Study of acoustic signals for the neutrino detector AMADEUS-ANTARES

UNIVERSIDAD POLITECNICA DE VALENCIA E S C U E L A P O L I T E C N I C A S U P E R I O R DE G A N D I A M a s t e r e n I n g e n i e r í a A c ú s t i c a “Study of acoustic signals for the neutrino detector AMADEUS-ANTARES” TESIS DE MASTER Autor: María Saldaña Coscollar Director/es: D. Miguel Ardid Ramírez D. Joan Martínez Mora GANDIA, 2011 2 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES Study of acoustic signals for the neutrino detector AMADEUS-ANTARES Autor: María Saldaña Coscollar Director1: Miguel Ardid Ramírez Director2: Joan Martínez Mora Resumen — En este se proyecto estudia aspectos acústicos pertenecientes al proyecto europeo ANTARES. Por una parte, se centra en el análisis de las señales acústicas recibidas en el sistema de posicionamiento AMADEUS-ANTARES. El programa desarrollado analiza las señales captadas por los sensores, las analiza de manera individual, aplicando una serie de filtros y transformaciones, extrayendo, la frecuencia, fase y amplitud propia de la señal. A continuación, se calcula el tiempo de subida y de bajada apropiado para la amplitud de la señal, con el que se obtiene su duración. La aplicación ha sido diseñada con objetivo de automatizar la detección y análisis de las señales recibidas del sistema, consiguiendo mayor precisión en la reconstrucción del posicionamiento de fuentes. Por otra parte, se lleva a cabo el diseño y calibración de un array lineal de ocho hidrófonos que genera, de manera artificial, el pulso acústico bipolar generado por el neutrino de energía ultra alta en su interacción con el agua. El objetivo del array es transmitir al detector acústico de ANTARES la señal bipolar con similares características en amplitud, forma y directividad a la señal acústica creada por el neutrino en el agua. La calibración ha sido desarrollada combinando métodos de procesado de señal junto con medidas experimentales. El array de hidrófonos es posicionado coherentemente en fase, a distancias conocidas. Aplicado el retraso correspondiente, la señal acústica bipolar es obtenida con mayor amplitud y directividad. La simulación del array acústico es estudiada con el objetivo de conocer la longitud y diseño idóneo para la correcta transmisión de la señal bipolar hasta el detector. Abstract — This academic work study acoustic aspects of the ANTARES detector European project. In one hand, the signal analysis processing for the acoustic signals received by the AMADEUS-ANTARES system. The software developed analyses the received signal of each sensor, applying specific filters and transformations in the frequency and time domain, getting the phase, frequency and amplitude of the signal. For the next step, it fixes the Rise Time and Fall Time appropriated for the signal and the signal duration is obtained. The application is developed for automating the detection and analysis of the received signals from the positioning system, achieving a higher accuracy. On the other hand, an eight lineal array hydrophone is designed and calibrated for generate the artificial Ultra High Energy (UHE) neutrino-induced pulses. The array goal is the transmission of the bipolar acoustic signal pulse mimicking (simulating) the neutrino acoustic signal in terms of amplitude, shape and directivity to the ANTARES detector. The calibration is developed using signal processing methods within experimental measures. The array hydrophone is located coherently in phase with known distances. Hence, the delay between them is applied to get a higher amplitude and directivity acoustic bipolar pulse. The acoustic array simulation is studied in order to know the length and design for the proper bipolar signal transmission to the detector. Autor: Maria Saldaña Coscollar, email: [email protected] Fecha de entrega: 5-09-2011 Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 3 INDEX I. Introduction (Spanish/English) .................................................................................................4 I.1. Background ............................................................................................................................5 I.2. Objectives ..............................................................................................................................5 I.3. Researching environment ......................................................................................................6 II. Introduction to the underwater neutrino telescopes ..............................................................6 II.1. UHE neutrino detector .........................................................................................................6 II.2. ANTARES detector ........................................................................................................... 13 II.3. The ANTARES Modules for Acoustic DEtection Under the Sea (AMADEUS) ..............14 III. AMADEUS acoustic data analysis .......................................................................................20 III.1. Introduction .......................................................................................................................20 III.2.Programming environment .................................................................................................21 III.3. Analysis execution ............................................................................................................21 III.3.1. Data Signals ..........................................................................................................21 III.3.2. Results and analysis .............................................................................................. 21 III.3.3. Analysis conclusion ..............................................................................................27 III.3.4. Final system methodology .................................................................................... 28 III.3.5. Study of problematic cases .....................................................................................29 III.4. Conclusions .......................................................................................................................32 IV. Acoustic array calibration .....................................................................................................32 IV.1. Introduction .......................................................................................................................32 IV.2. Methodology .....................................................................................................................33 IV.2.1. Signal processing techniques for single hydrophone calibration ..........................33 IV.3. Experimental setup and results..........................................................................................36 IV.3.1. Acoustic underwater system .................................................................................36 IV.3.2. Hydrophone array calibration ............................................................................... 38 IV.3.3. Hydrophone array test and prediction ...................................................................39 IV.4. Acoustic array simulation .................................................................................................43 IV.4.1. Introduction ...........................................................................................................43 IV.4.2. Array simulation modelling ...................................................................................43 IV.4.3. Comparison and results .........................................................................................45 IV.5. Conclusions .......................................................................................................................46 Acknowledgements ...................................................................................................................... 47 References .................................................................................................................................... 47 Appendix ...................................................................................................................................... 49 4 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES I. INTRODUCCIÓN I.1 PLANTEAMIENTO DEL PROBLEMA En esta tesina se desarrollan dos casos acústicos relacionados con el telescopio submarino de neutrinos implantado en el mar Mediterráneo, el detector ANTARES (Astronomy Neutrino Telescope and Abyss RESearch) [1]: el estudio de los datos acústicos recibidos en el sistema AMADEUS (Antares Modules for Acoustic DEtection Under the Sea) [2] para el sistema de posicionamiento, y por otra parte, la generación artificial del pulso acústico bipolar del neutrino en su interacción con el agua, mediante la calibración y diseño de un array lineal de hidrófonos. I.2 OBJETIVO Y ALCANCE DEL PROYECTO Los objetivos concretos del proyecto son los siguientes: - Análisis y estudio de señales acústicas recogidas por el sistema AMADEUS, desarrollando, una aplicación de procesado de la señal con el fin de mejorar la detección de señales acústicas del sistema de posicionamiento y mejorar su precisión y exactitud, así como, la búsqueda de posibles fuentes acústicas puntuales que puedan afectar en la detección del sistema de posicionamiento. - Calibración de un array lineal acústico compuesto por ocho hidrófonos encargado de generar artificialmente el pulso bipolar generado por el neutrino en su interacción con el mar. Alcanzando mayor precisión en la transmisión de datos debajo del agua, con el preciso control de amplitud y fase de la señal, siendo útil para la detección y calibración en el sistema AMADEUS-ANTARES. Estos estudios son llevados a cabo con el fin de comprobar y mejorar el estado de operación del sistema AMADEUS perteneciente al detector submarino de neutrinos ANTARES. I.3 CONTEXTO DE LA INVESTIGACIÓN Por una parte, para el análisis de las señales acústicas captadas por el sistema AMADEUS, ha sido necesario el estudio del procesado de señales acústicas y análisis de datos, por lo que, documentación [3], software y aplicaciones informáticas han sido requeridos. Principalmente, se ha utilizado el software matemático computacional Matlab (MathWorks) para el desarrollo del programa de análisis. Todo el material necesario en su desarrollo ha sido otorgado por el personal e instalaciones del Departamento de Física Aplicada (DFA) en colaboración con el Instituto de Investigación para la Gestión Integrada de Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 5 zonas Costeras (IGIC) de la Escuela Politécnica Superior de Gandía (EPSG), Universidad Politécnica de Valencia (UPV). Por otra parte, para la calibración del array acústico ha sido necesario el estudio de software y herramientas de procesado de señal para sistemas de simulación y generación de señales acústicas en el agua, e identificación de sistemas, utilizando para ello el software Maltlab (MathWorks) y LAbVIEW (National Instruments). Así como, en la utilización de la instrumentación, dispositivos y equipamiento utilizados en el desarrollo experimental en el tanque de agua del laboratorio.Este trabajo se ha llevado a cabo en el Departamento de investigación Research School of Computing, Engineering and Information Sciences (CEIS) perteneciente a Northumbria University (Newcastle, UK). Northumbria University participa en la colaboración ACorNE (Acoustic Cosmic Ray Neutrino Experiment) [4,5]. I. INTRODUCTION I.1 BACKGROUND This thesis develops the researching of two acoustic aspects from the neutrino submarine telescope installed in the Mediterranean Sea, the ANTARES (Astronomy Neutrino Telescope and Abyss RESearch) [1] detector where the AMADEUS (ANTARES Modules for the Acoustic Detection Under the Sea) [2] system is integrated: the study of the acoustic data received in the AMADEUS system for the positioning system, and on the other hand, the generation of the artificial Ultra High Energy (UHE) neutrino-induced pulse using an eight lineal array hydrophone. I.2 OBJECTIVES The specific goals of the project are: - To study and analysis of the acoustic signal taken by the AMADEUS system, developing a signal processing application. To the purpose of improving the acoustic signals detection for the positioning system in accuracy and precision, as well as the seeking of acoustic sources which could affect on the detection for the positioning system. - To calibrate an acoustic array of eight hydrophones which generates artificially the neutrino-induced bipolar pulse, achieving higher transmission data accuracy underwater, using the precise control of the amplitude and phase of the signal, being useful for the AMADEUS-ANTARES system calibration and detection. 6 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES This research has been done to check, test and improve the AMADEUS system operation state, belonged to the ANTARES neutrino submarine detector. I.3 RESEARCHING ENVIRONMENT On the one hand, for the acoustic analysis of the data recorded by the AMADEUS system, the study of acoustic analysis processing has been needed as theoretical information [3] as software and informatics application. Mainly, the Matlab (MathWorks) mathematical software has been used for the development of the analysis program.The needed material during the work progress has been given by the staff and installations of the ‘Departamento de Fisica Aplicada ‘(DFA) in collaboration with the ‘Instituto de Investigación para la Gestion de Zonas Costeras’ (IGIC) and the ‘Escuela Politécnica Superior de Gandía’ (EPSG), of the ‘Universidad Politécnica de Valencia’ (Valencia, Spain). On the other hand, for the acoustic array calibration has been needed the study of signal processing analysis software and features for the underwater acoustic signal generation and simulation, using the Matlab software and LabVIEW(National Instruments), as well as, for the instrumentation, devices and equipment used has been required during the experimental measures on the water tank laboratory. This research has been developed on the Research School of Computing, Engineering and Information Sciences (CEIS) of the Northumbria University (Newcastle,UK). Northumbria University is the member of Acoustic Cosmic Ray Neutrino Experiment (ACoRNE) collaboration [4,5]. II. INTRODUCTION TO THE UNDERWATER NEUTRINO TELESCOPES II.1. UHE NEUTRINO DETECTION II.1.1 Neutrino interest The interest for the observation and detection of very high energy (10 GeV-10 PeV) neutrinos has grown in recent years in order to study astrophysical objects of the Universe such as Active Galactic Nuclei, Gamma Ray Burst, Supernova remnant, or Dark Matter [6]. It could be very helpful as well in particle physics since it will allow testing astrophysical models through the neutrino interaction cross-sections or by comparing hadronic and electromagnetic processes. To obtain a good and complete description of the Universe it would be certainly interesting to observe and understand the internal composition of the Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 7 astrophysical object, to have observations and measurements in a large energy range and have the possibility of detecting particles from distant sources. All these aspects would be possible if there would be a messenger, which was easily produced in astrophysical sources, with no electrical charge, so it would not be affected by magnetic fields, and with small probability of interaction with the matter, so they can escape of dense media and travel long distances. Neutrino belongs to a family of particles called “Leptons” (from the Greek word leptos meaning small or fine). Leptons exist in three flavours and each forms a couplet of one charged and one neutral particle as shown in the below figure: Figure 1. The three doublet family charged and neutral leptons Neutrino is An elementary fermion particle without electric charge. There are three kinds of neutrinos: electron neutrino (Ve), muon neutrino (Vu) and tau neutrino (Vt). And it has a very little mass but it is not zero. Specific properties of the neutrino particle make it very interesting to do astrophysics. As it has been pointed before, the interaction probability of the neutrino with the matter is very small, meaning that it will not be affected in its trajectory through the Cosmos to the Earth. It can, as well, escape from dense media allowing us to study the inner working mechanism of astrophysical objects. In addition, it is a neutral particle so magnetic fields do not affect the path along its Universe. On the other hand, the low probability of interaction makes difficult to detect them, and therefore, it is essential that neutrino telescopes have necessarily big volume detection, especially for the case of energetic neutrinos where the fluxes are not so high. Neutrino detection is based on the measurement of the Cherenkov light induced by secondary leptons like muon produced by neutrino interactions with matter while passing across the Earth (Fig. (2)). Searches for cosmic ray neutrinos are ongoing in AMANDA [7], IceCube [8], ANTARES [1], NESTOR [9], NEMO [10], KM3NeT [11] and at Lake Baikal [12] detecting upward going muons from the Cherenkov light in either ice or water. In general, the muons detection from the Cherenkov light in either ice or water is sensitive to lower energies than discussed here since the Earth becomes opaque to neutrinos at very high energies. The experiments could detect almost horizontal higher energy neutrinos but have limited target volume due to the attenuation of the light signal in the media. Also, there are ongoing experiments to detect the neutrino interactions by either radio or acoustic emissions from the resulting particle showers because of the neutrino interaction in liquids [2, 13, 14]. These techniques, with 8 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES much longer attenuation lengths, allow very large target volumes utilising either large ice fields or dry salt domes for radio or ice fields and the oceans for the acoustic detection. II.1.2. Underwater neutrino telescope The main goal of underwater neutrino telescopes is the observation of astrophysical objects like Active Galactic Nuclei (AGN), Gamma Ray Bursts (GRBs), Microquasars, or Supernova Remnants. In fact, astrophysical objects that are able to accelerate protons and nuclei in a wide range of energy are possible candidates to be considered as neutrino sources due to a possible interaction of these accelerated cosmic rays with matter or radiation located in the surroundings of the source. Other interesting applications of underwater neutrino telescopes will be the indirect detection of dark matter using neutrinos generated by the annihilation of weakly interacting massive particles (WIMPs) captured by celestial bodies like the Earth, the Sun or even the Galactic Center. Finally, these detectors are also good for the understanding of some particle physics processes involving for example neutrino oscillations or neutrino interaction crosssections. In addition, and due to the particular location of underwater neutrino telescopes, these infrastructures offer the possibility to extend the scientific program beyond the astroparticle physics field by including in the facility different sensors for Earth and Sea studies. The continuous monitoring of these sensors in real time is a powerful tool for environmental studies, and to bring some light in the understanding of different multidisciplinary problems such as ocean dynamics, climate change, etc. Among the different kinds of neutrino interactions, the muon channel is the most suitable for astroparticle observation and the orientation and depth of the optical modules are selected to optimise the performance of this detection channel. Both, ice and water, neutrino telescopes have been designed taking advantage of the Earth as a shield, which stops all particles except neutrinos (very high energy neutrinos can also be stopped by the Earth). For this reason, we have to be able to detect upward-going muons that come from interactions and discriminate them from downward atmospheric muons. To reduce the huge amount of atmospheric muon background it is necessary that the neutrino telescope is installed in a deep site. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 9 Figure 2. Neutrino interaction and detection scheme. II.1.3. Deep-Sea Acoustic Neutrino Detection In order to test the feasibility of detecting such neutrinos by the acoustic technique it is necessary to understand the production, propagation and detection of the acoustic signal from the shower induced by an neutrino interaction in a medium, incorporing the true attenuation of the sound which has been found to be complex in nature in media such as sea water. Such complex attenuation causes dispersion of the acoustic signal and complicates both the propagation of the sound through the water and the signal shape at the detectors. This aspects has been explained in a previuos publication [14]. II.1.4. The Thermo-Acoustic model Measuring acoustic pressure pulses in huge underwater acoustic arrays is a promising approach for the detection of cosmic neutrino with energies exceeding 100 PeV. The pressure signals are produced by the particle cascades that evolve when neutrinos interact with nuclei in water. The resulting energy deposition in a cylindrical volume of a few centimetres in radius and several metres in length (from 5m to 10 m) lead a local heating of the medium which is instantaneous with respect to the acoustic process. This temperature change induces an expansion or contraction of the medium depending on its volume expansion coefficient. According to the thermo acoustic model, the accelerated motion of the heated volume a micro-explosion forms a pressure pulse of bipolar shape which propagates in the surrounding medium. 16 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES are positioned at heights of 380 m, 395 m, and 410 m above the sea floor. With this setup, the maximum distance between two acoustic storeys is 340 m AMADEUS hence covers three length scales: spacing of the order of 1m between sensors within a storey (i.e. an acoustic cluster); intermediate distances of 14.5 m between adjacent acoustic storeys within a line; and large scales from about 100 m vertical distance on the IL up to 340m between storeys on different lines. The sensors within a cluster allow for efficient triggering of transient signals and for direction reconstruction. The combination of the direction information from different acoustic storeys yields (after verifying the consistency of the signal arrival times at the respective storeys) the position of an acoustic source. The AMADEUS system includes time synchronization and a continuously operating data acquisition setup and is in principle scalable to a large-volume detector. Figure 4 A sketch of the ANTARES detector. The six acoustic storeys are highlighted and their three different setups are shown (see text for details). L12 and IL denote the 12th detection line and the Instrumentation Line, respectively. II.3.2. AMADEUS Acoustic Storeys Two types of sensing devices are used in AMADEUS: hydrophones and Acoustic Modules (AMs). The sensing principle is in both cases based on the piezo-electric effect. For the hydrophones, the piezoelements are coated in polyurethane, where as for the AMs they are glued to the inside of standard glass Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 17 spheres which are normally used for Optical Modules. The acoustic storeys on the IL house hydrophones only, where as the lower most acoustic Storey of Line 12 holds AMs. The hydrophones are mounted to point upwards, except for the central acoustic Storey of Line12, where they point downwards, an example is shown in the Fig.5. The sensitivity of the hydrophones is largely reduced at their cable junctions and therefore shows a strong dependence on the polar angle. The different configurations allow for investigating the anisotropy of ambient noise, which is expected to originate mainly from the sea surface. Three of the five storeys holding hydrophones are equipped with commercial models, dubbed “HTI hydrophones”, and the other two with hydrophones, developed and produced at the Erlangen Centre for Astroparticle Physics (ECAP). Figure 5. Drawing of a standard acoustic storey (acoustic cluster), with hydrophones. (a) (b) (c) Figure 6. Photographs of three different storeys of the AMADEUS system during their deployment: (a) A standard storey, equipped with hydrophones pointing up; (b) the central acoustic storey on Line 12 with the hydrophones pointing down; (c) the lowermost acoustic storey on Line 12 equipped with Acoustic Modules. 18 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES Three of the five storeys holding hydrophones are equipped with commercial hydrophones 2 and the other two with hydrophones developed and produced at the Erlangen Centre for Astroparticle Physics (ECAP). All acoustic sensors are tuned to be sensitive over the whole frequency range of interest from 1 to 50 kHz with a typical sensitivity around −145 dB ref 1V/μPa (including preamplifier) and to have a low noise level [22]. The sensitivity of a typical hydrophone is shown in Fig.7 as a function of frequency and the azimuthal angle [23]. The equivalent inherent noise level in the frequency range from 1 to 50 kHz is about 5.4 mPa for the AMADEUS hydrophones with the smallest such noise. This compares to 6.2 mPa of the lowest expected ambient noise level in the same frequency band for a completely calm sea [24], referred to as sea state 0 [25]. a) b) c) Fig 7. Typical sensitivity as a function of frequency for different polar angles, averaged over the azimuthal angle. Systematic uncertainties below 50 kHz are 2 to 3 dB; a) HTI hydrophone b) ECAP hydrophone c) AM sensor. Bipolar signals selected with the pulse shape recognition trigger typically have a signal-to-noise ratio exceeding 2 for a single sensor. Assuming a noise level of 10 mPa in the frequency range from 1 to 50 kHz, which is a typical level recorded at calm sea, i.e. sea state 0, a signal of 20 mPa can be detected. II.3.3. OnShore Data Processing An on-shore computer cluster is used to process and filter the data stream and store the selected events. The system is operating continuously and automatically, requiring only little human intervention. It currently consists of four server-class computers, of which two with a total of 12 cores are used for data Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 19 triggering. One of the remaining two computers is used to write the data to an internal 550 GB disk array, while the other is used to operate the software for the online control of the data acquisition and other miscellaneous processes and to provide remote access to the system via the Internet. The AMADEUS trigger searches the data by an adjustable software filter; the events thus selected are stored to disk. This way the raw data rate of about 1.5 TB/day is reduced to about 10 GB/day for storage. Currently, three trigger schemes are in operation [2, 26]: A minimum bias trigger which records data continuously for about 10 s every 60 min, a threshold trigger which is activated when the signal exceeds predefined amplitude, and a pulse shape recognition trigger. For the latter, a cross-correlation of the signal with a predefined bipolar signal, as expected for a neutrino-induced cascade, is performed. The trigger condition is met if the output of the cross-correlation operation exceeds a predefined threshold. With respect to a matched filter, this implementation reduces the run time complexity while yielding a comparable trigger performance. In total, 1.6 TB of data were recorded in 2008 and 3.2 TB in 2009. II.3.4. The Acoustic Positioning System The Acoustic Positioning System (APS) is composed by a set of transceiver hydrophones located at the bottom of each line, some autonomous transponders on the sea floor and five receiver hydrophones located at storeys 1, 8, 14, 22 and 25, to monitor some strategic points of the line and minimise the uncertainties in the shape reconstruction since the positions of these receiver hydrophones are the data input used in the mechanical model fits. Electronic boards manage the settings of the system, for emission, reception, 35 detection, filtering and communication. The emitters generate a tone burst signal of 2 ms in a specific frequency of the [40 kHz – 60 kHz] frequency range [21]. This frequency range is justified, in one hand, because it has an optimal spatial resolution and, on the other hand, by having an acceptable attenuation length at the distances of hundreds of meters. Measurements for the line shape reconstruction are performed every 2 minutes with the completion of an ‘acoustic run’. At the present time each acoustic run consists of fourteen cycles; each individual cycle being associated with the emission of an acoustic signal from a single acoustic emitter. For each cycle, the signal transit time and the intensity level to the different receivers is recorded. Since the time of emission is known, the sound travel time between emitter and each receiver for all lines can be computed. Taking into account the profile of sound velocity obtained by different acoustic sound velocimeters and CTDs located along the detector and considering the sound travel time obtained previously, it is possible to calculate the distances between emitter and receivers. The excellent time resolution of the system of few µs, allows obtaining the distances between fixed transducers with stability better than 1 cm. Finally, the final positions of the receiver hydrophones are obtained by an algorithm based on triangulation method applied to the distances 20 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES between emitters and receivers calculated for the fourteen cycles. This is an important task for the achieving an accurate positioning of unknown sources, studies for the positioning system reconstruction has been done in previous works [27]. The Fig.8 shows the radial displacement of all hydrophones on Line 11 during a 16 week period. Figure 8: Radial displacement of line 11 over 16 week period time. III. AMADEUS ACOUSTIC DATA ANALYSIS III.1. INTRODUCTION The precise determination of the relative positions of the Acoustic Storeys within ANTARES detector is a significant task of the AMADEUS system in order to provide an accurate positioning of unknown sources. Reconstruction methods have been analysed in previous studies [27]. The method used is to calculate the differences between the absolute time of signal emission and reception from several pingers to reconstruct the position of each hydrophone individually. The differences in arrival times of a pinger signal in the 6 hydrophones of a storey are used to reconstruct the direction of a pinger signal. Therefore, the adequate rise time and fall time determination for the reception signals at the hydrophones is an important measure to achieve precise signal detection, hence, the positioning reconstruction method can be improved in accuracy. In this work, first studies are done in order to know the proper rise time and fall time determination. Signal processing is applied to analyse the signals received in the hydrophones for the system positioning. III.2. PROGRAMING ENVIRONMENT Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 21 The processing, filtering and analysis of the acoustic data has been done using the software Matlab. This software provides functions really flexible and useful for the acoustic data managing. III.3. ANALYSIS EXECUTION III.3.1. Data Signals The AMADEUS data under studied are the signals recorded from the transceivers (‘pingers’) of the ANTARES acoustic positioning system. The data from the acoustic run 33925 is analysed for this study and the final application is used with data from the acoustic run 33928 for comparison and validation of the system analyses. The acoustic run 33925 has recorded 17 fast cycles at well-defined frequencies within the range from 44522 to 60235 Hz by 6 sensors hold on the same Acoustic Storey. The signals are tone burst of 2 ms, this time duration means 500 samples for each signal, considering that the used frequency sample to register the data was 250 kHz. Each sensor has received the same tones emitted by ‘pingers’ from different lines (positions). In the Fig.9 in left panel (a) the cycle receptions in the Sensor number 6 is shown and in right panel (b) one of the single signal reception is shown. 510 15 20 25 30 -0.2 -0.1 0 0.1 0.2 SENSOR 6: Received signals as function of time Time(s) Amplitude(normalized) 6.27846.27866.2788 6.279 6.27926.27946.27966.2798 6.28 6.28026.2804 -0.3 -0.2 -0.1 0 0.1 0.2 SENSOR 6: Received Signal 7 as function of time Time (s) Amplitude(normalized) (a) (b) Figure 9. (a) Received signals in the Sensor 6 at the storey. (b) Single received signal of Sensor 6. III.3.2. Results and analysis 22 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES The final system application for the determination of the signal processing parameters method is set studying the received signals with less added noise, called ‘clear’ signal, in order to simplify it. For the sensors 6 to 9 the most ‘clear’ signals are the number 5 and 7. The final system application is realized after a previous study which contains these steps: - Band Pass filtering and method - Frequency margin of the Band Pass Filter - Envelope method - Proper % for the rise time (Tr) and fall time (Tf) Band Pass Filtering and method: The first step is filtering the signals to avoid possible added noise from other frequency bands in order to get the interest signal under test. IIR filters are used for the Band Pass Filter study, although the FIR filters have a lineal phase and higher stability, a higher order is required, it means higher signal delay and computational cost (Figures). Hence, the comparison between a Tchebychev Filter and Butterworth Filter is realized, in order to know which one present the best relation between phase linearity level, amplitude and frequency response, as well as the required computational cost (Table 1). For this study, the wide frequency margin of 10 kHz is used. COMPARISON FILTERS Rs=10 Rp=3 Rs=10 Rp=0.1 Rs=40 Rp=3 Butter Tcheby Butter Tcheby Butterworth Tcheby Amplitude Response Very good Bad Good Good Very good Very bad Phase Linearity Very good Bad Good Bad Bad Very Bad Frequency response Very good Bad Good Good Very good Very good Table 1. Amplitude response, Phase linearity and frequency response comparison between Butterworth and Tchebychev Pass Band Filter. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 23 0 2 4 6 8 10 12 x 104 -200 0 200 Frequency (Hz) Phase (degrees) 0 2 4 6 8 10 12 x 104 -100 -50 0 Frequency (Hz) Magnitude (dB) Butterworth Bandpass Filter 41810.6996Hz - 51810.6996Hz 0 2 4 6 8 10 12 x 104 -200 0 200 Frequency (Hz) Phase (degrees) 0 2 4 6 8 10 12 x 104 -100 -50 0 Frequency (Hz) Magnitude (dB) TChebyshev Bandpass Filter 41810.6996Hz - 51810.6996Hz (a) (b) Figure 10. Phase and Magnitude of Band Pass Filter for 10 kHz of frequency margin and Rstop= 10 and Rpass=3: (a) Butterworth filter, (b) TChebychev filter. 6.2785 6.279 6.2795 6.28 6.2805 6.281 -0.2 -0.1 0 0.1 0.2 SENSOR 8 Band Pass Filtered Signal 7 Central Freq. : 46810.6996Hz Time (s) Amplitude(normalized) Original Butter Rp=3 Rs=10 Tcheby Rp=3 Rs=10 4 4.2 4.4 4.6 4.8 5 5.2 5.4 x 104 -40 -20 0 20 40 SENSOR 8 Band Pass Filtered Signal 7 Central Freq. : 46810.6996Hz Frequency (Hz) Amplitude(normalized) Original Butter Rp=3 Rs=10 Tcheby Rp=3 Rs=10 (a) (b) Figure 11. Signal 7 from Sensor 8 filtered by Butterworth and Tchebychev Band Pass filters, for Rp= 3 and Rs= 10 using 10 kHz margin frequency: (a) Normalized amplitude vs time, (b) Zoom of normalized amplitude vs frequency. Frequency margin of the Band Pass Filter: The frequency margin of the Band Pass Filter is studied for achieving a narrow Band Pass Filter which does not modify the original signal shape in terms of amplitude, phase and frequency. The Butterworth Band Pass Filter is analysed for three different margins: 10, 5 and 2.5 kHz (Fig. 12). 24 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES 6.2785 6.279 6.2795 6.28 6.2805 6.281 -0.2 -0.1 0 0.1 0.2 SENSOR 8 Filtered Signal 7 Band Pass at Central Freq. : 46810.6996Hz Time (s) Amplitude(normalized) F.P.B Margin 10KHz F.P.B Margin 5KHz F.P.B Margin 2.5KHz 4 4.2 4.4 4.6 4.8 5 5.2 x 104 -40 -20 0 20 40 SENSOR8 Filtered Signal 7 Paso Banda a Central Freq.: 46810.6996Hz Frequency (Hz) Amplitude(normalized) F.P.B Margin 10KHz F.P.B Margin 5KHz F.P.B Margin 2.5KHz (a) (b) Figure 12. Comparison between the Frequency margin: 10 kHz-5 kHz-2.5 k Hz, applying the Butterworth Band Pass Filter with Rp=3and Rs=10: (a) Normalized amplitude vs time;(b) Zoom of Normalized amplitude vs frequency;(c) Zoom of normalized amplitude vs frequency. Envelope method: The envelope of the signal is equivalent to its outline and an envelope detector connects all the peaks in this signal. Envelope detection has numerous applications in the fields of signal processing and communications, one of which is amplitude modulation (AM) detection. The following block diagram (Fig.13) shows the implementation of the envelope detection using the two methods. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 25 Figure 13. Block diagram of the envelope detection methods; the upper diagram shows the Squaring and Low Pass Filtering Method and the lower diagram shows the Hilbert Transform Method in the Time Domain. The obtaining of an accurate amplitude detection tool is an important task for the signal analysis of the positioning system detection. The two methods described above are studied and compared (Fig. 14). 6.279 6.2792 6.2794 6.2796 6.2798 6.28 6.2802 6.2804 6.2806 6.2808 6.281 0 0.05 0.1 0.15 0.2 SENSOR7: Normalized amplitude of signal 7 Time (s) Amplitude(normalized) Signal Hilbert Envelope SLP Filtering Envelope IIR n=10 SLP Filtering Envelope FIR n=10 6.279 6.2792 6.2794 6.2796 6.2798 6.28 6.2802 6.2804 6.2806 6.2808 6.281 0 0.05 0.1 0.15 0.2 SENSOR7: Normalized amplitude of signal 7 Time (s) Amplitude(normalized) Signal Hilbert Envelope SLP Filtering Envelope IIR n=5 SLP Filtering Envelope FIR n=5 (a) (b) 6.279 6.2792 6.2794 6.2796 6.2798 6.28 6.2802 6.2804 6.2806 6.2808 6.281 0 0.05 0.1 0.15 0.2 0.25 SENSOR7: Normalized amplitude of signal 7 Time (s) Amplitude(normalized) Signal Hilbert Envelope SLP Filtering Envelope IIR n=2 SLP Filtering Envelope FIR n=2 6.279 6.2795 6.28 6.2805 6.281 0 0.05 0.1 0.15 0.2 0.25 SENSOR7: Normalized amplitude of signal 7 Time (s) Amplitude(normalized) Signal Hilbert Envelope SLP Filtering Envelope IIR n=2 (c) (d) Figure 14. Zoom Envelope using Hilbert Method and Squaring-Low pass filtering Method (filter FIR and IIR): (a) Normalized amplitude vs time for order filter n= 10 ;(b) Normalized amplitude vs time for order filter n= 5; (c) Normalized amplitude vs time for order filter n=2. (d) Zoom Envelope using Hilbert Method and Squaring-Low pass filtering IIR of filter order=2, plotting the normalized amplitude vs time. Rise time and fall time: 32 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES IV.4. CONCLUSIONS The different analyses made and the results obtained permit to validate the correct running for the signal processing application to analyze the signals received for the positioning system detector at the AMADEUS system. This application allows obtaining accurate acoustic signal time and amplitude detection via signal processing. It can be considered as tool to improve the accuracy of the positioning system detection and as preliminary phase to realize a vast analysis from the received signal in order to study any possible external noise or behavior which could affect in the positioning detection. In future job, the application can be use as a helpful tool to realize an accurate signal detection as well as to realize a deeper signal analysis of the received data from the positioning system, hence it can be extended in order to get more results and aspects from the signal and system detection which can be of interest in ANTARES or future neutrino telescopes, such a KM3Net. IV. ACOUSTIC ARRAY CALIBRATION FOR GENERATION OF UHE NEUTRINOS IV.1. INTRODUCTION The evaluation of the acoustic neutrino detector behaviour in the field is a significant task under study. Hence, acoustic transmitters are needed to send artificial acoustic signatures from UHE neutrino interactions in the sea to the AMADEUS system. The transmitters have to be able to imitate the UHE neutrino acoustic pulse, in directivity (pancake) and shape (a short bipolar pulse). Furthermore the detector site is located at a depth of 2475 m underwater. Due these limitations, it is not an easy task to design a system able to transmit an acoustic bipolar pulse with such characteristics. The deployment and operation in site have to be considered in terms on the system design as well. The development of a compact transmitter using the parametric acoustic source technique has been studied to achieve the objective [28]. Previous studies in the calibration of single hydrophones for the UHE neutrino emulation were presented [29]. However, a single hydrophone only gives knowledge of the signal shape and the radiation is essentially omni-directional. In reality the pulse from a neutrino interaction should be highly direction. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 33 This can be emulated using a line array. In addition because of the coherent nature of detection, the energy in the transmit direction will depend on the square of the number of hydrophones used – significantly extending the range. Using hydrophone array also has the advantage that would allow precise calibration of existing and planned acoustic telescopes. Recent studies have predicted the acoustic bipolar pulse generated for the neutrino interaction with in the water at a frequency of over 23 kHz. This research realizes an 8 hydrophone array calibration for generation of Ultra High Energy neutrino-induced pulses at the frequency of 23 kHz. Signal processing techniques are applied to array hydrophone modeling and experimental data generated in a laboratory water tank are used and compared on the calibration system. An eight channel arbitrary waveform generation module has been built using 16 bit dsPIC digital signal controllers for driving hydrophone array transmitters. The NI PXI-6713 commercial module is used for validation. IV.2. METHODOLOGY The array calibration development can be classified in three phases. In the first phase, signal processing and measurements in water tank are used for the single hydrophone calibration. The characterisation system of the hydrophones is designed to generate the sought bipolar pulse. In a second phase, the eight transmitters are individually checked and validated for the calibration done. Then, the complete eight hydrophone array is placed in the water tank test with the purpose of studying the bipolar signal amplitude transmitted as function of the tension in the transducer, using both the eight channel PIC module and a commercial module for comparison. In a third phase, the eight hydrophones array is simulated by signal processing where is introduced the real array design side, and the environment conditions (Mediterranean sea attenuation, receiver distance), with the objective of predicting the pattern directivity and shape of the bipolar pulse at the detector, determining the correct 8 hydrophone’s array length which could send the signal with the 1 º ‘pancake’ directivity. IV.2.1 Signal processing techniques for hydrophone calibration Linear Time Invariant (LTI) signal processing can successfully be used to calibrate hydrophones and that standard signal processing techniques can be used to produce arbitrary acoustic waveforms. The LTI model for a transmit hydrophone consists of three parts: an RC electrical circuit coupled to a set of damped mechanical masses and an impedance matching network to couple the hydrophone to the water. This can be represented either in State Space or as a transfer function. 34 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES The output signal response of linear time invariant (LTI) system to any input signal can be determined by using convolution operator [7, 8]. It states that if we know the impulse response h (t) of a system, we can compute the output response y (t) to any input x (t) Alternatively, Fourier and inverse Fourier transforms can be used to transfer between domains (e.g. time to frequency and vice versa). When y (w) is the output response, h (w) is the system function, and x (w) is input signal, as shown in Fig.18. Figure 18. Block diagram of the LTI system to any input signal x(t) . where y (t) is output impulse response and the symbol * denotes convolution. Ideally a Dirac delta function would be used as a driving signal as x (ω) =1 and y (ω) =h (ω). In practice however as an impulse or delta function cannot be easily created, a step signal is chosen in order to generate step response of system. The step response of system is the time integral of its impulse response; hence the impulse response can be calculated from differentiation of known step response. As mentioned, the output response signal is resulted from the convolution between input signal and impulse response. Similarly, input signal can be calculated from de-convolution of the output response and system function as shown in (). It is then transferred to the time domain using inverse Fourier transform as shown in (Eq. (8)). )(/)()( hyx , )()( 2 1dex tj (8) The method, as described above, has three practical disadvantages. Firstly, the data acquisition needs ahigh signal-to-noise ratio as differentiation is used. Secondly, the step signal input must have a very fast rise time and also high signal to noise ratio. Lastly, the experiment needs to run in large volume or anechoic chamber to avoid reflections. The tank which is available in our workshop has limited volume and reflections soon arrive from the tank boundaries. To avoid the problems which are already mentioned above, signal processing techniques are used. The block diagram of signal processing techniques for calibration is shown in Fig .19. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 35 Figure 19. Block diagram of the signal processing techniques used for calibration. The calibration system is implemented in a rectangular glass water test tank of 1.52 m in length, 0.60 m in width and 0.60 m in depth. The eight hydrophones under calibration are identical robust spherical reference hydrophones (RESON TC 4033) used as transducers to transmit the desired acoustic signals. The RESON TC 4033 provides uniform omni-directional characteristics within the full useable frequency range of 1 Hz to 140 kHz. It has a factory-calibrated transmitting sensitivity of 120 dB ref 1µPa/V@ 1m at 23 kHz. Matlab computer software is used to create an ideal step signal which is injected to LabVIEW (NIUSB 6211) software, which has a 16 bit analogue-to-digital converter (ADC) sampled at 250 kHz. It is then amplified using a PA94 power amplifier chip (gain of 17) and it’s sent out to a single transmitter hydrophone in underwater. The output signal is detected by a wide range bandwidth hydrophone (a Bruel&Kjaer type 8106), general-purpose transducer for making absolute sound measurements over the frequency range from 7 Hz to 80 kHz with receiving sensitivity of -173dB ref 1V/µPa. It provides a flat frequency response and omni-directional characteristics over wide range. The output data from the receiver is the hydrophone step response which is recorded for a data acquisition module (Tektronik TDS 2002 B); it is shown in Fig. 20. in the left panel (a), which is the blue color response. The know input to the hydrophone, it means the step signal, is injected into the LTI software model and the outputs the real and model hydrophones are matched by reducing the sum squared error (Fig. 20, in hte left panel (a)). The order of the LTI model is gradually increased to ensure sufficient goodness on fit. A fifth-order scaled transfer function (TF) is fitted to this step response (Eq. (9)). The TF is a mathematical representation of the relationship between the input and output of a LTI system. The input driving signal for 23 kHz bipolar pulse is resulted from replacing fitted transfer function and the desired bipolar signal at 23 kHz. An inverse Fourier transform is then used to convert the input 36 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES driving signal from the frequency to time domain (Fig. 20). This is the driving pulse used as input in the hydrophone transmitter to send the 23 kHz bipolar pulse. 0 1 2 3 4 5 6 7 -1 -0.5 0 0.5 Time (ms) Amplituce (V) Actual response Fitted response -1000 -500 0 500 1000 -1.5 -1 -0.5 0 0.5 1 1.5 Time ( s) Amplitude (V) Phase/Amp Equalised Driving Pulse Bipolar Output (a) (b) Figure 20.(a) Fitting the step response simulated with the measured for the hydrophone transfer function;(b) Input driving pulse and output 23 kHz bipolar pulse simulated. The transfer function H(s) obtained is: 2.436 s² - 0.1226 s - 0.08506 --------------------------------------------------------------------- 0.002913 s 5 + 0.01982 s⁴ + 0.5279 s³ + 2.434 s²+ 18.65 s + 45.9 (9) IV.3. EXPERIMENTAL SETUP AND RESULTS IV.3.1. Acoustic underwater detection system In the experiment, the acoustic underwater detection system comprises of the dsPIC 8 channel arbitrary waveform generator, a National Instruments PXI-6713 8 channel commercial high speed analog output, an eight channel power amplifier, eight TC4033 transmitter hydrophones, a B&K-8106 received hydrophone, a B&K-2690 conditioning amplifier and data acquisition modules (TDS2002B storage oscilloscope, NI PCI-6254 16 bit resolution commercial data acquisition). The block diagram of acoustic underwater detection and the real system at the laboratory are illustrated in Fig. 21and Fig, 22 respectively. Magnitude and Phase Equalised Bipolar driving pulse Step response fitting of transmitter hydrophone Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 37 8 Channel Power Amplifier (PA94) Array of acoustic transmitter (8 of TC4033) An acoustic receiver (B&K-8106) Conditioning amplifier (B&K 2690) Data acquisition (NI PCI-6254 and TDS2002B) Acoustic Signal dsPIC MCU 8CH DAC NI PXI-6713 8CH DAC O/P Figure 21.Block diagram of acoustic underwater detection. Figure 22. Acoustic underwater detection at the laboratory. The 60x152x60 cm3 water tank is used for experiment of the acoustic array at the laboratory. Due to the limitation of dimension of water tank, the hydrophone array is installed in horizontal direction in order to test in along the longest dimension. The water tank and hydrophone installation are illustated in Fig 23. 25 5 for each hydrophone 50 25 60 150 60 55 10 10 10 *All dimensions in centimeter Water level Figure 23. Dimension of water tank and hydrophone installation. 38 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES IV.3.2. Hydrophone array calibration The first experiment is the calibration of each hydrophone using the same driving pulse and distance. The signal output voltage from dsPIC and NI PXI-6713 are set to 4.1 V. The driving pulse is then amplified by power amplifier with a gain of 17. The pulse height after the power amplifier is about 70 V which is sufficient to drive the hydrophones in the laboratory and the field. The testing for calibration of hydrophone is started with dsPIC module. Hydrophone number 1 (H1) is firstly tested and then H2 is also tested but uses the same distance (30 cm between transmitter and receiver) with H1. The received bipolar signals from H1 to H8 are plotted in the Fig. 24 The plot starts from H1 signal in the left hand side until the H8 signal in the right hand side. The hydrophones are extremely well matched. The NI PXI-6713 is also used to calibrate hydrophones to validate the result with dsPIC module. The regular calibration is done as same as the dsPIC module at the same distance. The result of received signal using PXI-6713 is plotted in Fig. 25. 0.01 0.02 0.03 0.04 0.05 0.06 0.07 -1 -0.5 0 0.5 1 Time (ms) Amplitude (Vpeak) Received signal from H1-8 (dsPIC) H1 H8 Figure 24. Results of regular calibration for the eight hydrophones with the dsPIC module. Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 39 0.01 0.02 0.03 0.04 0.05 0.06 0.07 -1 -0.5 0 0.5 1 Time (ms) Amplitude (Vpeak) Received signal from H1-8 (PXI-6713) H1 H8 Figure 25. Results of regular calibration for the eight hydrophones with NI PXI-6713. IV.3.3. Hydrophone array test and prediction The second experiment is the study of the hydrophone array operation, in order to understand the direction and shape of the signals at the receiver. In this experiment, the receive hydrophone is set at three positions. The first position is directly set to the centre of hydrophone transmitters (0 degree of pancake direction). Another two positions are set in the left and bottom side with 10 cm from the centre point to examine off axis behavior. The three positions of received hydrophone is illustrate in Fig 26. The measurements start with setting hydrophone receiver at the centre point. A bipolar signal is individually sent from hydrophone number 1 (H1) until hydrophone number 8 (H8) to know the amplitude of received signal for each hydrophone. In the next step, signals from H1-H8 are simultaneously sent to the receiver. In this step, it is very important to set the delay time between each hydrophone transmitter to synchronize the pulses at the receiver. As the distance between hydrophones are 5 cm, and the velocity of sound in water approximately 1500 m/s, the delay time is set to 33.3 s. The measurements are repeated with the same procedure but the receiver is moved to the left and bottom points. The PXI-6713 is also tested with the same step to validate the result with dsPIC module. The result data from measurements is recorded in Table 4. The measurement values are presented that the case of sending eight hydrophones at the centre position can achieve a highest value because it is the best synchronous position of eight signals when compares to the left and bottom positions. These are displayed in Table 4. If the hydrophones were truly 40 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES omni-directional then the “Left” and “bottom” column should show identical values. The variations are however slight being of the order of a few percent. Moreover, in the case of individual sending hydrophone at the centre position, the results show the relation between the falls off sound amplitude with the distance. The relative pressure amplitude and relative distance of each hydrophone position on the experiment is compared with the theoretical model (1/r α pressure). The comparison of relation between pressure and distance of data from theory, dsPIC and PXI-6713 modules is plotted in Fig. 27. 50 10 Centre Left Bottom H1 H2 H3 H4 H5 H6 H7 H8 5 for each hydrophone Delay 33.3 us for each hydrophone 10 * All dimensions in centimeter Figure 26. Three positions of received hydrophone. No of Hydrophone PXI-6713 module dsPIC module Centre Left bottom Centre Left bottom H1 1.54 1.46 1.56 1.68 1.54 1.58 H2 1.64 1.60 1.66 1.76 1.62 1.60 H3 1.84 1.76 1.74 1.86 1.74 1.72 H4 1.98 1.84 1.80 1.96 1.92 1.86 H5 2.30 2.06 2.08 2.30 2.04 2.10 H6 2.46 2.24 2.30 2.48 2.22 2.30 H7 2.56 2.28 2.36 2.58 2.30 2.32 H8 2.86 2.64 2.78 2.88 2.70 2.74 H1-8 16.4 15.0 15.8 16.6 14.8 16.00 *All units are in Vp-p Table 4. Comparision received signal between dsPIC and ni pxi-6713 modules Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 41 1 1.2 1.4 1.6 1.8 0.5 0.6 0.7 0.8 0.9 1 Relative Distance Relative Sound Pressure Theory PXI-6713 dsPIC Figure 27. The relation between distance and pressure. The Fig. 28 shows the received signal at the centre point by increase the number of hydrophone transmitters from two hydrophones until eight hydrophones (lowest amplitude to highest amplitude). The graph is shown that received signals are determined by addition received signal amplitude of each hydrophone. The comparison of received signal between dsPIC and PXI module for 8 hydrophones is plotted in Fig. 29. The receiver is placed in the centre. The result shows that it is little different between the signal which come from dsPIC and PXI modules. 0 0.02 0.04 0.06 -8 -6 -4 -2 0 2 4 6 8 Time (ms) Amplitude (Vpeak) Received signal from Hydrophones H1-2 H1-8 H1-8 H1-7 H1-6 H1-5 H1-4 H1-3 H1-2 Figure 28.Bipolar pulse receptions from two till eight hydrophones emitting. 48 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES [8] A. Achterberg, First Year Performance of the Icecube NeutrinoTelescope, Astropart. Phys. 26 (2006) 155173;http://arxiv.org/abs/astro-ph/0604450. [9] G. Aggouras, Operation and performance of the NESTOR test detector, Nucl. Instrum. And Meth. A 552 (2005) 420-439. [10] E. Migneco,Recent achievements of the NEMO project, Nucl. Instrum. and Meth. A 588 (2008) 111-118. [11] E. de Wolf, KM3NeT, A new generation neutrino telescope, Nucl. Instrum. and Meth. A 588 (2008) 86-91. [12] V. Aynutdinov, The Baikal neutrino experiment: Status, selected physics results, and perspectives, Nucl. Instrum. and Meth. A 588 (2008) 99-106. [13] J. G. Learned, Acoustic Radiation By Charged Atomic Particles In Liquids: An Analysis, Phys. Rev. D 19, (1979) 3293. [14] S.Bevan, Study of the acoustic siganture of UHE neutrino interactions in water and ice, Nucl.Instr. and Meth.A 607 (2009) 398-411. [15] G. A. Askaryan, Sov. J. Atom Energy 3, 921 (1957). [16] S. Bevan, Simulation of Ultra High Energy Neutrino Interactions in Ice and Water. Astroparticle Physics 28 (2007) 366–379. [17] R. E. Francois, G. R. Garrison, Sound absorption based on ocean measurements: Part I:Pure water and magnesium sulfate contributions, Journal of the Acoustical Society of America 72(3) (1982) 896-907. R. E. Francois, G. R. Garrison, Sound absorption based on ocean measurements: Part II:Boric acid contribution and equation for total absorption, Journal of the Acoustical Society of America 72(6) (1982) 1879-1890. [18] M. A. Ainslie and J. G. McColm, A simplified formula for viscous and chemical absorption in seawater, Journal of the Acoustical Society of America 103(3) (1998) 1671-1672. [19] L. N. Lieberman, The Origin of Sound Absorption in Water and in Sea Water, The Journal of the Acoustical Society of America 20(6) (1948) 868-873. [20] V. Niess and V. Bertin, Underwater Acoustic Detection of Ultra High Energy Neutrinos, Astropart. Phys. 26(4-5) (2006) 243-256;http://arxiv.org/abs/astro-ph/0511617. [21] P.Keller, Acoustic Positioning system for the Deep-Sea ANTARES Neutrino telescope, sensorcomm, pp243-247 (2007). [22] C. Naumann, Proc. of the Int. Workshop (ARENA 2005), World Scientific Publishing, Singapore (2006) 92; preprint: arXiv:astro-ph/0511243. [23] C. Naumann, Ph.D. Thesis, Univ. Erlangen-Nürnberg, FAU-PI4-DISS-07-002, 2007. [24] K. Graf, Experimental studies within ANTARES towards acoustic detection of ultra-high energy neutrinos in the deep sea, Ph.D. thesis, Univ. Erlangen-N¨urnberg, 2008, FAU-PI1-DISS-08-001. [25] R.J. Urick, Ambient Noise in the Sea, Undersea Warfare Tech. Office, Naval Sea and Systems Command,Dept. of the Navy, Washington, D.C(1984). Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 49 [26] M. Neff, Studie zur akustischen Teilchendetektion im Rahmen des ANTARES-Experiments: Entwicklung und Integration von Datennahmesoftware, Diploma thesis, Univ. Erlangen-N¨urnberg, 2007, FAU-PI1-DIPL-07003. [27] M.Bou-Cabo, PhD. Thesis: Acoustics for underwater neutrino telescopes. Universidad Politécnica de Valencia(2011.01). [28] M. Ardid, S.Adrian, et. at. R&D studies for the development of a compact transmitter able to mimic the acoustic signature of a UHE neutrino interaction. Nucl. Instr. and Meth. A (2010), doi:10.1016/j.nima.2010.11.139. [29] W. Ooppakaew,S. Danaher, Hydrophone Calibration Based on Microcontrollers for Acoustic Detection of UHE Neutrinos. Nucl.Instr. and Meth.A (2011),doi:10.1016/j.nima.2010.11.142. [30] S. T. Karris, Signals and Systems with MATLAB Applications,California: Orchard Publications(2003). [31] V. K. Ingle, J. G. Proakis, Digital Signal Processing Using MATLAB. USA: Bookware Companion Series (2007). [32] W.Ooppakaew, S.Danaher, M.Saldaña, Acoustic Array Calibration and Signal Processing for UHE Neutrinos Generation, MARSS 2011. Northumbria University and Universidad Politécnica de Valencia, APPENDIX Appendix 1. Matlab code with the fynal system application for the received signal análisis from the AMADEUS-ANTARES positioning system. clear all clc close all %% ------------- ******* APLICABLE DEL SENSOR 6 al 9 ******** --------- %% [x,fs,bits]=wavread('sensor7.wav'); % x: muestras de la señal en el tiempo,fs:frecuencia de muestreo,bits=número de bits para codificar 1 muestra fs ts=1/fs bits s=7;%NUM SENSOR (para plots) I=0.15*0.25; % inicio (muestras) señal 1º recepción F=0.15*0.05; % final(muestras) señal 1º recepción %tamaño(muestras) total 1º recepción fin1=300000; num_muestrasI=zeros; num_muestrasF=zeros; cont1=1; 50 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES cont=1; % ------------------------------- RECEPCIÓN 1 a la 11 ----------------------- - % % ------------------ ENVOLVENTE ------------------% hilbert=hilbert(x); envelopeH=abs(hilbert); for i=1:1:11 % ------------------- Deteccion de cada senyal ---------------- % if i>=2 cont2=1; cont3=1; ini(i)=ini_r(i-1)+150000; fin(i)=fin_r(i-1)+220000; for j=ini(i):1:fin(i) %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(j)>I num_muestrasI(cont2)=j; cont2=cont2+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for k=ini_fin:1:fin(i) if x(k)<=F & x(k)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont3)=k; cont3=cont3+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F else cont2=1; cont3=1; for j=1:1:fin1 %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(j)>I num_muestrasI(cont2)=j; Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 51 cont2=cont2+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for k=ini_fin:1:fin1 if x(k)<=F & x(k)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont3)=k; cont3=cont3+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F end %-------------- Calculo de la amplitud propia de la señal ------------% for l=(ini_r(i)-100):1:fin_r(i) %muestras de ini y fin de r1 z(l)=envelopeH(l); end ts=1/fs; rango_estable=envelopeH(ini_r(i)+100:fin_r(i)-100); amplitud(i)=mean(rango_estable); I=amplitud(i)*0.25; % inicio (muestras) señal 1º recepción F=amplitud(i)*0.3; for p=1:1:fin1 %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(p)>=I num_muestrasI(cont1)=p; cont1=cont+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción 52 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for q=ini_fin:1:fin1 if x(q)<=F & x(q)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont)=q; cont=cont+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F %creamos el nuevo vector para que posteriormente guarde la señal recibida 1ª entre los margenes del eje horizontal correctos ini_analisis(i)=ini_r(1); ini_analisis(i)=ini_r(i)-200; fin_r(i)=fin_r(i)+150; %z=x(ini_r1:1:fin_r1); for l=ini_analisis(i):1:fin_r(i) za(l)=x(l); hilbert1(l)=hilbert(l); envelopeH1(l)=envelopeH(l); end %**************************************************************% %****** Filtrado Paso-Banda ***************% %**************************************************************% %******** Transformada de Fourier de la señal ************% t=[ini_analisis(i):1:fin_r(i)].*ts; %rango temporal de la señal f=1./(t.*ts);%rango frecuencial de la señal N=length(za(ini_analisis(i):fin_r(i)));%longitud de la señal frec =[(-1/ts)/2:(1/ts)/(N-1):((1/ts)/2)];%rango frecuencial de la tranformada z1a= fftshift(fft(za(ini_analisis(i):fin_r(i)))); figure,subplot(1,2,2),plot(t,za(ini_analisis(i):fin_r(i))),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en tiempo']) subplot(1,2,1),plot(frec,z1a),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en frecuencia(transformada)']) Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 53 %********* Detección de la Frecuencia de la señal, Fp y Fs *******% m=max(z1a) p=find(z1a==m) %En el punto donde capta el pico en la transformada se encuentra su frecuencia frec_senyal=frec(p); %por si coge el valor máximo(correspondiente a la frecuencia) en la parte %imaginaria, quitarle el valor negativo: if frec_senyal<0 frec_senyal=-frec_senyal else frec_senyal=frec_senyal end %-------- Aplicamos Filtro Paso Banda de 35 a 65 KHz a una frecuencia de unos 60KHz ---- % fp_min=40000;%fpass minimo fp_max=60000;%fpass maximo fs_min=35000;%fstop minimo con 2.5KHz de margen fs_max=65000;%fstop maximo con 2.5KHz de margen Wp = [fp_min fp_max]/fs;%Frecuencias que deseas pasar Ws = [fs_min fs_max]/fs;%Tolerancia que ofreces a la horquilla de frecuencias Rp = 1; Rs = 10;%Tolerancias de la banda de rizado [n,Wn] = buttord(2*Wp,2*Ws,Rp,Rs)%Definicion [b,a] = butter(n,Wn); figure,freqz(b,a,bits,fs)%Parametros y ordenes del filtro + numero de puntos + Frecuencia de muestreo title(['Butterworth Bandpass Filter ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]) zaf1(ini_analisis(i):fin_r(i))= filter(b,a,za(ini_analisis(i):fin_r(i)));%aplicamos filtro a la señal z1af1= fftshift(fft(zaf1(ini_analisis(i):fin_r(i)))); figure,subplot(1,2,1),plot(frec,z1af1),title(['ondafiltrada' num2str(i) 'F.Paso Banda (frecuencia): ' num2str((fp_min)) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(1,2,2),plot(t,zaf1(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) 'F.Paso Banda (tiempo): ' num2str((fp_min)) 'Hz - ' num2str((fp_max)) 'Hz' ]); % ---- Aplicamos filtro Banda Paso entre 40-60KHz a la hilbert senyal -----% % --- para obtener la envolvente de la senyal filtrada ---- % 54 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES hilbert1=hilbert1(ini_analisis(i):fin_r(i)); filtro_h(ini_analisis(i):fin_r(i))= filter(b,a,hilbert1); %********* Detección de la Frecuencia de la señal de nuevo , Fp y Fs *******% m=max(z1af1) p=find(z1af1==m) %En el punto donde capta el pico en la transformada se encuentra su frecuencia frec_senyal=frec(p); %por si coge el valor máximo(correspondiente a la frecuencia) en la parte %imaginaria, quitarle el valor negativo: if frec_senyal<0 frec_senyal=-frec_senyal else frec_senyal=frec_senyal end fp_min=frec_senyal-2500%fpass minimo con 2.5KHz de margen desde la frecuencia propia fp_max=frec_senyal+2500%fpass maximo 2.5KHz de margen desde la frecuencia propia fs_min=fp_min-10000%fstop minimo con 10KHz de margen desde la frecuencia propia fs_max=fp_max+10000%fstop maximo con 10KHz de margen desde la frecuencia propia %---------- Rango de Banda de paso de margen 5KHz con frec central=frecuencia de la señal ------% Wp = [fp_min fp_max]/fs;%Frecuencias que deseas pasar Ws = [fs_min fs_max]/fs;%Tolerancia que ofreces a la horquilla de frecuencias Rp = 1; Rs = 10;%Tolerancias de la banda de rizado [n,Wn] = buttord(2*Wp,2*Ws,Rp,Rs)%Definicion [b,a] = butter(n,Wn); figure,freqz(b,a,bits,fs)%Parametros y ordenes del filtro + numero de puntos + Frecuencia de muestreo title(['Butterworth Bandpass Filter ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]) y(ini_analisis(i):fin_r(i))= filter(b,a,zaf1(ini_analisis(i):fin_r(i))); ondafiltrada = fftshift(fft(y(ini_analisis(i):fin_r(i)))); figure subplot(1,2,1),plot(frec,ondafiltrada),title(['ondafiltrada' num2str(i) '(frecuencia): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(1,2,2),plot(t,y(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) '(tiempo): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 55 figure,subplot(2,2,2),plot(t,za(ini_analisis(i):fin_r(i))),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en tiempo']) subplot(2,2,1),plot(frec,z1a),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en frecuencia(transformada)']) subplot(2,2,3),plot(frec,ondafiltrada),title(['ondafiltrada' num2str(i) '(frecuencia): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(2,2,4),plot(t,y(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) '(tiempo): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); figure set(gca,'fonts',8) plot(frec,z1a,frec,z1af1,frec,ondafiltrada),title(['SENSOR' num2str(s) 'OndaFiltrada' num2str(i) '(frecuencia) Paso Banda a Frec.Central: ' num2str(frec_senyal) 'Hz - ']) legend('S. Original' ,'S. F.P.B 40-60KHz','S. F.P.B Margen 5KHz') figure set(gca,'fonts',8) plot(t,za(ini_analisis(i):fin_r(i)),t,zaf1(ini_analisis(i):fin_r(i)),t,y(ini_ analisis(i):fin_r(i))),title(['SENSOR' num2str(s) 'OndaFiltrada' num2str(i) '(tiempo) Paso Banda a Frec.Central: ' num2str(frec_senyal) 'Hz - ']) legend('S. Original' ,'S. F.P.B 40-60KHz','S. F.P.B Margen 5KHz') pause % --------------- Método 'HILBERT TRANSFORM' ----------% % ---- Aplicamos filtro Banda Paso 5KHz a la hilbert senyal filtrada -----% % --- entre 40-60KHz para obtener la envolvente de la senyal filtrada ---- % filtro_h= filter(b,a,filtro_h(ini_analisis(i):fin_r(i))); envelopeH_filtrado(ini_analisis(i):fin_r(i))=abs(filtro_h); fase=unwrap(angle(filtro_h)); %fase fx=diff(unwrap(angle(filtro_h)))./diff(t)/(2*pi); %frecuencia %------------------ amplitud ------------ % figure,subplot(3,1,1),plot(t,y(ini_analisis(i):fin_r(i)),t,envelopeH_filtrado (ini_analisis(i):fin_r(i)),'r:'),title(['SENSOR' num2str(s) ':Muestras y Envolvente(Mediante HILBERT)de la Señal recibida ' num2str(i) ' en el tiempo parte estable']); xlabel('t(segundos)'); ylabel('x(t) normalizado') grid on 56 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES %------------------ fase ------------ % subplot(3,1,2),plot(t,fase),title(['SENSOR' num2str(s) ':Fase de la Señal recibida ' num2str(i) ' en el tiempo']); xlabel('t(segundos)'); ylabel('fase (t)') grid on %s=[z1 z]; %s=[z za];%Aqui quedan guardadas todas las señales su número de muestras correspondiente %------------------ frecuencia ------------ % subplot(3,1,3),plot(t(2:end),fx),title(['SENSOR' num2str(s) ':Frecuencia de la Señal recibida ' num2str(i) ' en el tiempo']); xlabel('t(segundos)'); ylabel('frecuencia (t)') grid on % -------------------- Cálculo del tiempo de subida --------------------% % Calculamos de nuevo la amplitud, en la señal Filtrada, en la parte central % de la senyal, ya que, es normalmente, donde tiene un nivel de amplitud % mas estable rango_estable=envelopeH_filtrado(ini_analisis(i)+300:fin_r(i)-250); amplitud(i)=mean(rango_estable) % ---------------------- y1(ini_analisis(i):fin_r(i))=envelopeH_filtrado(ini_analisis(i):fin_r(i)); a2=amplitud(i)*0.25; i2=find(y1>=a2); i2(i)=min(i2); format long Ts_25(i)=i2(i)*ts y1=0; %--------------------------- % -------------------- Cálculo del tiempo de bajada --------------------% ini_Tb=ini_analisis(i)+350; fin_Tb=fin_r(i);%-50? %asi nos hemos desplazado a la parte final para el cálculo del Tb y2=envelopeH_filtrado(ini_Tb:fin_Tb);% se tiene que tener en cuenta que el 1ºvalor de y2 realmente corresponde a la muestra ini_Tb,para el posterior cálcuo del tiempo a3=amplitud(i)*0.75; f3=find(y2<=a3); f3(i)=min(f3)+ini_Tb; format long Tb_75(i)=f3(i)*ts Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 57 duracion_muestras25_75(i)=-f2(i)+f4(i); duracion_tiempo25_75(i)=(duracion_muestras25_75(i)).*ts; y2=0; pause end % ------------------------------- RECEPCIONES 12,13,14,15 ------------------- ----- % ini_r(11)=4030000; fin_r(11)=4030520; %Nos desplazamos al siguiente grupo de recepciones for i=12:1:15 % ------------------- Deteccion de cada senyal ---------------- % if i>=2 cont2=1; cont3=1; ini(i)=ini_r(i-1)+150000; fin(i)=fin_r(i-1)+220000; for j=ini(i):1:fin(i) %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(j)>I num_muestrasI(cont2)=j; cont2=cont2+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for k=ini_fin:1:fin(i) if x(k)<=F & x(k)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont3)=k; cont3=cont3+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F else 64 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES ini_Tb=ini_analisis(i)+350; fin_Tb=fin_r(i);%-50? %asi nos hemos desplazado a la parte final para el cálculo del Tb y2=envelopeH_filtrado(ini_Tb:fin_Tb);% se tiene que tener en cuenta que el 1ºvalor de y2 realmente corresponde a la muestra ini_Tb,para el posterior cálcuo del tiempo a3=amplitud(i)*0.75; f3=find(y2<=a3); f3(i)=min(f3)+ini_Tb; format long Tb_75(i)=f3(i)*ts duracion_muestras25_75(i)=-f2(i)+f4(i); duracion_tiempo25_75(i)=(duracion_muestras25_75(i)).*ts; y2=0; pause end % ------------------------------- RECEPCIONES 16,17 ------------------------ % %ini_r(15)=ini(15)+2400000;% nos desplazamos 2400000=9.65 segundos después, al último grupo de señales recibidas %fin_r(15)=fin(15)+2400000; ini_r(15)=7300000; fin_r(15)=7251800; for i=16:1:17 % ------------------- Deteccion de cada senyal ---------------- % if i>=2 cont2=1; cont3=1; ini(i)=ini_r(i-1)+150000; fin(i)=fin_r(i-1)+220000; for j=ini(i):1:fin(i) %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(j)>I num_muestrasI(cont2)=j; cont2=cont2+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 65 % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for k=ini_fin:1:fin(i) if x(k)<=F & x(k)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont3)=k; cont3=cont3+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F else cont2=1; cont3=1; for j=1:1:fin1 %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(j)>I num_muestrasI(cont2)=j; cont2=cont2+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for k=ini_fin:1:fin1 if x(k)<=F & x(k)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont3)=k; cont3=cont3+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F end %-------------- Calculo de la amplitud propia de la señal ------------% 66 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES for l=(ini_r(i)-100):1:fin_r(i) %muestras de ini y fin de r1 z(l)=envelopeH(l); end ts=1/fs; rango_estable=envelopeH(ini_r(i)+100:fin_r(i)-100); amplitud(i)=mean(rango_estable); I=amplitud(i)*0.25; % inicio (muestras) señal 1º recepción F=amplitud(i)*0.3; for p=1:1:fin1 %for desde inicio hasta un margen asegurado de la recepción de la 1 señal:300000*Ts=1.2segundos if x(p)>=I num_muestrasI(cont1)=p; cont1=cont+1;%asi conocemos en qué muestra esta el inicio de la 1 recepción % dejamos un margen desde el inicio para asegurar que campte la amplitud correspondiente a la finalizacion de la señal end end ini_r(i)=num_muestrasI(1);%asi coge la primera muestra que ha captado >= que I ini_fin=ini_r(i)+500; for q=ini_fin:1:fin1 if x(q)<=F & x(q)>0 %La condición de x(i)>0 , ya que, sino cogia números negativos de la señal x num_muestrasF(cont)=q; cont=cont+1; end end fin_r(i)=num_muestrasF(1);%asi coge la primera muestra que ha captado <= que F %creamos el nuevo vector para que posteriormente guarde la señal recibida 1ª entre los margenes del eje horizontal correctos ini_analisis(i)=ini_r(1); ini_analisis(i)=ini_r(i)-100; fin_r(i)=fin_r(i)+150; %z=x(ini_r1:1:fin_r1); for l=ini_analisis(i):1:fin_r(i) Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 67 za(l)=x(l); hilbert1(l)=hilbert(l); envelopeH1(l)=envelopeH(l); end %**************************************************************% %****** Filtrado Paso-Banda ***************% %**************************************************************% %******** Transformada de Fourier de la señal ************% t=[ini_analisis(i):1:fin_r(i)].*ts; %rango temporal de la señal f=1./(t.*ts);%rango frecuencial de la señal N=length(za(ini_analisis(i):fin_r(i)));%longitud de la señal frec =[(-1/ts)/2:(1/ts)/(N-1):((1/ts)/2)];%rango frecuencial de la tranformada z1a= fftshift(fft(za(ini_analisis(i):fin_r(i)))); figure,subplot(1,2,2),plot(t,za(ini_analisis(i):fin_r(i))),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en tiempo']) subplot(1,2,1),plot(frec,z1a),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en frecuencia(transformada)']) %********* Detección de la Frecuencia de la señal, Fp y Fs *******% m=max(z1a) p=find(z1a==m) %En el punto donde capta el pico en la transformada se encuentra su frecuencia frec_senyal=frec(p); %por si coge el valor máximo(correspondiente a la frecuencia) en la parte %imaginaria, quitarle el valor negativo: if frec_senyal<0 frec_senyal=-frec_senyal else frec_senyal=frec_senyal end %-------- Aplicamos Filtro Paso Banda de 35 a 65 KHz a una frecuencia de unos 60KHz ---- % fp_min=40000;%fpass minimo fp_max=60000;%fpass maximo 68 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES fs_min=35000;%fstop minimo con 2.5KHz de margen fs_max=65000;%fstop maximo con 2.5KHz de margen Wp = [fp_min fp_max]/fs;%Frecuencias que deseas pasar Ws = [fs_min fs_max]/fs;%Tolerancia que ofreces a la horquilla de frecuencias Rp = 1; Rs = 10;%Tolerancias de la banda de rizado [n,Wn] = buttord(2*Wp,2*Ws,Rp,Rs)%Definicion [b,a] = butter(n,Wn); figure,freqz(b,a,bits,fs)%Parametros y ordenes del filtro + numero de puntos + Frecuencia de muestreo title(['Butterworth Bandpass Filter ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]) zaf1(ini_analisis(i):fin_r(i))= filter(b,a,za(ini_analisis(i):fin_r(i)));%aplicamos filtro a la señal z1af1= fftshift(fft(zaf1(ini_analisis(i):fin_r(i)))); figure,subplot(1,2,1),plot(frec,z1af1),title(['ondafiltrada' num2str(i) 'F.Paso Banda (frecuencia): ' num2str((fp_min)) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(1,2,2),plot(t,zaf1(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) 'F.Paso Banda (tiempo): ' num2str((fp_max)) 'Hz - ' num2str((fp_max)) 'Hz' ]); % ---- Aplicamos filtro Banda Paso entre 40-60KHz a la hilbert senyal -----% % --- para obtener la envolvente de la senyal filtrada ---- % %hilbert1(ini_analisis(i):fin_r(i))= filter(b,1,hilbert1(ini_analisis(i):fin_r(i)));%aplicamos filtro al hilbert de la señal,para que luego la envolvente se corresponda con la de la señal filtrada hilbert1=hilbert1(ini_analisis(i):fin_r(i)); filtro_h(ini_analisis(i):fin_r(i))= filter(b,a,hilbert1); %********* Detección de la Frecuencia de la señal de nuevo , Fp y Fs *******% m=max(z1af1) p=find(z1af1==m) %En el punto donde capta el pico en la transformada se encuentra su frecuencia frec_senyal=frec(p); %por si coge el valor máximo(correspondiente a la frecuencia) en la parte %imaginaria, quitarle el valor negativo: if frec_senyal<0 frec_senyal=-frec_senyal else frec_senyal=frec_senyal end % else % end fp_min=frec_senyal-2500%fpass minimo con 2.5KHz de margen desde la frecuencia propia Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 69 fp_max=frec_senyal+2500%fpass maximo 2.5KHz de margen desde la frecuencia propia fs_min=fp_min-10000%fstop minimo con 10KHz de margen desde la frecuencia propia fs_max=fp_max+10000%fstop maximo con 10KHz de margen desde la frecuencia propia %---------- Rango de Banda de paso de margen 5KHz con frec central=frecuencia de la señal ------% Wp = [fp_min fp_max]/fs;%Frecuencias que deseas pasar Ws = [fs_min fs_max]/fs;%Tolerancia que ofreces a la horquilla de frecuencias Rp = 1; Rs = 10;%Tolerancias de la banda de rizado [n,Wn] = buttord(2*Wp,2*Ws,Rp,Rs)%Definicion [b,a] = butter(n,Wn); figure,freqz(b,a,bits,fs)%Parametros y ordenes del filtro + numero de puntos + Frecuencia de muestreo title(['Butterworth Bandpass Filter ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]) y(ini_analisis(i):fin_r(i))= filter(b,a,zaf1(ini_analisis(i):fin_r(i))); ondafiltrada = fftshift(fft(y(ini_analisis(i):fin_r(i)))); figure subplot(1,2,1),plot(frec,ondafiltrada),title(['ondafiltrada' num2str(i) '(frecuencia): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(1,2,2),plot(t,y(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) '(tiempo): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); figure,subplot(2,2,2),plot(t,za(ini_analisis(i):fin_r(i))),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en tiempo']) subplot(2,2,1),plot(frec,z1a),title(['Señal' num2str(i) '-Sensor' num2str(s) '- : Onda en frecuencia(transformada)']) subplot(2,2,3),plot(frec,ondafiltrada),title(['ondafiltrada' num2str(i) '(frecuencia): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); subplot(2,2,4),plot(t,y(ini_analisis(i):fin_r(i))),title(['ondafiltrada' num2str(i) '(tiempo): ' num2str(fp_min) 'Hz - ' num2str(fp_max) 'Hz' ]); figure set(gca,'fonts',8) plot(frec,z1a,frec,z1af1,frec,ondafiltrada),title(['SENSOR' num2str(j) 'OndaFiltrada' num2str(i) '(frecuencia) Paso Banda a Frec.Central: ' num2str(frec_senyal) 'Hz - ']) legend('S. Original' ,'S. F.P.B 40-60KHz','S. F.P.B Margen 5KHz') figure set(gca,'fonts',8) plot(t,za(ini_analisis(i):fin_r(i)),t,zaf1(ini_analisis(i):fin_r(i)),t,y(ini_ analisis(i):fin_r(i))),title(['SENSOR' num2str(s) 'OndaFiltrada' num2str(i) '(tiempo) Paso Banda a Frec.Central: ' num2str(frec_senyal) 'Hz - ']) legend('S. Original' ,'S. F.P.B 40-60KHz','S. F.P.B Margen 5KHz') 70 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES pause % --------------- Método 'HILBERT TRANSFORM' ----------% % ---- Aplicamos filtro Banda Paso 5KHz a la hilbert senyal filtrada -----% % --- entre 40-60KHz para obtener la envolvente de la senyal filtrada ---- % filtro_h= filter(b,a,filtro_h(ini_analisis(i):fin_r(i))); envelopeH_filtrado(ini_analisis(i):fin_r(i))=abs(filtro_h); fase=unwrap(angle(filtro_h)); %fase fx=diff(unwrap(angle(filtro_h)))./diff(t)/(2*pi); %frecuencia %------------------ amplitud ------------ % figure,subplot(3,1,1),plot(t,y(ini_analisis(i):fin_r(i)),t,envelopeH_filtrado (ini_analisis(i):fin_r(i)),'r:'),title(['SENSOR' num2str(s) ':Muestras y Envolvente(Mediante HILBERT)de la Señal recibida ' num2str(i) ' en el tiempo parte estable']); xlabel('t(segundos)'); ylabel('x(t) normalizado') grid on %------------------ fase ------------ % subplot(3,1,2),plot(t,fase),title(['SENSOR' num2str(s) ':Fase de la Señal recibida ' num2str(i) ' en el tiempo']); xlabel('t(segundos)'); ylabel('fase (t)') grid on %s=[z1 z]; %s=[z za];%Aqui quedan guardadas todas las señales su número de muestras correspondiente %------------------ frecuencia ------------ % subplot(3,1,3),plot(t(2:end),fx),title(['SENSOR' num2str(s) ':Frecuencia de la Señal recibida ' num2str(i) ' en el tiempo']); xlabel('t(segundos)'); ylabel('frecuencia (t)') grid on % -------------------- Cálculo del tiempo de subida --------------------% % Calculamos de nuevo la amplitud, en la señal Filtrada, en la parte central % de la senyal, ya que, es normalmente, donde tiene un nivel de amplitud % mas estable rango_estable=envelopeH_filtrado(ini_analisis(i)+300:fin_r(i)-250); amplitud(i)=mean(rango_estable) Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 71 % ---------------------- y1(ini_analisis(i):fin_r(i))=envelopeH_filtrado(ini_analisis(i):fin_r(i)); a2=amplitud(i)*0.25; i2=find(y1>=a2); i2(i)=min(i2); format long Ts_25(i)=i2(i)*ts y1=0; %--------------------------- % -------------------- Cálculo del tiempo de bajada --------------------% ini_Tb=ini_analisis(i)+350; fin_Tb=fin_r(i);%-50? %asi nos hemos desplazado a la parte final para el cálculo del Tb y2=envelopeH_filtrado(ini_Tb:fin_Tb);% se tiene que tener en cuenta que el 1ºvalor de y2 realmente corresponde a la muestra ini_Tb,para el posterior cálcuo del tiempo a2=amplitud(i)*0.75; f2=find(y2<=a2); f2(i)=min(f2)+ini_Tb;%sumamos ini_Tb ya que la 1 casilla de y2(valor del tiempo) corresponde a la casilla ini_Tb format long Tb_50(i)=f2(i)*ts%el número de casilla(valor tiempo normalizado)*ts= valor del tiempo(desnormalizado) que corresponde al alcance del % de amplitud durante el tiempo de bajada duracion_muestras25_75(i)=-f2(i)+f4(i); duracion_tiempo25_75(i)=(duracion_muestras25_75(i)).*ts; y2=0; pause end Appendix 2. Acoustic Array Calibration and Signal Processing for UHE Neutrinos Generation. Acoustic Array Calibration and Signal Processing for UHE Neutrinos Generation 72 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES W.Ooppakaew, S.Danaher School of CEIS Northumbria University Newcastle, UK wichian.oop[email protected] [email protected] M.Saldaña IGIC Universidad Politécnica de Valencia Valencia, Spain [email protected] Abstract— The calibration and design of an eighthydrophone linear array has been developed for the generation of emulated Ultra High Energy (UHE) neutrino-induced pulses. The acoustic array simulates the acoustic pulse created from a neutrino interaction in water by generating a coherently emitted acoustic bipolar pulse. The calibrator is developed using signal processing methods using experimental measurements to characterise the hydrophone system. An 8 channel PIC hydrophone array module has been built for processing and control. A linear array simulation for the neutrino pulses production has been studied in order to predict the directivity and shape of the acoustic bipolar pulse at the ANTARES neutrino detector. Keywords: Array hydrophone calibration; UHE neutrinos; Acoustic detection; PIC module; array simulation. INTRODUCTION In order to evaluate the actual behaviour of acoustic neutrino detector in the field, acoustic transmitters are needed to send artificial acoustic signatures from UHE neutrino interactions in the sea. The transmitters have to be able to imitate the UHE neutrino acoustic pulse, in directivity (pancake) and shape (a short bipolar pulse). Furthermore the detector site is located at a depth of 2475 m underwater. Due these limitations, it is not an easy task to design a system able to reproduce an acoustic bipolar pulse with such characteristics. The deployment and operation in site have to be considered in terms on the system design as well. The use of a linear array of omnidirectional transmitters has been studied and developed to achieve the proposed objective. AIM This research utilizes an eight-hydrophone calibrated array for the generation of Ultra High Energy neutrino-induced pulses at the frequency of 23 kHz. Signal processing techniques are applied to array hydrophone modeling and experimental data generated in a laboratory water tank are used and compared to the calibration system. Using such an array has the advantage that would allow precise calibration of existing and planned acoustic neutrino telescope. An acoustic generation module of 8 channels has been built using PIC microcontrollers for process and control [1]; it will be used to improve the speed and precision of data transmission in underwater communication using the precise control of the amplitude and phase of the signal. DEEP-SEA ACOUSTIC NEUTRINO DETECTION The ANTARES module for Acoustic Detection under the sea (AMADEUS) The ANTARES (Astronomy with a Neutrino Telescope and Abyss environmental RESearch) Collaboration [2] built a large effective area water Cherenkov detector in the deep Mediterranean Sea optimised for the detection of muons from highenergy astrophysical neutrinos. A large amount of Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 73 sensors of different kinds had to be deployed and connected in deep sea to become a network of coordinated underwater sensors looking for neutrinos and monitoring the deep sea. The AMADEUS project [3] was conceived to perform a feasibility study for a potential future large scale acoustic neutrino detector. For this purpose, a dedicated array of acoustic sensors was integrated into the ANTARES neutrino telescope. The Thermo-Acoustic model Measuring acoustic pressure pulses in huge underwater acoustic arrays is a promising approach for the detection of cosmic neutrino with energies exceeding 100 PeV. The particle cascades that evolve when neutrinos interact with nuclei in water produce pressure signals. The resulting energy deposition in a cylindrical volume of a few centimetres in radius and several metres in length (from 5m to 10 m) lead a local heating of the medium, which is near instantaneous with respect to the acoustic process. This temperature change induces an expansion or contraction of the medium depending on its volume expansion coefficient. According to the thermo acoustic model, the accelerated motion of the heated volume—a micro-explosion—forms a pressure pulse of bipolar shape, which propagates in the surrounding medium. The wave equation describing the pulse is: 2 2 2 2 2 2),(),(1 ),( t tr C t trp c trp p s    (1) Hear ),( trp  denotes the pressure at a given place and time, s c the speed of sound in the medium, p C its specific heat capacity, the thermal expansion coefficient and ),( tr  the energy deposition density of the particles. The spatial and temporal distribution of is not accessible to laboratory experiments at the relevant primary neutrino energies [4,5] and hence is subject to uncertainties. Simulations depend on both the extrapolation of parameterisations for lower energies into this regime and on the transcription of simulations for extended air showers to showers in water. Monte Carlo simulation of neutrino interactions in water for a cascade energy has been developed in previous studies [6,7], where is demostrated if the energy deposition is modelled using Monte Carlo points with a density proportional to energy, the pressure will be a scaled derivative histogram of the flight times to the observer. Attenuation in seawater The acoustic attenuation in seawater is almost totally caused by absorption. In the 1-100 kHz region it is dominated by a chemical relaxation process that is connected with the associationdissociation of magnesium sulphate ( 4 MgSO ) ions under the pressure of the sound wave. Below 1 kHz a similar mechanism involving boric acid ( ))(( 3 OHB is responsible for much of the observed attenuation. Taken together these mechanisms result in an attenuation of acoustic waves in seawater and a velocity of sound, which are both frequency dependent. Experimentally however whereas it is straightforward to estimate the magnitude of |a| of the complex attenuation it is very difficult to determine the phase angle of a and no current measurements exist in the literature. The magnitude of the attenuation however is well 80 Study of acoustic signals for the neutrino detector AMADEUS-ANTARES [4] K.Greisen,Phys.Rev.Lett 16(1966)748. [5] G:T Zatsepin, V.A kuz´min,JETP Lett,4(1906)78;G.T.Zatsepin, V.A.Kuz´min, Z.Eksp.Teor.Fiz(1966)114. [6] S.Bevan, “Study of the acoustic siganture of UHE neutrino interactions in water and ice” Nucl.Instr. and Meth.A 607 (2009) 398-411. [7] S. Bevan “Simulation of Ultra High Energy Neutrino Interactions in Ice and Water” et al. / Astroparticle Physics 28 (2007) 366–379 [8] R. E. Francois, G. R. Garrison, “Sound absorption based on ocean measurements: Part I:Pure water and magnesium sulfate contributions,” “Sound absorption based on ocean measurements: Part II:Boric acid contribution and equation for total absorption” [9] M. A. Ainslie and J. G. McColm “A simplified formula for viscous and chemical absorption in seawater” Journal of the Acoustical Society of America 103(3) (1998) 1671-1672. [10] L. N. Lieberman, “The Origin of Sound Absorption in Water and in Sea Water” The Journal of the Acoustical Society of America 20(6) (1948) 868-873. [11] V. Niess and V. Bertin, “Underwater Acoustic Detection of Ultra High Energy Neutrinos”, Astropart. Phys. 26(4-5) (2006) 243-25 Study of acoustics signal for the neutrino detector AMADEUS-ANTARES 81