Prediction Of Shear Wave Velocity Research Articles

Shear wave velocity prediction based on 1DCNN-BiLSTM network with attention mechanism

The Shear wave (S-wave) velocity is an essential parameter in reservoir characterization and evaluation, fluid identification, and prestack inversion. However, the cost of obtaining S-wave velocities directly from dipole acoustic logging is relatively high. At the same time, conventional data-driven S-wave velocity prediction methods exhibit several limitations, such as poor accuracy and generalization of empirical formulas, inadequate exploration of logging curve patterns of traditional fully connected neural networks, and gradient explosion and gradient vanishing problems of recurrent neural networks (RNNs). In this study, we present a reliable and low-cost deep learning (DL) approach for S-wave velocity prediction from real logging data to facilitate the solution of these problems. We designed a new network sensitive to depth sequence logging data using conventional neural networks. The new network is composed of one-dimensional (1D) convolutional, bidirectional long short-term memory (BiLSTM), attention, and fully connected layers. First, the network extracts the local features of the logging curves using a 1D convolutional layer, and then extracts the long-term sequence features of the logging curves using the BiLSTM layer, while adding an attention layer behind the BiLSTM network to further highlight the features that are more significant for S-wave velocity prediction and minimize the influence of other features to improve the accuracy of S-wave velocity prediction. Afterward, the nonlinear mapping relationship between logging data and S-wave velocity is established using several fully connected layers. We applied the new network to real field data and compared its performance with three traditional methods, including a long short-term memory (LSTM) network, a back-propagation neural network (BPNN), and an empirical formula. The performance of the four methods was quantified in terms of their coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). The new network exhibited better performance and generalization ability, with R2 greater than 0.95 (0.9546, 0.9752, and 0.9680, respectively), RMSE less than 57 m/s (56.29, 23.18, and 30.17 m/s, respectively), and MAE less than 35 m/s (34.68, 16.49, and 21.47 m/s, respectively) for the three wells. The test results demonstrate the efficacy of the proposed approach, which has the potential to be widely applied in real areas where S-wave velocity logging data are not available. Furthermore, the findings of this study can help for a better understanding of the superiority of deep learning schemes and attention mechanisms for logging parameter prediction.

Shear Wave Velocity Prediction Based on the Long Short-Term Memory Network with Attention Mechanism

Shear wave velocity (VS) is a vital prerequisite for rock geophysics. However, due to historical, cost, and technical reasons, the shear wave velocity of some wells is missing. To reduce the deviation of the description of underground oil and gas distribution, it is urgent to develop a high-precision neural network prediction method. In this paper, an attention module is designed to automatically calculate the weight of each part of the input value. Then, the weighted data are fed into the long short-term memory network to predict shear wave velocities. Numerical simulations demonstrate the efficacy of the proposed method, which achieves a significantly lower MAE of 38.89 compared to the LSTM network’s 45.35 in Well B. In addition, the relationship between network input length and prediction accuracy is further analyzed.

Comparing Artificial Intelligence Algorithms with Empirical Correlations in Shear Wave Velocity Prediction

Accurate estimation of shear wave velocity (Vs) is crucial for modeling hydrocarbon reservoirs. The Vs values can be directly measured using the Dipole Shear Sonic Imager data; however, it is very expensive and requires specific technical considerations. To address this issue, researchers have developed different methods for Vs prediction in underground rocks and soils. In this study, the well logging data of a wellbore in the Iranian Aboozar limestone oilfield were used for Vs estimation. The Vs values were estimated using five available empirical correlations, linear regression technique, and two machine learning algorithms including multivariate linear regression and gene expression programming. Those values were compared with the real Vs data. Furthermore, three statistical indices including correlation coefficient (R2), root mean square error (RMSE), and mean absolute error (MAE) were used to evaluate the effectiveness of the applied techniques. The mathematical correlation obtained by the GEP algorithm delivered the most accurate Vs values with R2 = 0.972, RMSE = 0.000290, and MAE = 0.000208. Compared to the available empirical correlations, the obtained correlation from the GEP approach uses multiple parameters to estimate the Vs, thereby leading to more precise predictions. The new correlation can be used to estimate the Vs values in the Aboozar oilfield and other geologically similar reservoirs.

Prediction of geomechanical bearing capacity using autoregressive deep neural network in carbon capture and storage systems

Carbon Capture and Storage (CCS) field is growing rapidly as a means to mitigate the accumulation of greenhouse gas emissions. However, the geomechanical stability of CCS systems, particularly related to bearing capacity, remains a critical challenge that requires accurate prediction models. In this research paper, we investigate the efficacy of employing an Autoregressive Deep Neural Network (ARDNN) algorithm to predict the geomechanical bearing capacity in CCS systems through shear wave velocity prediction as an index for bearing capacity evaluation of deep rock formations. The model utilizes a dataset consisting of 23,000 data points to train and test the ARDNN algorithm. Its scalability, use of deep learning techniques, automatic feature extraction, adaptability to changes in data, and versatility in various prediction tasks make it an attractive option for accurate predictions. The results demonstrate exceptional performance, as evidenced by an R-squared value of 0.9906 and a mean squared error of 0.0438 for the test data compared to the measured data. This research has significant practical implications for effectively predicting geomechanical stability in CCS systems, thus mitigating potential risks associated with their operation.

An improved integration strategy for prediction of shear wave velocity using petrophysical logs: Integration of spatiotemporal and small sample nonlinear feature

Shear wave velocity (Vs) plays an important role in seismic data inversion and accurately determining petrophysical parameters. However, owing to the high acquisition cost of S-wave logging, Vs-related data cannot be obtained from all wells, especially the older ones. Therefore, developing an efficient and fast intelligent prediction method for Vs using existing petrophysical logging data is necessary. Owing to the advantages of conventional machine learning (CML) and deep learning (DL) methods, this study developed a data-driven integrated learning (IL) method for Vs prediction, which uses sensitive logging data as input, builds multiple sample sets, analyzes the performance differences of the CML and DL methods for predicting Vs in each dataset, and optimizes the support vector machine (SVM) in the CML method and convolutional neural network-long short-term memory (CNN-LSTM) in the DL method to construct an integrated CNN-LSTM-SVM neural network (ICLSNN). The spatiotemporal feature extraction capability of CNN-LSTM was used to effectively extract the deposition information of the formation from the logging data, and the small sample nonlinear feature extraction capability of SVM was used to alleviate the limitations of overfitting in DL for data-driven intelligent Vs prediction. The ICLSNN showed a higher Vs prediction accuracy than SVM or CNN-LSTM alone. The Vs prediction results of the single well in the discovery without training were better than those of the other IL and empirical models, proving the strong generalization applicability of the proposed model. This method compensates for the limitations of individual machine learning or DL models for Vs prediction and provides a novel research idea. The effective implementation of this scheme will facilitate the efficient prediction of Vs in areas with few wells.

Quasi-site-specific prediction of shear wave velocity from CPTu

A rapid and reasonable estimation of the local site shear wave velocity (Vs) profile is indispensable for site assessment. For soil deposits, researchers have developed numerous empirical correlations between Vs and piezocone penetration test (CPTu) based on a global/regional (generic) database. However, these transformations are usually biased for local sites because the generic database contains all available data without regard to site-specific differences, and the transformation uncertainty is usually non-negligible due to the complicated geology factors and measurement errors. Moreover, it is challenging to construct a site-specific correlation due to the sparse in-situ measurements. To this end, this study develops a quasi-site-specific CPTu-Vs transformation model that algorithmically combines both the generic database and site-specific measurements to provide reliable Vs prediction with uncertainty description. A generic database containing 3116 cases of 6 parameters is collected from different countries, including corrected cone tip resistance, sleeve frictional resistance, depth, soil behavior index, pore pressure ratio, and Vs. As the significant impact of soil type on CPTu-Vs correlation, independent Johnson transformation (JS) for different soil types is first adopted to normalize the generic database as a whole, ensuring the constructed model is applicable to any soil type site. Three state-of-art quasi-site-specific models are investigated for CPTu-Vs transformation: probabilistic multiple regression (PMR), the hybridization method (HYD), and the hierarchical Bayesian model (HBM). Empirical correlations and the recently-developed data-driven model (XGBoost) are also considered to illustrate the novelty of the quasi-site-specific model. The applications on three sites with different soil types show that considering soil type in JS improves model accuracy and reduces the transformation uncertainty. The HBM with multiple basic CPTu parameters performs best regarding accuracy, uncertainty, and robustness.

Shear wave velocity prediction using bidirectional recurrent gated graph convolutional network with total information embeddings

Shear wave velocity is an essential elastic rock parameter for reservoir characterization, fluid identification, and rock physics model building. However, S-wave velocity logging data are often missing due to economic reason. Machine learning approaches have been successfully adopted to overcome this limitation. However, they have shortcomings in extracting meaningful spatial and temporal relationships. We propose a supervised data-driven method to predict S-wave velocity using a graph convolutional network with a bidirectional gated recurrent unit (GCN-BiGRU). This method adopts the total information coefficient to capture non-linear dependencies among well-log data and uses graph embeddings to extract spatial dependencies. Additionally, the method employs a bidirectional gated mechanism to map depth relationships in both upward and backward directions. Furthermore, the prediction performance is increased by an unsupervised graph neural network to handle outliers and the generation of additional features by the complete ensemble empirical mode decomposition with additive noise method. Finally, the GCN-BiGRU network is compared with Castagna’s empirical velocity formula, support vector regression, long-short-term memory (LSTM), GRU, and BiGRU methods over the North Sea open dataset. The results show that the proposed method performs better predicting S-wave velocity than the other ML and empirical methods.

Experimental study on shear wave velocity of sand-gravel mixtures considering the effect of gravel content

Sand-gravel mixtures are special engineering geological materials between soils and fractured rocks. This study performs a series of bending element tests to systematically investigate the shear wave velocity (Vs) of the sand-gravel mixtures, establish an effective evaluation method, and assess the influence of relative density and effective confining pressure on mixtures with a wide range of gravel contents. The results showed that the shear wave velocity increases and then decreases with the increase in gravel content and increases with the rise in relative density and effective confining pressure. Furthermore, a shear wave velocity prediction model is proposed in this study based on the intergranular contact state theory, including the stress parameter (n) and skeleton void ratio. The stress parameter can be described by a power function considering the uniformity coefficient. The model serves as a reference guide for estimating the shear wave velocity of sand-gravel mixtures with a wide range of gravel contents.

Shear wave velocity prediction based on deep neural network and theoretical rock physics modeling

Shear wave velocity plays an important role in both reservoir prediction and pre-stack inversion. However, the current deep learning-based shear wave velocity prediction methods have certain limitations, including lack of training dataset, poor model generalization, and poor physical interpretability. In this study, the theoretical rock physics models are introduced into the construction of the labeled dataset for deep learning algorithms, and a forward simulation of the theoretical rock physics models is utilized to supplement the dataset that incorporates geological and geophysical knowledge. This markedly increases the physical interpretability of the deep learning algorithm. Theoretical rock physics models for two different types of reservoirs, i.e., conventional sandstone and tight sandstone reservoirs, are first established. Then, a full-sample labeled dataset is constructed using these two types of theoretical rock physics models to traverse the elasticity parameter space of the two types of reservoirs through random variation and combination of parameters in the theoretical models. Finally, based on the constructed full-sample labeled dataset, four parameters (P-wave velocity, clay content, porosity, and density) that are highly correlated with the shear wave velocity are selected and combined with a deep neural network to build a deep shear wave velocity prediction network with good generalization and robustness, which can be directly applied to field data. The errors between the predicted shear wave velocity using the deep neural network and the measured shear wave velocity data in the laboratory and the logging data in three real field work areas are less than 5%, which are much smaller than the errors predicted by both Han’s and Castagna’s empirical formula. Furthermore, the prediction accuracy and generalization performance are better than those of these two common empirical formulas. The forward simulation based on theoretical models supplements the training dataset and provides high-quality labels for machine learning. This can considerably improve the interpretability and generalization of models in real applications of a machine learning algorithm.

Empirical formulas for shear wave velocity prediction and their uncertainties: a case study of thirteen alluvium test sites in the Taipei Basin

Empirical formulas for shear wave velocity prediction and their uncertainties: a case study of thirteen alluvium test sites in the Taipei Basin

Prediction of Shear Wave Velocity Based on a Hybrid Network of Two-Dimensional Convolutional Neural Network and Gated Recurrent Unit

Compressional and shear wave velocities (Vp and Vs, respectively) are important elastic parameters to predict reservoir parameters, such as lithology and hydrocarbons. Due to acquisition technologies and economy, the shear wave velocity is generally lacking. Over the last few years, some researchers proposed deep learning algorithms to predict the shear wave velocity using conventional logging data. However, these algorithms focus either on spatial feature extraction for different physical properties of rocks or on sequential feature extraction in the depth direction of rocks. Only focusing on feature extraction in a direction of rocks might lead to a decrease in prediction accuracy. Therefore, we propose a hybrid network of a two-dimensional convolutional neural network and the gated recurrent unit (2DCNN-GRU), which can establish more complex nonlinear relationships between the input and output data based on the spatial features extracted by 2DCNN and the sequential features extracted by GRU. In this study, two cases are used to validate the reliability and prediction accuracy of the proposed network. Comparing the prediction results of 2DCNN, GRU, and the proposed network, the proposed network shows better performance. Meanwhile, for improving the prediction accuracy of the proposed network, the relationship is analyzed between the prediction accuracy of the proposed network and the length of the input sample.

Research on Prediction Method of Volcanic Rock Shear Wave Velocity Based on Improved Xu–White Model

Volcanic rock reservoirs have received extensive attention from scholars all over the world because of their geothermal, mineral, and oil and gas resources. Shear wave velocity is the essential information for AVO (amplitude variation with offset) analysis and the reservoir description of volcanic rocks. However, due to factors such as cost, technical reasons, and so on, shear wave velocity is not provided in many logging data. This paper proposes a shear wave velocity prediction method suitable for the conventional logging of volcanic rocks. Firstly, the Xu–White model is improved. The probability distributions formed by the prior information of the logging area are used to initialize the key petrophysical parameters in the model instead of the fixed parameter value to establish the statistical petrophysical model between the logging curve and shear wave velocity. Then, based on the Bayesian inversion method, the simulated P-wave velocity is matched with the actual P-wave logging data to calculate the key petrophysical parameters, and is then used for S-wave velocity prediction. The method is applied to the actual logging data of the No. 5 structure in Nanpu Sag, eastern China. The prediction effect of shear wave velocity is better than that of the conventional method, indicating the feasibility and effectiveness of this method. This study will provide more accurate shear wave velocity data for the exploration and development of volcanic reservoirs.

Prediction of Shear Wave Velocity in Fine-grained Soils From Cone Penetration Test Data: Toward a Global Approach

Shear wave velocity ([Formula: see text]) is often used for the calculation of the shear modulus to combat sampling issues in clays and minimize the cost of laboratory testing. Several studies reported correlations between [Formula: see text] and direct and indirect measurements of the cone penetration test (CPT), yet those proposed for fine-grained formations are generally of a local nature. This study sets out to propose a global approach for the estimation of [Formula: see text] for these soils. Seven regressions were developed using a total of 632 CPT data measurements collected from 55 different sites in 12 countries. The correlations utilizing the tip resistance and depth statistically outperform those using other CPT data. A total of 165 CPT points—collected from the state of New Jersey, USA, and Mexico City, Mexico—were used for verification. It is instrumental to analyze CPT data before building global correlations for gauging applicability and understanding limitations. The adopted approach paves the way for more methodical and rigorous global regressions.

Automatic prediction of shear wave velocity using convolutional neural networks for different reservoirs in Ordos Basin

Shear wave velocity (S-wave velocity) has great significance for reservoir characterization and can effectively reduce the ambiguity in seismic interpretation. However, owning to its high cost and technical difficulties, it is usually difficult to obtain for the whole region. Therefore, the prediction of S-wave velocity currently has turned into an urgent task. Thus, the S-wave velocity was predicted using 1D convolutional neural network (1D-CNN) with the well logs. In the first step, the traditional Xu-White and Xu-Payne model were used to carry out the inversion of bulk modulus, shear modulus and density, and illustrated the phenomenon for the traditional rock-physics model. Then, 1D-CNN, including one input layer, four convolutional layers, three fully connected layers and one output layer, was proposed, in which three types of well logs (conventional well logs, array induction well logs and gamma ray spectral well logs) sensitive to the S-wave velocity were determined through the well logs analysis. Finally, based on the established 1D-CNN model, the influence of different well logs combination on the S-wave velocity prediction was analyzed, indicating that the addition of different well logs can improve the accuracy of S-wave velocity. And different machine learning methods (MLs), including back-propagation (BP) and support vector regression (SVR), were also compared, showing that the predicted accuracy of 1D-CNN has been improved by 4.2% on average. Two cases for sandstone and carbonate reservoirs also showed that 1D-CNN proposed can achieve higher accuracy and better performance than traditional MLs and rock-physics models. Furthermore, the proposed method can be applied to the other oil and gas exploration fields, improving the exploration accuracy and increasing the production of oil and gas.

Improved prediction of shear wave velocity for clastic sedimentary rocks using hybrid model with core data

Accurate measurement of acoustic velocities of sedimentary rocks is essential for prediction of rock elastic constants and well failure analysis during drilling operations. Direct measurement by advanced logging tools such as dipole sonic imager is not always possible. For older wells, such data are not available in most cases. Therefore, it is an alternate way to develop a reliable correlation to estimate the shear wave velocity from existing log and/or core data. The objective of this research is to investigate the nature of dependency of different reservoir parameters on the shear wave velocity (Vs) of clastic sedimentary rocks, and to identify the parameter/variable which shows the highest level of dependency. In the study, data-driven connectionist models are developed using machine learning approach of least square support vector machine (LSSVM). The coupled simulated annealing (CSA) approach is utilized to optimize the tuning and kernel parameters in the model development. The performance of the simulation-based model is evaluated using statistical parameters. It is found that the most dependency predictor variable is the compressional wave velocity, followed by the rock porosity, bulk density and shale volume in turn. A new correlation is developed to estimate Vs, which captures the most influential parameters of sedimentary rocks. The new correlation is verified and compared with existing models using measured data of sandstone, and it exhibits a minimal error and high correlation coefficient (R2 = 0.96). The hybridized LSSVM-CSA connectionist model development strategy can be applied for further analysis to predict rock mechanical properties. Additionally, the improved correlation of Vs can be adopted to estimate rock elastic constants and conduct wellbore failure analysis for safe drilling and field development decisions, reducing the exploration costs.

Shear wave velocity prediction based on LSTM and its application for morphology identification and saturation inversion of gas hydrate

Compressional wave velocity (Vp) and shear wave velocity (Vs) are important geophysical parameters for studying gas hydrate systems. However, the Vs parameter is not commonly used in logging-while-drilling because of the high costs of acquiring it. To address this issue, we propose a long short-term memory (LSTM) neural network for predicting a complete shear wave velocity curve at wells in Alaminos Canyon Block 21 (AC 21) based on the limited measurement of Vs logs. The LSTM method performs better than the LSF method when comparing the Vs curves generated by them with the measured one. The measured Vp and predicted Vs curves were applied to determine the distribution of hydrates using the ratio of Vp increment to Vs increment, and we projected velocity samples into the cross plot of modelled Vp and Vs by assuming pore- and fracture-filling hydrate reservoirs to identify the hydrate morphologies. The pore- and fracture-filling gas-hydrate-bearing reservoirs were proved at 541–633 feet below sea floor (fbsf) and 1200–1500 fbsf in hole AC 21-A, respectively. However, the gas hydrate saturations estimated using the joint usage of Vp and Vs at shallow and deep depths are less than 10% (at most 30% locally) and 10%–20%, respectively. Because less Vs data is measured, a transfer learning technique is introduced to estimate a Vs curve at well AC 21-B. Comparing velocity samples with the theoretical velocity curves, the morphology of the hydrate in well AC 21-B is observed to be a horizontal fracture-filling type, and its gas hydrate saturation is approximately 10%–20% at the depth intervals of hydrate occurrence. According to our numerical calculations, the hydrate occurrence at well AC 21-B shows a better continuity than in well AC 21-A. • A LSTM model is developed to predict a complete VS curve in JIP wells. • Morphologies and distributions of gas hydrates are determined. • Gas hydrate saturation is estimated based on the jointly usage of Vp and Vs data. • transfer learning to predict a Vs curve and further analyse gas hydrates.

Estimation of S-wave velocity from wire-line logs for organic-rich rocks

Existing shear-wave velocity prediction methods for organic-rich rocks have not simultaneously taken the effects of organic matter (kerogen) distribution and pore shape on the velocity into consideration. To solve this problem, the current paper proposes a shear-wave velocity prediction method for organic-rich rocks based on the Kuster–Toksöz (KT) rock physics model and chaotic quantum particle swarm optimisation (CQPSO) algorithm. In this method, the organic-rich rocks are considered equivalent to a mixture of minerals, kerogen particles and fluid-containing pores, in which kerogen particles and pores are both equivalent to ellipsoid-shaped inclusions. The effect of kerogen distribution and pore shape on velocity is described according to the change in the aspect ratio of ellipsoids. We used the error between the predicted and measured P-wave velocities to establish the inverse objective function, and introduced the CQPSO algorithm to calculate the equivalent kerogen particle and the equivalent pore aspect ratios. Based on this, the shear-wave velocity is predicted using inverted parameters. Compared with three existing single-parameter adaptive methods, the prediction results show that the proposed method is superior and can be satisfactorily implemented for laboratory measurement and on logging data.

Semi-empirical correlation of shear wave velocity prediction in the Yellow River Delta based on CPT

The Yellow River Delta is one of the most abundant offshore energy areas in China, and the construction of offshore projects is rapidly developing. In this paper, four typical regions of the Yellow River Delta in different periods were selected as research objects. In-situ cone penetration tests (CPTs), geotechnical testing, electron microscope scanning testing, and soil type indexing (Ic) was used to classify and identify the soil layers. At the same time, a modified calculation model based on CPT for the shear wave velocity of fine-grained seabed soil was sought and verified with the measured values. The results showed that: (1) the obtained soil layers are mainly classified by clay and silty soil, which is in line with the conclusions of previous scholars on the classification of the Yellow River Delta soil layer. (2) The measured shear wave velocity is generally between 100 m/s and 250 m/s, and it gradually increases with depth. (3) The depth, cone resistance, and soil behavior index are directly related to the calculation of the shear wave velocity in this area. (4) Discrete analysis of measured and calculated shear wave velocity provides theoretical support for the change of shear wave velocity under different conditions.

Numerical Investigation of the Effect of Heterogeneous Pore Structures on Elastic Properties of Tight Gas Sandstones

Strong heterogeneity of pore microstructures leads to complicated velocity-porosity relationships in tight sandstone that cannot be well explained by conventional empirical formulas. To better understand the effect of complex pore structures on elastic properties of tight gas sandstone, we compared three rock physics models. In the first model, we used a single aspect ratio value to quantify varied pore geometry in the tight sands. In the second model, complex pore space was equivalent to the combination of high-aspect-ratio round pores (stiff pores) and low-aspect-ratio compliant microcracks (soft pores). In the third multiple pore-aspect-ratio model, pore spaces are represented using a set of pores with varied values of aspect ratio following statistical normal distribution. Modeling results showed that complex velocity-porosity relationships could be interpreted by the variations in pore aspect ratio in the first model, by the fraction of soft pores in the second model, and by the mean value and variance in the third statistical model. For a given mean value in the third model, higher variance of the multiple pore-aspect-ratio indicated stronger heterogeneity of pore spaces. Further studies on rock physical inversion showed that, compared with the first single pore-aspect-ratio model, the second dual-pore model gave better prediction in shear wave velocity by regarding the soft pore fraction as a fitting parameter. This finding revealed that the dual-pore model could be a more realistic representation of tight sandstone. The third statistical model showed comparable precision in the prediction of shear wave velocity compared with the dual-pore model; however, uncertainty existed for simultaneously determining mean value and variance of pore aspect ratio. On the basis of the dual-pore model, we evaluated the elastic modulus of dry frames of the tight sandstone using logging data in a borehole. Compared with empirical formulas, such as the Krief methods, the method in this paper provided a more rigorous way to determine elastic properties of dry frames for the tight sandstone. Comparisons of rock physical modeling methods offer a better understanding of the microstructures controlling the elastic behaviors of tight gas sandstone.

Shear wave velocity estimation utilizing statistical and multi-intelligent models from petrophysical data in a mixed carbonate-siliciclastic reservoir, SW Iran

The conjugation of two or more Artificial Intelligent (AI) models used to design a single model that has increased in popularity over the recent years for exploration of hydrocarbon reservoirs. In this research, we have successfully predicted shear wave velocity (Vs) with higher accuracy through the integration of statistical and AI models using petrophysical data in a mixed carbonate-siliciclastic heterogeneous reservoir. In the designed code for multi-model, first Multivariate Linear Regression (MLR) is used to select the more relevant input variables from petrophysical data using weight coefficients of a suggested function. The most influential petrophysical data (Vp, NPHI, RHOB) are passed to Ant colony optimization (ACOR) for training and establishing initial connection weights and biases of back propagation (BP) algorithm. Afterward, BP training algorithm is applied for final weights and acceptable prediction of shear wave velocity. This novel methodology is illustrated by using a case study from the mixed carbonate-siliciclastic reservoir from one of the Iranian oilfields. Results show that the proposed integrated modeling can sufficiently improve the performance of Vs estimation, and is a method applicable to mixed heterogeneous intervals with complicated diagenetic overprints. Furthermore, predicted Vs from this model is well correlated with lithology, facies and diagenesis variations in the formation. Meanwhile, the developed AI multi-model can serve as an effective approach for estimation of rock elastic properties. More accurate prediction of rock elastic properties in several wells could reduce uncertainty of exploration and save plenty of time and cost for oil industries.