Evaluation of Rock Acoustic Properties by a Parametric Spectral Analysis A. Lachouri 1 N. Doghmane 2* , T. Bouden 3 , B. Harhoud 2 and H. Boumaaraf 1 1 Laboratoire d’Automatique Appliquée LAASK, Université de Skikda. 2* Dpt. d’Electronique, Faculté des Sciences de l’ingénieur, Université de Annaba. 3 Dpt d’électronique Université de Jijel ALGERIA Abstract: In this paper, a new approach on non-destructive evaluation by ultrasonic signal is proposed. This work is based on a parametric spectral analysis. Indeed, the ultrasonic velocity measurement in rocks is often obtained by transmitting ultrasonic pulses through the specimen. The longitudinal waves (P) are chosen in the experiment analysis (the case of transversal waves (S) will be tested in other studies). This measurement is simple and can allow an easy identification of the acoustics parameters of the rock. The rock acoustic characteristics give several indications on its state of alteration, the grains size and micro crakes. Keywords: Ultrasonic, Attenuation, Porosity, Micro cracks, parametric estimation, PSD, Prony, Resolution. 1. Introduction Ultrasonic rocks characterization techniques can be used for the recognition of the internal constitution of specimen (in this case: The rock). This is possible by using existing pulse method and appropriate digital signal processing. Different parameters extracted from the processed ultrasound have been considered in the literature, with variable degrees of success. The present work deals with rock characterizations using ultrasonic nondestructive evaluation method. The ultrasonic velocity measurement in rocks is often obtained by transmitting ultrasonic pulses through the materials (rock sample) [1, 2]. This measurement is easy, fast and offers the possibility for the identification and the estimation of rock acoustic parameters. These parameters can be used to give significant indications on the state of alteration, grain size and micro cracks of the sample (rock) [3] etc… 2. Problem position Digital signal processing involves techniques that improve our understanding of information contained in received ultrasonic data. Normally, this signal is measured and viewed in the time domain. When the frequency content of the signal is of interest, it makes sense to study the signal in the frequency domain (spectral analysis). The spectral analysis of the ultrasonic signals can be used as a processing technique to permit, then, a non- destructive diagnosis of the rocks. The goal is to evaluate several acoustic characteristics of the rocks. Generally, the diagnosis is done by the calculation of the attenuation of the ultrasonic signal transmitted through the rock (attenuation and/or quality factor) or by the measurement of the delay of this same signal (for wave velocity measurement inside the solid). This method, which is known as quantitative analysis, does not provide enough information and some times its results are not reliable. The purpose of this study is to improve the characterization using a spectral estimation with linear prediction for a sharp and accurate analysis. In others words, the defaults, lacunae and pores, which may exist in the samples, are identified and isolated. The method is based on the emission reception of an acoustic longitudinal wave passing through rocks. This paper contains a description of quantitative analysis method of rock acoustic properties and a description of qualitative analysis method based on autoregressive (AR) spectral estimation. 3. Description of the experiment A bloc diagram of the experiment set-up is given in fig.1. In these tests, the pulse transmission method is used. The experiment set-up allows the study of attenuation coefficient and longitudinal wave velocity obtained from the pulse transmitted through cylindrical samples of rocks. The pulses generated from an ultrasonic generator and transducers (transmitter and receiver) of 150KHz, 300KHz and 500KHz central frequencies. Acoustic contact, between transducers and samples, is made with couplant (grease) and a weight of about 0.2Mpa is also applied to improve wave propagation. The waveform of the signal is visualized by a digital scope and transferred to the computer by the liaison RS232C as digital data file for further analysis. Proceedings of the 5th WSEAS Int. Conf. on SIGNAL, SPEECH and IMAGE PROCESSING, Corfu, Greece, August 17-19, 2005 (pp123-127)