks.

In structural magnetic resonance (MR) imaging, motion artefacts, low resolution, imaging noise and variability in acquisition protocols frequently degrade image quality and confound downstream analyses. Here we report a foundation model for the motion correction, resolution enhancement, denoising and harmonization of MR images. Specifically, we trained a tissue-classification neural network to predict tissue labels, which are then leveraged by a 'tissue-aware' enhancement network to generate high-quality MR images. We validated the model's effectiveness on a large and diverse dataset comprising 2,448 deliberately corrupted images and 10,963 images spanning a wide age range (from foetuses to elderly individuals) acquired using a variety of clinical scanners across 19 public datasets. The model consistently outperformed state-of-the-art algorithms in improving the quality of MR images, handling pathological brains with multiple sclerosis or gliomas, generating 7-T-like images from 3 T scans and harmonizing images acquired from different scanners. The high-quality, high-resolution and harmonized images generated by the model can be used to enhance the performance of models for tissue segmentation, registration, diagnosis and other downstream tasks.

Real-Time Structural Health Monitoring of Bridges Using Convolutional Neural Network-Based Loss Factor Analysis for Enhanced Energy Dissipation Detection

This study introduces a novel method for real-time structural health monitoring (SHM) of bridges using a convolutional neural network (CNN) model that leverages loss factor analysis. As bridge structures deteriorate over time, often accelerated by operational loads, maintaining structural integrity and safety becomes critical. The proposed approach utilizes the concept of a loss factor, which represents the process of energy dissipation across different vibration states, as a key indicator of structural health. This factor is computed from vibration energy spectra, which include amplitude and frequency components, processed through the CNN model for high sensitivity to structural changes. The results demonstrate that the energy dissipation of the bridge during operation can be categorized into signals from three distinct sources: structural responses, defects-related indicators, and noise interference. By monitoring variations in the loss factor over time, the model effectively identifies early signs of structural deterioration, which is critical for timely maintenance interventions. The study also highlights the adaptability to different load conditions and environmental factors, ensuring robust performance in various operational scenarios. The findings underscore the potential of the CNN model to transform SHM practices by enhancing early defect detection, supporting preventive maintenance, and ultimately extending the lifespan of bridge infrastructure.

The use of geometrical nonlinear local resonators to enhance the vibration control performance of metamaterial structures

Abstract Metamaterials has become an exciting solution for structural noise and vibration reduction in many engineering applications. Acoustic metamaterials are structures built using repetitive assemblies of identical elements to explore either Bragg-scattering or localized resonance to control mechanical waves. They present frequency bands in which waves do not freely propagate, named bandgaps, allowing acoustic and vibration attenuation at one or multiple targeted frequency ranges. If properly designed and implemented, the inclusion of a nonlinear stiffness can result in resonance frequency shifts that broaden the attenuation frequency band. This work investigates the influence of an array of lightweight nonlinear local resonators (NLR) realized via additive manufacturing on the low-frequency bandgap formation of a metamaterial beam. The NLR uses the high-static-low-dynamic stiffness (HSLDS) concept, which introduces geometric nonlinearity. The proposed model is validated through simulation and experimental analysis. The NLR results in a broadened and shifted attenuation band to lower frequencies without the necessity of adding extra mass to the resonator. This investigation contributes to understanding improvements provided by nonlinear elements for vibration control in metastructures.

Hybrid Noise Control in Helicopter Cabin with Piezoelectric Stack Periodic Strut

Improving the vibration isolation performance of the helicopter’s main gearbox strut can reduce the vibration level transmitted to the aircraft body and the level of structural noise radiation. This article proposes a piezoelectric stack periodic strut (PSPS) scheme based on the piezoelectric stack and rubber periodic arrangement. Piezoelectric stack actuators can achieve active vibration and noise control, and the periodic structure of the piezoelectric stack actuators and rubber also has wideband stopband characteristics for reducing vibration and noise. Therefore, PSPS has active and passive hybrid control capabilities, which has advantages in vibration and noise reduction performance, the power of active control, volume and mass, etc. This article presents a full-parameter model that can describe the electromechanical coupling dynamic performance of PSPS. The correctness of the model was verified by the finite element software, and the concept of optimal isolation potential of PSPS was proposed and evaluated. Finally, the hybrid control capability of PSPS for single-frequency and wide-frequency noise was verified through experiments. The results show that compared with the noise level of the cabin using an ordinary strut, PSPS has wideband vibration reduction and noise reduction performance above 500 Hz, with a maximum noise attenuation exceeding 25 dB. When applying active noise control to the PSPS on the helicopter model, a control effect of 12.63–15.56 dB at 900 Hz was achieved.

Exposure to industrial noise and professional hearing loss among employees of enterprises in the Murmansk region

Occupational sensorineural hearing loss (OSNHL) is one of the oldest and at the same time urgent problems of occupational medicine. In this regard, the characteristics of the development and prevalence of OSNHL among enterprise workers in the Murmansk region became the aim of the study. We analyzed the data of the social and hygienic monitoring «Working conditions and occupational morbidity», as well as the data of the register of newly identified occupational diseases in the Murmansk region in 2003–2022. It was found that in 2003–2022, noise was the most prevalent harmful production factor (20.6 %), and OSNHL was the most prevalent occupational disease (15.5 %) at the region’s enterprises. 770 cases of OSNHL were diagnosed for the first time, mainly in men (96.0 %), miners (65.1 %) and metallurgical workers (17.5 %). In 308 (40.0 %) workers, ONSHL was the only disease, and in 462 (60.0 %) workers, OSNHL was combined with other diseases, primarily diseases of the musculoskeletal system (n=256) and vibration disease (n=54). Over 20 years, there was a decrease in the following indicators: the share of noise in the overall structure of harmful industrial factors (p<0.001); the risk of exposure to noise (OR=1.13; 95 % CI 1.12–1.15; p<0.001); the proportion of OSNHL in the overall structure of occupational pathology (p<0.001). On the contrary, there was an increase in the share of workers with OSNHL in combination with other diseases (p<0.001) and in the share of OSNHL with mild hearing impairment (p=0.016). The observed 32.8 % decrease in the number of OSNHL cases was due to a decrease in the number of workers, and not to the effect of preventive measures. The data obtained, despite the achieved positive dynamics of hygienic and clinical indicators, demonstrate the need to continue active measures to reduce the level of industrial noise and prevent occupational hearing loss, especially among workers in metallurgical enterprises.

Contrastive message passing for robust graph neural networks with sparse labels

Graph Neural Networks (GNNs) have achieved great success in semi-supervised learning. Existing GNNs typically aggregate the features via message passing with the aid of rich labels. However, real-world graphs have limited labels, and overfitting weakens the classification ability when labels are insufficient. Besides, traditional message passing is sensitive to structure noise, such as perturbations on edges. The performance of GNNs would drop sharply when trained on such graphs. To mitigate these issues, we present a noise-resistant framework via contrastive message passing. Except for the limited labelled nodes as supervision widely used in GNNs, we model the topology structure by graph likelihood as extra supervision. Specifically, we first propose contrastive graph likelihood, which is defined as a product of the edge likelihood on all connected node pairs. Then, we apply two unfolding updated steps via descent iterations. The first step updates the features in a single view with the aid of initializing the edge probability. Then the second step applies a binary edge for homophily and heterophily views, respectively. The homophily view applies attractive force to pull the positive-connected nodes close; otherwise, the heterophily view utilizes repulsive force to push away the negative-connected nodes. Extensive experiments show that our method has superior performance on semi-supervised node classification tasks with sparse labels and excellent robustness under perturbations in structure.

Experimental and numerical analysis of rubber isolator dynamic stiffness under hydrostatic pressure

Due to their function in damping and attenuating vibrations, as well as their relatively low cost, rubber isolators have widespread applications in industries such as aerospace, automotive and maritime. As components with distinct nonlinear behavior, accurately predicting the dynamic characteristics of isolators is essential for overall structural design and vibration noise prediction. Although extensive research has been conducted on the static and dynamic performance of rubber isolators, there has been limited investigation into performance under certain specialized application scenarios, such as hydrostatic pressure environments. These environments are indeed real, for instance, isolators employed in the bow ballast tank of underwater vehicle to mitigate vibrations during weapon launch processes. In such instances, isolators are subjected not only to the influence of added mass due to the water medium during vibration but also to increasing hydrostatic pressure on the isolator surface with increasing depth. This study introduces an original experimental apparatus capable of measuring the dynamic stiffness of isolators while simulating hydrostatic pressure conditions. The constitutive model parameters governing the hyperelastic and viscoelastic properties of rubber were identified via nonlinear tests, serving as input parameters for numerically predicting the dynamic stiffness within a water medium environment. Furthermore, a comprehensive analysis was conducted to evaluate the impacts of preload, water medium, boundary conditions, and hydrostatic pressure on the dynamic stiffness of the isolator. Results indicate that preload tends to reduce peak dynamic stiffness in the shear directions, while the water medium significantly increases high-frequency dynamic stiffness in the shear directions. Within the range of 1 MPa, the impact of increasing hydrostatic pressure on dynamic stiffness can be largely disregarded. When calculating the dynamic stiffness of the isolator, it is crucial to consider the actual installation environment and set the acoustic boundary conditions of the surrounding water domain accordingly.

Investigation on the characteristics of flow-induced vibration and noise of acoustic window with embedded compound acoustic black holes

In recent decades, acoustic black hole (ABH) structures have shown significant potential for suppressing structural vibrations and noise radiation. However, the ABH structure has not been applied to panel–cavity noise control under heavy fluid loading. In this study, a hybrid numerical approach was used to analyse the vibro-acoustic response suppression mechanism of a compound acoustic black hole (CABH) acoustic window under turbulent boundary layer (TBL) excitation. Both sides of the acoustic window interact with the water. The results show that the local vibration modes of the CABH lead to a strong interaction between the structure and damping material. The amplitude of the high-order modal wavenumber spectra of the CABH was small, which weakened the spatial coupling between the TBL pressure and the structure. The coupling strength between the CABH acoustic window and the internal sound field was reduced, causing noise suppression inside the cavity. To further examine the noise reduction performance of the CABH panel cavity system, a flow-induced experiment with a large-scale CABH acoustic window was carried out in a large cavitation channel, and the results showed significant hydrodynamic self-noise reduction.

Evaluating bridge-borne noise contribution degree for a railway steel–concrete composite box girder

A train running across a bridge usually produces higher noise than on the ground. This increase in noise varies considerably from one bridge to another, but the noise contribution of different bridges is unclear. A hybrid finite element– boundary element–statistical energy analysis method is used to investigate the vibroacoustic characteristics of a steel–concrete composite (SCC) bridge induced by running trains. The numerical procedure is verified by using a scaled SCC bridge measured in the laboratory. The mechanism of vibration transmission and vibroacoustic comparisons between a 40-m span SCC bridge and a traditional all-concrete bridge are investigated, and the results show that higher vibration levels (12–20 dBA for different components) and sound levels (12 dBA on average) arise with the former. At a train speed of 300 km/h, the structural noise contribution of the SCC bridge to the total noise emission from the train pass-by is 67% in the near field. When the train speed reaches 385 km/h, the structural noise contribution can still reach 27%. The contribution of the deck and bottom plate to the structural noise was the highest, accounting for approximately 37%.

ANALYSIS OF NOISE CHARACTERISTICS OF TRAFFIC FLOWS ON THE EXAMPLE OF KRASNODAR

In this article, the authors consider the noise characteristics of traffic flows in urban conditions. The unevenness of traffic flows in time and space entails a change in noise effects on the environment. The main source of noise pollution is the internal combustion engine, which produces aerodynamic noise and structural noise. Noise level measurements are carried out separately for vehicles in motion and for stationary vehicles. The noise level is directly influenced by the mode of movement of cars, the speed of traffic flow, and the intensity of traffic flow. Deterioration of working and leisure conditions with a high level of traffic noise can lead to a decrease in labor productivity and the occurrence of nervous disorders. Therefore, the protection of the population from the noise generated by transport is not only of social, but also of economic importance. Purpose: to analyze the level of noise impacts of traffic flow in urban conditions. Method and methodology. The methods of observation, statistical analysis, and synthesis were used in the article. Results: noise level measurements were carried out on the selected section of the road network, and calculated data were obtained based on the declared characteristics of the traffic flow. Scope of application of the results: research activities to improve the environmental safety of the traffic flow.

Persistently active El Niño–Southern Oscillation since the Mesozoic

The El Niño-Southern Oscillation (ENSO), originating in the central and eastern equatorial Pacific, is a defining mode of interannual climate variability with profound impact on global climate and ecosystems. However, an understanding of how the ENSO might have evolved over geological timescales is still lacking, despite a well-accepted recognition that such an understanding has direct implications for constraining human-induced future ENSO changes. Here, using climate simulations, we show that ENSO has been a leading mode of tropical sea surface temperature (SST) variability in the past 250 My but with substantial variations in amplitude across geological periods. We show this result by performing and analyzing a series of coupled time-slice climate simulations forced by paleogeography, atmospheric CO2 concentrations, and solar radiation for the past 250 My, in 10-My intervals. The variations in ENSO amplitude across geological periods are little related to mean equatorial zonal SST gradient or global mean surface temperature of the respective periods but are primarily determined by interperiod difference in the background thermocline depth, according to a linear stability analysis. In addition, variations in atmospheric noise serve as an independent contributing factor to ENSO variations across intergeological periods. The two factors together explain about 76% of the interperiod variations in ENSO amplitude over the past 250 My. Our findings support the importance of changing ocean vertical thermal structure and atmospheric noise in influencing projected future ENSO change and its uncertainty.

Modal analysis and structural noise control of vehicle body frame

The structural noise inside the vehicle cabin is mainly low-frequency vibration, which is closely related to the modal characteristics of the vehicle frame. The finite element method was used to simulate and calculate the body frame under free modal conditions, and the first four effective modal shapes were obtained. The calculation error of the natural frequencies was verified through modal experiments. Taking structural stiffness into account, a sound-structure coupled model was established. The suspension connection point was selected as the excitation point, and a one-way dynamic load was applied to obtain the noise and vibration responses of the front and rear rows. Based on the modal analysis results, the top-roof reinforcement scheme was adopted to verify the noise suppression effect of the structure. The results show that the optimized structure can effectively suppress structural noise, which plays an important role in improving the NVH characteristics.