Acoustic Signal Detection

Wiebe, P. H., Chu, D., Kaartvedt, S., Hundt, A., Melle, W., Ona, E., and Batta-Lona, P. 2010. The acoustic properties of Salpa thompsoni. – ICES Journal of Marine Science, 67: 583–593. Aggregations of the salp Salpa thompsoni were encountered during the Antarctic krill and ecosystem-studies cruise on the RV “G.O. Sars” from 19 February to 27 March 2008. The salp's in situ target strength (TS), size, number of individuals in aggregate chains, and chain angle of orientation were determined. Shipboard measurements were made of Salpa thompsoni's material properties. Individual aggregates were mostly 45.5–60.6 mm in mean length; relatively rare solitaries were ∼100 mm. Chains ranged from 3 to at least 121 individuals, and in surface waters (<20 m), they showed no preferred angle of orientation. Sound-speed contrast (h) ranged from 1.0060 to 1.0201 and density contrast (g) estimates between 1.0000 and 1.0039. The in situ TS distributions peaked between −75 and −76 dB at 38 kHz,...

This paper proposes an experimental setup composed of two interconnected vertical arrays: one transmit only array (TOA) and a receive only array (ROA). It is shown that using the time reversal principle, where the ocean is used as a spatial matched filter, signal energy can be simultaneously focused on each ROA hydrophone and thus obtain the detector output by simple summation of the received energy over the array. This setup effectively configures a multistatic system with several transmitters and several receivers coherently processed both in time and space. Simulations using a ray trace propagation model combined with a naive scattering formulation show that the obtained empirical detection probability is close to the theoretical optimal bound derived assuming space -time white Gaussian noise. This system was tested in a sea trial that took place during September 2007 in the Hopavågen Bay near Trondheim, Norway. The actual setup was composed of a 2 sources TOA on a shallow area of 8 m depth near the shore and a 16 receivers ROA approximately 100 m apart in a 8 to 25 m depth range dependent duct. The transmitted signals were 200 ms duration LFMs with 3 kHz bandwidths in two frequency bands centered on 5 and 10 kHz. The results obtained in realistic conditions show that a 1.7 m 2 hard plate could be detected when placed across the barrier. Several results are presented and compared with the theoretical values. Although the system can be significantely improved, in particular by using more populated source and receiving arrays, it is believed that these results can be reproduced at sea in harbor like conditions.

Free metallic particles, inside gas insulated switchgear (GIS) apparatus are considered hazardous to the insulation system, can be detected using acoustic sensors placed on the enclosure. From acoustic measurements, it is rather easy to determine the partial discharge pulses caused by loose or fixed particles. One can readily recognise the motion modes of the free particles by studying the acoustic signal. This paper presents the analysis of free metallic particle motion modes in GIS apparatus from acoustic measurements. The parameters such as maximum acoustic amplitude, number of acoustic pulses, mean value of maximum amplitude of 10 measurements, number of pulses which exceed the specified % of maximum amplitude characterising partial discharge activity of Al as well as Cu particles of different shapes are investigated

Image processing, as well as acoustic signal detection, have had major enhancements over the years, and this is due to AI. In the past, most algorithms involved using basic signal processing where features needed to be extracted manually and then various rules were applied when the data grew large. Deep learning models, for example, provide a durable solution to ventilation by eliminating the need for manual feature engineering as well as improving the detection rate in areas of health, surveillance and even industrial applications. This paper offers a comprehensive analysis of the emerging innovation driven by Advanced Intelligence in the field of image processing and the detection of acoustic signals with regard to the substrate patterns identified by AI technologies such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), as well as other sophisticated algorithms. The paper also describes how AI, when combined with image processing and acoustic detection, can add more value to the results being produced. Due to the large number of cases and training data, patterns can be learned and are as follows: image classification, object detection, process anomaly detection in industrial systems, as well as acoustic event recognition in noisy environments. The paper aims to provide an understanding of the AI methodologies adopted in both domains and, to this end, offers examples of specific industries and rationales for their implementation of these technologies. An extensive discussion of the basics of neural networks and their modifications is provided, with emphasis on the application of those structures for automated image feature extraction and acoustic pattern recognition. We also study the issues of comparison, accuracy, computational complexity, and the ability of AI models to function in similar conditions. This article also seeks to present how AI models can be enhanced by integrating image processing with acoustic signal detection methods and should produce possible research directions for increasing AI performance. Finally, the authors recap the main findings, provide information about advanced methods in their field, and show some possible future uses in self-driving cars, robots and drones, and meteorological monitoring.

An analysis of drone detection methods was carried out in this paper. There is a reasoned choice of the optimal method for implementation based on the analysis of acoustic signal analysis is present. These methods provide the measurement of drone coordinates and complement each other depending on the flight conditions. The structure of a distributed sensor network oriented on the online detection of drones and their tracking has been synthesized. The developed system provides the appropriate complexity as well as large flexibility and scalability.

The acoustic waves caused by partial discharges inside a dielectric can be detected by sensors of acoustic emissions (EA) and analyzed in the time domain. The experimental results that are presented, demonstrate the capability of online detection of these sensors, in the vicinity of a cable, splice or terminal accessory. The EA sensors are immune to electromagnetic interference, are a non-destructive detection method and guarantee galvanic decoupling from the voltage network; This technique of detection of partial discharges can be applied as a test method for preventive or predictive maintenance (maintenance based on the condition), in equipment or installations of medium and high voltage in service and represents an alternative to electrical detection systems, conventional or not, which continue to be based on the detection of current pulses.

En la medición y ensayos de descargas parciales (DP) en equipamiento de potencia en ambiente industrial, existen variadas fuentes de ruido que se superponen a la señal de DP, alterando los resultados obtenidos. Estas perturbaciones pueden llevar a erróneas interpretaciones de la fuente u origen de DP. El daño que sufre la aislación eléctrica del dispositivo depende de la fuente de DP, por lo tanto, la separación e identificación de las fuentes de DP y ruido eléctrico son cruciales para efectuar un diagnóstico más preciso. Para mejorar la identificación de fuentes de DP dos métodos son propuestos y comparados: mapas Power-Ratio (PR) y Time-Frequency (TF). Ambos crean mapas bi-dimensionales, donde cada punto representa una DP, y cada coordenada una ecuación matemática correspondiente a su tipo de mapa. Estos métodos son estudiados para la separación de múltiples fuentes de DP que ocurren simultáneamente.

In this paper, two probabilistic adaptive algorithms for jointly detecting active users in a DS-CDMA system are reported. The first one, which is based on the theory of hidden Markov models (HMM's) and the Baum-Wech (BW) algorithm, is proposed within the CDMA scenario and compared with the second one, which is a previously developed Viterbi-based algorithm. Both techniques are completely blind in the sense that no knowledge of the signatures, channel state information, or training sequences is required for any user. Once convergence has been achieved, an estimate of the signature of each user convolved with its physical channel response (CR) and estimated data sequences are provided. This CR estimate can be used to switch to any decision-directed (DD) adaptation scheme. Performance of the algorithms is verified via simulations as well as on experimental data obtained in an underwater acoustics (UWA) environment. In both cases, performance is found to be highly satisfactory, showing the near-far resistance of the analyzed algorithms.

The identification of fast and slow waves propagating through trabecular bone is a challenging task due to temporal wave overlap combined with the high attenuation of the fast wave in the presence of noise. However, it can provide valuable information about bone integrity and become a means for monitoring osteoporosis. The objective of this work is to apply different coded excitation methods for this purpose. The results for single-sine cycle pulse, Golay code, and chirp excitations are compared. It is shown that Golay code is superior to the other techniques due to its signal enhancement while exhibiting excellent resolution without the ambiguity of sidelobes.

Electrical trees are one of the main degradation processes leading to failure of high voltage polymeric insulation. Electrical trees grow under the effect of partial discharges (PD), which can be measured and analyzed for condition monitoring of electrical insulation. In this paper, techniques that are normally used for classification of PD and noise separation were explored in their ability to determine the stage of growth of electrical trees: Spectral Power Clustering Technique (SPCT), Time-Frequency (TF) maps and Chromatic Technique (CT). The techniques allowed to analyze PD signals captured in ultra-high frequency (UHF) range with an antenna during tree growth. Laboratory treeing-samples were made of epoxy resin and trees were generated at six different excitation frequencies: 0.1, 10, 50, 150, 250 and 350 Hz. The results showed that two parameters, part of SPCT and TF maps, were sensitive to the tree progression and showed a consistent relation with the length of the tree: the low-frequency power ratio and the equivalent bandwidth. These two parameters were selected to create a new map, proposed for the characterization of electrical tree growth, which is more consistent and robust than the original separation maps. It was found that the low-frequency content of PD pulses proportionally increased with tree propagation. INDEX TERMS Chromatic technique, electrical trees, electrical insulation, partial discharges, source separation techniques, spectral power clustering technique, time-frequency.

La monitorización y medida de las descargas parciales en las redes eléctricas e instalaciones de alta tensión para evaluar el estado del aislamiento de los activos es un tema de gran actualidad, sin embargo, no existen criterios ni procedimientos para evaluar las características técnicas de estos sistemas de medida y diagnóstico. Los sistemas de cable se monitorizan con sensores de tipo HFCT por tener alcances de medida de varios kilómetros cada sensor. En las instalaciones GIS, los sensores invasivos de tipo UHF son ampliamente utilizados, al presentar alta sensibilidad para detectar pequeñas DP pese a su corto su alcance de pocos metros y para los transformadores de potencia se utilizan las capacidades de las bornas capacitivas, como condensador de acoplamiento acoplado a una impedancia de medida. Aunque aparentemente no existe una solución universal para resolver el mismo problema “detección del defecto del aislamiento eléctrico”, en realidad no son bien conocidas las capacidades...

A soundfield image is a data structure that efficiently encodes and represents the wave field as captured by a microphone array. Its representation is based on the directional plenacoustic function, which is defined as the radiance of the acoustic paths (rays) that cross the segment that the array lies upon. The soundfield image can be processed "as is" to develop a variety of applications. In its original formulation, the soundfield image is based on a Euclidean parameterization that can accommodate a limited range of rays and is suitable for managing a single array only. In this paper, we generalize this methodology to the case of multiple microphone arrays deployed in space. The use of multiple arrays allows us to capture truly global information on the sound field but requires us to rethink the ray space, and adopt a global representation of the acoustic rays based on projective geometry. After introducing the new parameterization, we present two examples of applications: the estimation of the mutual poses of two or more arrays (self-calibration); and the localization of multiple acoustic sources. The effectiveness of these applications is proven through simulations as well as real data experiments.

In this work a non-standardized method for on-line diagnostic of high capacity motors based on Ultra Wide Band (UWB) Partial Discharge (PD) measuring techniques is theoretically supported, implemented and applied. The method developed is non invasive and consist in the measurement of PD in the main motor power supply cables and its ground shields connection using a near field sensor (Rogowski coil, clamp-on type) with a bandwidth of 2 to 40 MHz. The measured signals by the sensor, are stretched electronically, digitized and fed to a conventional Partial Discharge digital detector. The UWB PD detection system used displays the output in a Pulse repetition frequency-Charge-Phase angle (N-Q-Φ) PD pattern. The results obtained in thirteen 2500 H.P., 13.8 kV motors performed during operating in an Oil Pumping facility are presented, including its analysis and comparison with values and PD patterns reported in literature.