International Research Journal of Earth Sciences______________________________________ ISSN 2321–2527Vol. 2(9), 7-14, October (2014) Int. Res.J. Earth Sci. International Science Congress Association 7 Prediction of Rock Mechanical Parameters as a Function of P-Wave Velocity Abbas Abbaszadeh Shahri1, 3*, Arsham Gheirati2 andMaria EsperssonFaculty of Civil Engineering, Islamic Azad University, Rudehen Branch, Rudehen, Tehran, IRAN Department of Geophysics, Islamic Azad University, Chalous branch, Chalous, IRAN 1,3Uppsala University, Box 534, Construction Engineering Department, SE-75121 Uppsala, SWEDENAvailable online at: www.isca.in, www.isca.me Received 21th June 2014, revised 24th September 2014, accepted 13th October 2014 AbstractDue to non destructive and easy method of P-wave velocity (V) measurements in field and laboratory conditions, and also its relation to mechanical parameters of the material, it has increasingly been conducted to determine the physical properties of rock materials. In this paper an experimental study of the measurement of P-wave velocity, uniaxial compressive strength (UCS), Schmidt hammer test (N), porosity (n), saturated and dry density () and elasticity modulus (E) for two types of rocksincluding the sandstone and schist at a selected site in West of Iran in Hamedan province were performed. According to available reports of Geological Survey of Iran (GSI), in most of areas of Zagros fault zone including our selected area, particularly in Sanandaj – Sirjan zone, the majority of rock types are consisting of sandstones and metamorphed Jurassic rocks in green schist. Therefore, in the present paper, we aim to determine reliable empirical predictive models as a function of P-wave velocity to estimate the rocks properties for these two rock types. For this purpose, 30 samples consisting of 9 schist and 21 sandstones were tested in laboratory. By application of statistical analysis and student t-test the computed regression coefficient value evaluated and showed that these obtained empirical relations can be applied for the West of Iran. To verify and validate our results a detailed comparison between results of this study by other researchers were conducted by plotting graphs. The result of comparison shows good compatibility with each other. Keywords: Rock mechanical parameters; P-wave velocity, predictive models, Zagors fault zone. Introduction In different condition and purposes, there are several developed approach and different modelling techniques for design of engineering structures using intact and rock mass index properties. In this case, application of ultrasonic techniques because of simplicity and non destructivity in field and laboratories1,2 are increasingly being used in various fields of mining, geotechnical, civil and underground engineering. Rock type, density, grain size and shape, porosity, anisotropy, porewater pressure, clay content, confining pressure and temperature are effective factors on p-wave velocity but weathering, alteration, bedding planes and joint properties consisting of roughness, filling material, water, dip and strike have an important influence on the seismic velocity. For site characterization, mining and civil engineering applications, having a strong laboratory database of rock mechanical and engineering properties will be very useful. In practice, such a database is not available and moreover, discontinuity and variable nature of rock masses increases difficulties of directly obtain the specific design parameters of interest. Application of these proposed correlations is of interest, mainly due to the fact that rock index tests have the advantages of being relatively fast and economical. However, obtained correlations are not constant and can be varied with rock types. Hence, several researchers have been established and propose empirical equations between the V, petrophysical and mechanical parameters of the rocks4-29. In this paper, we aim to establish predictive models for rock mechanical parameters as a function of V from collected and tests rock samples from the various boreholes of the Khorram rud earth dam site in Hamedan province in west of Iran. V, UCS, Schmidt hammer rebound test (N), density (dry,saturated), porosity and elastic modulus were the determined properties of samples in this study. Validation of obtained models in this study was investigated by conducting detailed comparison with available literature reviews for application of rock engineering. The samples were picked from various depths in each borehole and all samples were calibrated to be in standards format of ISRM30. Material and Methods By considering the figure-1, Zagros Mountains follow a NW to SE pattern. A common way to divide this large area is considering two parts including Northern and Southern Zagros. Northern Zagros includes Iranian provinces of West Azerbayejan, Kurdistan, Hamedan, Kermanshahan, Ilam, Lorestan, Khuzestan and Chaharmahal va Bakhtiari. Southern Zagros covers provinces of Kohgiluye va Buyerahmad, Fars, Bushehr and Hormozgan. Geologically, Zagros Mountains consist of high and folded Zagros as two major parts. High Zagros forms the north eastern mountains and folded Zagros stands in the south and west of the high Zagros. Heading east, high Zagros faces inner highlands of the Zagros Mountains also known as Sanandej - Sirjan Zone. Folded Zagros instead ends in the Persian Gulf in south Khuzestan plain in south west and Mesopotamia in west. International Research Journal of Earth Sciences____________________________________________________ ISSN 2321–2527 Vol. 2(9), 7-14, October (2014) Int. Res. J. Earth Sci. International Science Congress Association 8 Figure1 Zagros fault characteristics and location of the studied area Hamedan Province which located in the west of Iran is mostly mountainous in part of the Zagros range, especially high folded Zagros31. The site of Khorramrud embankment dam is located in 48º 18´ E and 34º 38´ N geographical coordinates, in 30 Km of southwest of Hamedan and near the Oshtoran village in west of Iran as shown in figure1. This clay core earth dam has a crest length of 755m, 55m height from the river base and 11.53Mm reservoir capacity. A total of 13 boreholes up to depth of 80m with maximum depth of ground water table of 18.7m in the various location of site were drilled but the information of 10 boreholes because of better quality core samples were selected for analysis as shown in table-1. According to executed exploratory drilling, field investigations, surveying and acquisition of discontinuities information and prepared petrological thin sections, two types of lithology including sandstone and schist were recognized and structure properties of rock mass have been performed and analyzed. The sandstones of this area can separate into red sandstone (slightly weathered) and gray sandstone. Table-1 Drilled exploratory boreholes in the studied area Borehole Depth and GWT (m) Sample properties Depth Sample No. Rock type KB2: dam axis 55 (18.7) 24.28 - 25.00 51912 schist 52.50 - 53.00 51914 schist KB3:dam axis 35 (10.5) 28.00-28.50 51921 schist BH2: dam axis 65 (9.70) 41.20 - 41.70 51924 schist BH2(A): dam axis 60 (8.20) 54.00 - 54.30 42992 schist 47.53 - 47.93 42991 sandstone 32.50 - 32.80 42989 sandstone KB1: right support 30 (---) 18.20 - 18.75 42944 sandstone 24.68 - 25.00 42943 sandstone BH1: right support 75 ( --- ) 43.10 - 43.70 42950 sandstone BH3: dam axis 75 (8.30) 42.20 - 42.75 42933 sandstone 48.68 - 49.00 42934 sandstone BH3(A): dam axis 80 (---) 24.72 - 25.00 42946 sandstone 23.00 - 23.33 42945 sandstone BH4: left support 30 (---) 11.20 - 11.50 42949 sandstone 14.40 - 14.70 42950 sandstone BH5: left support 65 (---) 15.30 - 15 - 60 42938 schist 30.40 - 30.80 42942 schist 32.40 - 32.95 42940 schist 42.60 - 43.00 42941 sandstone 60.57 - 60.95 42939 sandstone International Research Journal of Earth Sciences____________________________________________________ ISSN 2321–2527 Vol. 2(9), 7-14, October (2014) Int. Res. J. Earth Sci. International Science Congress Association 9 Table-2 Obtained results of the tested samples in the studied area Rock type (number of samples) UCS (MPa) N E (GPa) Vp (Km/s) n (%) sat (gr/cm) (gr/cm) max min max min max min max min max min max min max min sandstone (21) 95 20 22.56 10.18 3.75 3.10 95 20 22.56 10.18 3.75 3.10 95 20 schist (9) 45.2 37.5 9.6 9.13 3.43 3.35 45.2 37.5 9.6 9.13 3.43 3.35 45.2 37.5 Table-3 Statistical analyses of the measured parameters Rock Type UCS (Mpa) E(Gpa) Vp(Km/s) N n(%) sat (gr/cm) d (gr/cm) Sandstone 50.65±25.41 13.62±5.8 3.44±0.19 13.71±2.8 1.27±0.11 2.66±0.03 1.82±0.05 Schist 42.33±3.2 9.34±0.15 3.39±0.02 11.6±1.4 1.54±0.4 2.68±0.01 1.81±0.02 From this area, 30 picked samples were inspected to ensure that it would provide standard testing specimens without macroscopic defects, alteration zones and fractures according ISRM suggested methods and then tested for UCS, V, porosity, elasticity modulus, Schmidt hammer test and density. Obtained results of the measured samples and statistical analyses of them are given in tables-2 and 3 respectively. By considering the distribution of tested data, there is a probable relation between the UCS and other mechanical parameters in this area. Therefore as presented in figures-2 and 3, we plotted the contour line of the tested parameters for the selected area which shows same stretch in contour trends. This can improve the existence of probable relations between the mechanical parameters. Linear regression analysis is a statistical process for estimating the relationships among variables and aims to describe the output variable y through a linear combination of one or more input variables. Due to several advantages such as simplicity, providing adequate and interpretable description of how the inputs affect the outputs, linear models were largely developed and are still good reasons to study and use them since they are the foundation of more advanced methods. Moreover, for prediction purposes they can often outperform fancier nonlinear models, especially in situations with small numbers of training data or a low signal-to-noise ratio. Finally, linear models can be applied to transformations of the inputs and therefore be used to model nonlinear relations. In this study, as presented in table-3, at the first of all, the results of the executed tests were determined and after statistical analysis, the range, the mean and the standard deviation values for each measured properties were calculated. By application of the least square regression analysis method and using Matlab curve fitting tool and curve expert mathematical software the linear and power regression among the UCS, density, porosity and elastic modulus of samples as a function of V and mainly to derive reliable, empirical approaches for the determination of UCS were expressed. To determine whether there is any relation between UCS and other parameters or not the Student t-test is performed and a relation is observed. International Research Journal of Earth Sciences____________________________________________________ ISSN 2321–2527 Vol. 2(9), 7-14, October (2014) Int. Res. J. Earth Sci. International Science Congress Association 10 Figure-2 Variation of V and UCS in the selected area on base of the tested sandstone samples Figure-3 Variation of V and UCS in the selected area on base of the tested schist samples The t-test compares the computed values with tabulated values using null hypothesis32. According to the t-test, when computed t-value is greater than tabulated t-value, the null hypothesis is rejected and obtained correlation coefficient (r) is acceptable. Also, observed level of significance is often used in hypothesis test33. In this case, as p-value is smaller than level ofsignificance ( = 0.05), the null hypothesis is rejected. Therefore, it means that there is a relation between the correlated parameters and this shows that r-value is significant. The equation of the best fit line, the 95% confidence limits, and the correlation coefficient (r) which indicates of how well the model ts the data, were determined for each regression as shown in figures 4 to 7. The high value obtained r-value in the introduced relationships in these figures can be a reasonable evidence for the correctness of these empirical relationships. ERROR: stackunderflow OFFENDING COMMAND: ~ STACK: