Source Model Research Articles

Eolization: A project on spatial auralisation of wind turbine noise

The increase of onshore wind farm projects in Europe has heightened concerns regarding the noise emitted by wind turbines and the potential nuisance to nearby residents. To address this issue, advanced tools are required for accurately assessing and predicting the noise levels and perceived annoyance from wind turbines. Physics-based auralization emerges as a promising approach for recreating with fidelity wind turbine noise in controlled environments. To enhance the immersive experience, this study proposes to extend the current methodologies with the spatial representation of wind turbine noise using ambisonics. In this work, the sound emitted by each turbine is computed using an extended source model. The propagation is simulated with a parabolic equation method accounting for the flow surrounding the wind turbines. Finally, the signal is auralized and spatialized using spherical harmonic decomposition. This methodology provides a complete framework for predicting and reproducing the three-dimensional time-varying sound field around the listener. Overall, these improvements offer valuable insights into the spatial characteristics of wind turbine noise.

Flight test for the noise source modeling of the optionally piloted personal air vehicle

An acoustic flight test is carried out in order to obtain the noise hemisphere of the optionally piloted personal air vehicle (OPPAV). The OPPAV is one-seater, research urban air mobility (UAM) being developed by Korea Aerospace Research Institute. Noise radiated from the aircraft is measured using a microphone array consisting of a total of 78 microphones. Flight logs such as the aircraft position, velocity, and attitude are recorded synchronously with the acoustic signals. Measured acoustic signals are dedopplerized and divided into a number of segments. Noise source model for the OPPAV flight is constructed using these segments. The noise source model will be used for the assessment of noise annoyance during urban flight of the UAMs.

Methodology for determining new CNOSSOS-EU vehicle sound power coefficients using type approval data

CNOSSOS-EU is a standardized framework for strategic noise mapping across the European Union which member states are required to conduct every five years. The framework provides an approach to assess noise from different sources, including road traffic noise. Road traffic noise is primarily caused by vehicles. In CNOSSOS-EU, the vehicle exterior noise is modeled using fixed sound power coefficients. These coefficients were derived from measurements conducted up to 2006. Although significant progress has been made in reducing vehicle exterior noise, the existing model does not account for this technical progress. This paper first reviews the evolution of exterior noise legislation and its impact on vehicle noise levels. It then presents a new procedure for integrating the technical progress into vehicle source modeling in CNOSSOS-EU by incorporating type approval data. For demonstration purposes, this procedure is applied to up-to-date type approval data of M1 vehicles, resulting in a more accurate representation of current noise emissions. This research contributes to the refinement of the CNOSSOS-EU model, aligning it with the current state of technology and providing a more precise basis for environmental noise assessment.

Large aperture microphone array measurement at operating airport for sound source model creation of J-FRAIN

In 2022, authors conducted microphone array measurements at an operating airport to create sound source models for commercial aircraft based on the noise prediction framework J-FRAIN. A similar measurement was also performed by our colleagues in 2019, but it was a trial to establish the concept of the framework and was done using a microphone array in which the number of microphones was reduced to about half. In order to improve the accuracy of the sound source models, the microphone array with a 30-meter diameter consisting of 195 microphones mounted on the ground was used in 2022. Recording-based simultaneous noise measurement using sound level meters was also performed around the airport to validate prediction results in terms of time histories of 1/3 octave band sound pressure levels. This paper describes an overview of the microphone array measurements conducted in 2022 and a case study of noise prediction around the airport using sound source models created from the data of microphone array measurements conducted in 2022. Predicted results are compared with 1/3 octave band data measured using sound level meters mounted at a height of 4 m above the ground at sites of noise monitoring stations around the airport.

Noise modeling of UAM around the urban vertiport

The present study aims to investigate the noise distribution around urban vertiport caused by UAM(Multicopter) and helicopter. Noise modelings were carried out using SoundPLAN software at the some vertiports in Seoul, Korea. Sound source model of UAM and helicopter were used as a form of hemisphere. Noise predictions were undertaken with three different operations including depart, level flight and landing. In order to evaluate the noise impact on the buildings around the vertiport, noise levels at the outer wall of nearby buildings were calculated as the three noise units including Leq, Lmax and Lden. Also, noise maps were drawn around the vertiport. As a result, it was found that average noise level of UAM is approximately 12 dB lower than the helicopter noise levels. Regarding the noise frequency distribution, most high noises are made at the mid-frequencies during landing hours while low frequency noise level around 63 Hz was dominant at take-off time.

Sound propagation modelling in a street canyon by combining an image source model and impulse response measurements

This research introduces an innovative approach to achieve a computationally efficient model for calculating impulse responses in an urban environment, more precisely a street canyon, for auralization purposes. The image source method has shown its merits before, but when used for auralization, it typically lacks realism. In this study the image source method is used to create a basic impulse response of a street canyon, enriched by a stochastic response based on an envelope function. While the image source model accounts for the most prominent reflections, the stochastic part is due to all scattered sound energy in the street originating mainly from facade irregularities and street furniture. To obtain realistic frequency dependent envelope functions for various street canyon geometries, the function is based on measured impulse responses in nearly 100 inner-city street canyons in the city of Ghent, which were conducted by Thomas et al. (2013). By the proposed approach, we are able to generate impulse responses for street canyons suitable for sound perception research in an efficient way. More studies are needed to verify the applicability of the model for streets outside the Ghent dataset.

Drone fleet noise impact calculation - a methodology

The introduction of drones into an already dense urban infrastructure is expected to have an important impact on the population. The goal of the SESAR MUSE project, " Measuring U-Space Social and Environmental Impact ", is to develop a set of key performance indicators, methods and tools for the comprehensive and rigorous assessment of the impact of UAM operations on the liveability and quality of life in European cities. One focus domain is drone-generated noise. In the frame of this project, a number of ONERA tools is chained in order to create noise maps, from scratch, for realistic drone fleet scenarios. Hence, in this paper a methodology is presented where a flight mechanics tool, noise source models, an en-route tool and an open-source urban noise propagation tool are chained, allowing complete freedom with regard to choice and number of drones, concept of operations and type of urban environment.

Numerical modeling of infrasounds radiation and propagation in hurricanes

Infrasound can be useful for monitoring natural or human-induced energetic events in the atmosphere. For high-energy sources, the infrasonic signals can propagate over large distances around the globe, given the low rate of atmospheric absorption. This study focuses on the numerical simulation of infrasound propagation in realistic large-scale three-dimensional (3D) environments. More precisely, we are interested in the simulation of infrasound, called microbaroms, which are radiated from extreme weather conditions such as hurricanes. The simulation of hurricane-induced microbaroms relies on three main components: i) a hurricane model describing the atmospheric wind and pressure fields, ii) a microbarom source model, and iii) an infrasound propagation model. 3D simulations have been performed on a rectangular mesh. Azimuthal deviation caused by the hurricane and the stratified atmosphere can be seen in 3D. Further studies are needed to improve each model, and the correlation between them. The ultimate goal is to be able to simulate radiated infrasound propagation for any meteorological conditions.

A new noise source modelling tool for shooting sheds using novel edge diffraction methodology

Shooting ranges may have a shooting shed to protect the shooter against weather and to reduce noise to neighbours. The noise reduction depends on both shielding and reflection effects and can be difficult to measure due to complex surroundings. In this paper, a new tool will be presented to predict the influence of the shed on the emitted noise. The tool produces a substitute source which models the combined effect of the weapon and the shed, meant for efficient use in regular noise mapping software. It combines advanced edge diffraction methodology for the body of the shed, with a boundary element solution for minor areas representing openings or absorbers in the shed. With this approach we cover the full spectral range (up to 10 kHz) at low calculation time, with an intended precision comparable to traditional models based on boundary element methods. The tool is also intended for opposite use, by reciprocity, for instance to model quietness on the backside of a house by a road.

An easy-to-use analytical model for standing column wells operating with bleed

Accurate assessment of subsurface heat transport is vital for the design of standing column well systems. When bleed is utilized, the mathematical problem is governed by coupled heat transfer and groundwater flow within a radially convergent flow field, making the development of models challenging. While numerical models exist to simulate these transient processes, the absence of simple and accessible analytical solutions limits their broader application. This study addresses this gap by developing a novel easy-to-use analytical model to accurately represent the heat advection–diffusion problem in standing column wells operating with bleed. The proposed model combines the well-known infinite line source model with an innovative scaling function, inspired from the field of solute transport, through a simple convolution product. Notably, the developed model depends on only three dimensionless parameters: dimensionless advection time, Péclet number, and bleed ratio. Rigorous validation against two distinct sets of reference numerical solutions demonstrated the model’s efficiency, accuracy, and reliability across a broad spectrum of nine physical parameters. Key results include relative root mean square errors on the order of 10−3 across 200 reference solutions, confirming the model’s robustness. These findings highlight the model’s potential to significantly advance both research and practical applications in the design and optimization of standing column wells.

Damage mechanics coupled with a transfer learning approach for the fatigue life prediction of bronze/steel diffusion welded bimetallic material

The bronze/steel diffusion welded (BSDW) bimetallic material is often applied in the rotors of piston pumps to withstand complex alternating loads under high-speed operating conditions. Although diffusion welding is a type of solid-phase welding method to achieve high-quality material connections, the fatigue problems still deserve our attention, especially the very high cycle fatigue (VHCF) and high cycle fatigue (HCF) problems. However, due to the high cost of obtaining data, it is necessary to find an efficient and high-precision fatigue life prediction method for diffusion welded materials with a small sample size. In this study, a novel method continuum damage mechanics − transfer learning method (CDM-TLM) for fatigue life prediction of BSDW material is proposed based on the transfer learning (TL) and continuum damage mechanics − finite element method (CDM-FEM). In comparison with the test results, the predicted values of BSDW material fatigue life all fall within the twice error band of the median values of the test life. The influence of frozen layers during TL and training samples in source and target models on the prediction performance is further discussed. CDM-TLM is an effective life prediction method for high-precision life prediction of BSDW material with a small sample size.

Improvements to EXSIM in Ground Motion Simulation for Earthquakes Explained by Double‐Corner‐Frequency Source Model

AbstractThe stochastic EXtended finite‐fault ground‐motion SIMulation algorithm (EXSIM) has been widely applied in simulating and predicting broadband strong ground‐motion. However, an increasingly number of researchers have found that EXSIM may overestimate ground‐motions at low frequencies for some large‐magnitude earthquakes and/or thrust earthquakes, for which the far‐field source model has been explained by a double‐corner‐frequency model. Despite controversy, the double‐corner‐frequency model is now being accepted as one of the main categories of the far‐field source model. This study demonstrated the limited applicability of EXSIM to earthquakes explained by the double‐corner‐frequency source model, by presenting the equivalence between motions generated by EXSIM and those generated by EXSIM's point‐source version, SMSIM, which adopts the ω‐square single‐corner‐frequency model. Furthermore, two improvements to EXSIM have been proposed: (a) the incorporation of the asperity‐distributed stress‐drop compound faults model and (b) the hybrid application of EXSIM with the proposed model. The effects of the two improvements have been verified by comparing EXSIM‐generating motions with recorded ground‐motions for the 2013 Mw 6.7 Lushan thrust earthquake. Significantly, consistent simulation accuracy has been achieved across high‐ and low‐frequency bands as well as in far‐ and near‐fields. The consistent accuracy of the improved EXSIM in simulating high‐ and low‐frequency ground motions enables its direct and independent application to broadband ground motion simulations. Moreover, the first validation of this consistent accuracy in both near‐ and far‐field scenarios further enhances its application in earthquake engineering practices.

An Integrated Imitation and Reinforcement Learning Methodology for Robust Agile Aircraft Control with Limited Pilot Demonstration Data

In this paper, we present a methodology for constructing data-driven maneuver generation models for agile aircraft that can generalize across a wide range of trim conditions and aircraft model parameters. Maneuver generation models play a crucial role in the testing and evaluation of aircraft prototypes, providing insights into the maneuverability and agility of the aircraft. However, constructing the models typically requires extensive amounts of real pilot data, which can be time-consuming and costly to obtain. Moreover, models built with limited data often struggle to generalize beyond the specific flight conditions covered in the original dataset. To address these challenges, we propose a hybrid architecture that leverages a simulation model, referred to as the source model. This open-source agile aircraft simulator shares similar dynamics with the target aircraft and allows us to generate unlimited data for building a proxy maneuver generation model. We then fine-tune maneuver generation model to the target aircraft using a limited amount of real pilot data. Our approach combines techniques from imitation learning, transfer learning, and reinforcement learning to achieve this objective. To validate our methodology, we utilize real agile pilot data provided by Turkish Aerospace Industries (TAI). By employing the F-16 as the source model, we demonstrate that it is possible to construct a maneuver generation model that generalizes across various trim conditions and aircraft parameters without requiring any additional real pilot data. Our results showcase the effectiveness of our approach in developing robust and adaptable maneuver generation models for agile aircraft.

Monte Carlo simulated correction factors for high dose rate brachytherapy postal dosimetry audit methodology

Background and PurposeFull-scatter conditions in water are impractical for postal dosimetry audits in brachytherapy. This work presents a method to obtain correction factors that account for deviations from full-scatter water-equivalent conditions for a small plastic phantom Material and MethodsA 16 × 8 × 3 cm phantom (PMMA) with a radiophotoluminescent dosimeter (RPLD) at the centre and two catheters on either side was simulated using Monte Carlo (MC) to calculate correction factors accounting for the lack of scatter, non-water equivalence of the RPLD and phantom, source model and backscatter for HDR 60Co and 192Ir sources. ResultsThe correction factors for non-water equivalence, lack of full scatter, and the use of PMMA were 1.062 ± 0.013, 1.059 ± 0.008 and 0.993 ± 0.009 for 192Ir and 1.129 ± 0.005, 1.009 ± 0.005 and 1.005 ± 0.005 for 60Co respectively. Water-equivalent backscatter thickness of 5 cm was found to be adequate and increasing thickness of backscatter did not have an influence on the RPLD dose. The mean photon energy in the RPLD for four HDR 192Ir sources was 279 ± 2 keV in full scatter conditions and 295 ± 1 keV in the audit conditions. For 60Co source the corresponding mean energies were 989 ± 1 keV and 1022 ± 1 keV respectively. ConclusionsCorrection factors were obtained through the MC simulations for conditions deviating from TG-43, including the amount of back scatter, and the optimum audit set up. Additionally, the influence of different source models on the correction factors was negligible and demonstrates their generic applicability.

Acoustic Emission simulation: assessing the influence of AE source modeling and addressing the issue of dimensionality in the propagation medium

A thorough understanding of simulating Acoustic Emission (AE) requires the source modeling and propagation medium to be considered respectively. Crack initiation and propagation simulation as an AE source can be performed by the Successive Node Release (SNR) method and the Direct Node Release (DNR) method. Both are compared to show how the source model affects the AE signals. The SNR method is commonly used to simulate the crack initiation and propagation. It is necessary to consider the influence of the crack velocity profile on the crack initiation [1]. However, for the DNR method, all nodes are instantaneously released once the critical crack length is attained. Hence, it is not necessary to consider the velocity profile. Instead, it is sufficient to define the average crack velocity. However, the kinetic energy varies when employing different methods, leading to a variance in the source energy. We thus assess the suitable conditions to apply the DNR method to represent crack initiation and investigate the influence of the node release methods on the AE signals. Secondly, a comparative assessment of the Pencil-lead Break test on PMMA plates is conducted using both two-dimensional (2D) and three-dimensional (3D) finite element simulations. This analysis is valuable for understanding the deficiencies and constraints of the 2D simulation. However, there is a significant difference in the AE signals between 2D and 3D simulations, suggesting that the 2D simulation is unable to capture the AE information during wave propagation, in spite of its lower computational cost. Therefore, to obtain extensive and precise information from the PLB test, it is crucial to have an extensive understanding of the limits of the 2D simulation. This understanding should take into account factors such as the distance between the sensor and the source, as well as the descriptors chosen to characterize the AE.

Gravity.jl: Fast and accurate gravitational lens modeling in Julia

We present Gravity.jl, a new proprietary software for the modeling of gravitational lens systems. Gravity.jl is written in the Julia programming language, and is designed to be fast, accurate, and flexible. It can be used to model gravitational lens systems composed of multiple lensing planes, and to perform Bayesian inference on the lens model parameters. In this paper we present the theoretical and statistical ideas behind the code, and we describe its main features. In this first paper of the series, we focus on the modeling of point-like and small extended sources, for which we can linearize the lens equation. We show a practical use of Gravity.jl on a galaxy-scale lens, and we compare the results with those obtained with other codes. We also show how Gravity.jl can be used to perform Bayesian inference on cosmological parameters.

Modeling and application of array light sources in ultraviolet communication systems

Reliable communication over long distances using ultraviolet (UV) light is challenging due to the strong scattering and atmospheric absorption of UV light. However, with advancements in UV LED technology, the performance of transmitters can be significantly enhanced by employing an array of light sources, which improves the overall quality of communication. Despite these advancements, the impact of various array configurations—such as shape, quantities, and other parameters—on light intensity distribution and UV optical communication remains largely unexplored. In this work, we first modeled the light intensity distribution of arrayed light sources with varying shapes, quantities, and other parameters. We improved the traditional Monte Carlo method to better accommodate these distributions, enabling more accurate simulations. Subsequently, we obtained simulation results for different array configurations. Building on these findings, we developed a UV LED array communication system that achieved a bit error rate (BER) of 10−6 for information transmission over a distance of 500 meters. This research provides valuable insights into the long-distance transmission capabilities of UV light using arrayed light sources.

Modeling of Coupled Harmonic Current Source for Grid-Connected Inverters

Modeling of Coupled Harmonic Current Source for Grid-Connected Inverters

NAS-BNN: Neural Architecture Search for Binary Neural Networks

Binary Neural Networks (BNNs) have gained extensive attention for their superior inferencing efficiency and compression ratio compared to traditional full-precision networks. However, due to the unique characteristics of BNNs, designing a powerful binary architecture is challenging and often requires significant manpower. A promising solution is to utilize Neural Architecture Search (NAS) to assist in designing BNNs, but current NAS methods for BNNs are relatively straightforward and leave a performance gap between the searched models and manually designed ones. To address this gap, we propose a novel neural architecture search scheme for binary neural networks, named NAS-BNN. We first carefully design a search space based on the unique characteristics of BNNs. Then, we present three training strategies, which significantly enhance the training of supernet and boost the performance of all subnets. Our discovered binary model family outperforms previous BNNs for a wide range of operations (OPs) from 20M to 200M. For instance, we achieve 68.20% top-1 accuracy on ImageNet with only 57M OPs. In addition, we validate the transferability of these searched BNNs on the object detection task, and our binary detectors with the searched BNNs achieve a novel state-of-the-art result, e.g., 31.6% mAP with 370M OPs, on MS COCO dataset. The source code and models will be released at https://github.com/VDIGPKU/NAS-BNN.

Pseudo-label refinement via hierarchical contrastive learning for source-free unsupervised domain adaptation

Source-free unsupervised domain adaptation aims to adapt a source model to an unlabeled target domain without accessing the source data due to privacy considerations. Existing works mainly solve the problem by self-training methods and representation learning. However, these works typically learn the representation on a single semantic level and barely exploit the rich hierarchical semantic information to obtain clear decision boundaries, which makes it hard for these methods to achieve satisfactory generalization performance. In this paper, we propose a novel hierarchical contrastive domain adaptation algorithm that exploits self-supervised contrastive learning on both fine-grained instances and coarse-grained cluster semantics. On the one hand, we propose an adaptive prototype pseudo-labeling strategy to obtain much more reliable labels. On the other hand, we propose hierarchical contrastive representation learning on both fine-grained instance-wise level and coarse-grained cluster level to reduce the negative effect of label noise and stabilize the whole training procedure. Extensive experiments are conducted on primary unsupervised domain adaptation benchmark datasets, and the results demonstrate the effectiveness of the proposed method.