Reservoir Physical Parameters Research Articles

Time-varying quantitative characterization of reservoir microscopic pore structure based on digital core

The reservoir physical parameters (such as porosity, permeability, phase permeability, etc.) will change in the process of water injection development with the continuous flushing effect of high magnification water drive, which may have adverse effects on development if they cannot be accurately characterized. In this paper, we try to quantitative characterize the time -varying features of pore structure from the microscopic perspective. We used high-resolution 3D micro-CT (MCT) images to supplement the analysis of the physical properties of the rocks, and we used 3D pore network modeling technology to generate a 3D model of the rock samples. Then the relevant static micro-pore parameters are derived to study the changes of rock properties in different periods of development. Pore parameters were obtained to study the changes in porosity, permeability and pore size distribution of the rocks under different expulsion multiples, and to visualize the time-varying pattern of the microstructure of the rock samples under different stages of water injection. Under low-multiple water-driven conditions (10 PV), the physical properties of the rock samples did not change significantly, with an average facies change of only 1.98%. When reaching high multiples of water drive (more than 500PV), the change rate reaches more than 15%. And with the increase of water drive multiples, the distribution of pore radius and throat radius of rock samples showed an obvious right-shift trend, and the pores and throats increased, which was consistent with the results of CT scanning photographs. Equation fitting of the simulation results reveals that the core micro-parameters show an overall logarithmic relationship with the change of water drive. In this paper, a new application of Micro CT is provided to quantify the microstructure of water-injected reservoirs. By adopting this method, the effect of water-drive multiplicity on reservoir structure can be illustrated from a microscopic perspective. Meanwhile, compared with previous studies, this paper further investigates the effect of time-variation based on the qualitative analysis of the effect of time-variation of water-driven in reservoirs, and regresses time-variation curves and formulas through the changes in multiple water-driven states, which fills in the gap of the current lack of quantitative research and provides new understanding and optimization of the development of oilfields. The study will provide a new perspective for understanding and optimizing oilfield development.

A Study on the Accurate Equation Inversion Method for Uncoupled Interface Parameters of OA Media Considering Initial Stress

Underground media are generally affected by the initial stress of the overlying strata. At the same time, the uncoupled interface is an important influencing factor of seismic response, as well as an essential reservoir environment, reservoir space, and migration channel. It is significant to explore the parameters of the uncoupled interface under stress for identifying oil and gas reservoirs. As an effective means of predicting reservoir physical parameters, pre-stack seismic inversion uses information on uncoupled interface fractures under stress to improve the accuracy of reservoir parameter prediction, overcome the problem of insufficient interface parameter prediction methods, and promote oil and gas reservoir exploration and development. Based on the precise equations for directly characterizing uncoupled interface fracture weakness parameters, this paper uses the ANNI inversion method to realize the nonlinear inversion method of uncoupled interface fracture parameters. The feasibility and practicability of predicting underground medium parameters from seismic response characteristics have been verified through model testing and actual work area application.

Research on the approximate equation inversion method for the physical properties parameters of OA media with decoupling of stress and fracture effects

Pre-stack seismic inversion is an effective method for predicting reservoir physical parameters. At the same time, azimuth seismic data is a reliable source of information for predicting underground reservoir elastic parameters, fracture parameters, and interface parameters and has rich underground information. However, it is difficult to determine whether the anisotropy of the reservoir is caused by stress or the original fractures. Therefore, distinguishing the influence of fractures and initial stress in the reservoir from azimuth seismic data is the key to predicting reservoir parameters. Because of the problem of insufficient consideration of the influence of stress and fracture weakness in OA media, based on Born's first-order scattering theory and Bayesian inversion framework, an approximate equation for the reflection coefficient decoupling fracture anisotropy and stress-induced anisotropy was constructed, and a linear inversion method based on the approximate equation was developed, realizing the exploration of methods for predicting fracture weakness and stress action parameters. Model testing and actual work area application have verified the feasibility and practicability of predicting underground medium parameters from seismic response characteristics.

A data-driven dual-optimization hybrid machine learning model for predicting carbon dioxide trapping efficiency in saline aquifers: Application in carbon capture and storage

Geological carbon sequestration (GCS) is a promising method to reduce global carbon emissions. To better understand the CO2 trapping mechanism, the CO2 sequestration process was simulated through a physics-based reservoir simulator, and the residual trapping (RTI) and solubility trapping (STI) efficiencies were analyzed. Through the experiments, a database containing 4543 data points with reservoir physical parameters and rock properties was established. In this study, a kernel extreme learning machine (KELM) model based on variational mode decomposition (VMD) and dung beetle optimization optimizer (DBO) was proposed to perform the prediction task. Through the VMD, the original data sequences were decomposed and combined, and the obtained subsequences were executed modeling prediction and reconstruction respectively. The reliability of the prediction model was confirmed to be greater than 97.71% through domain analysis. In addition, it was applied to a tight gas reservoir in China in to confirm its validity in field applications. Comparative results indicated a high agreement between the prediction results and the simulation results, suggesting that the proposed method can complement numerical simulations and help users understand the application of machine learning in carbon capture and storage (CCS).

Revised inflow performance relationship for productivity prediction from marine sandy hydrate reservoirs in Nankai Trough

Productivity prediction is crucial for the safe and efficient exploitation of marine methane hydrate. As a feasible method, inflow performance relationship (IPR) can characterize productivity results and provide a theoretical basis for optimization of depressurization schemes. In this study, we propose two modified IPR models based on field data from limited trial production and numerical fitting in the Nankai Trough. Prediction results are obtained by depressurization, together with IPR curves under different recovery ratios. According to the conventional formulas, the reservoir pressure of the deliverability equation is revised, considering the dual contributions of hydrate dissociation and pressure drop. The comparison results show that the modified Vogel formula and Fetkovich formula can suitably describe the IPR for the single vertical well and horizontal well, respectively. Sensitivity analyses of various reservoir physical parameters indicate that changes in porosity and hydrate saturation cause diverse values for the undetermined coefficient A, primarily due to the alterations in the key coefficient pr′ in the revised Vogel formula for the vertical well scenario. However, absolute permeability has only a slight influence on the shape of the IPR curve. In contrast, absolute permeability plays a crucial role in affecting the Fetkovich-modified IPR curves for horizontal wells. Productions under varying porosities exhibit a consistent linear trend on the Fetkovich-modified IPR curve because of the comparable impact of gas production rate and pressure square variance. Besides, changing hydrate saturation can also have some effect on the IPR. Our research provides a feasible productivity prediction methodology for the safe and efficient extraction of similar hydrate reservoirs.

Five-dimensional facies-driven seismic inversion for igneous reservoirs based on rock-physics modelling

Abstract The igneous reservoir has become an important exploration target for increasing reserves and production of oil and gas in the Junggar Basin. However, igneous reservoir exploration is restricted because the seismic exploration of high-quality igneous reservoirs is difficult and the anisotropy induced by high-angle fractures cannot be neglected. To implement the characterization of igneous reservoirs, we first study the correlation between anisotropy parameters and physical properties of igneous rock, and we propose a 5D facies-driven inversion method based on rock physics, which means we employ 3D seismic data at different incidence angles and azimuths to implement the estimation of a hydrocarbon reservoir constrained by the igneous rock facies. We also present an anisotropic igneous rock-physics model, in which micro-petrophysical characteristics, strong heterogeneity of skeleton minerals, and pore structures are considered. As a reasonable initial model is important for seismic inversion, we propose a facies-driven modelling seismic inversion method in which we use facies obtained on the basis of the difference between rock composition, reservoir physical parameters, and elastic parameters of different lithofacies igneous rocks to constrain the seismic inversion. Finally, we present a step seismic inversion method of using seismic data to estimate multi-parameters of horizontal transversely isotropic media. Therefore, the comprehensive processes of rock-physics modelling, inversion model establishment, and reservoir prediction of high-quality igneous rocks are proposed in this study, which demonstrates effective application for igneous reservoirs in China.

Advancing Quantitative Seismic Characterization of Physical and Anisotropic Properties in Shale Gas Reservoirs with an FCNN Framework Based on Dynamic Adaptive Rock Physics Modeling

Quantitative seismic methods are crucial for understanding shale gas reservoirs. This study introduces a dynamic adaptive rock physics model (DARPM) designed to systematically quantify the relationship between physical parameters and elastic parameters within shale formations. The DARPM uniquely adapts to changes in formation dip angle, allowing adaptive reservoir property assessment. An innovative adaptive rock physics inversion methodology is subsequently proposed to compute values for reservoir physical and seismic anisotropy parameters. This is achieved using well log data and building upon the foundation laid by the established DARPM. We introduce the RPM-FCNN (rock physics model—fully connected neural network) framework, seamlessly integrating the DARPM with the corresponding inversion results into a comprehensive model. This framework facilitates a quantitative analysis of the nonlinear relationship between elastic and reservoir physical parameters. Utilizing the trained RPM-FCNN framework, the spatial distribution of reservoir and seismic anisotropic characteristics can be precisely characterized. Within this framework, the organic matter mixture aspect ratio indicates the continuity of organic matter, while the organic matter porosity reveals the maturity of organic matter. Simultaneously, seismic anisotropy characteristics signify the degree of stratification within the reservoirs. This method, therefore, establishes a robust foundation for identifying favorable areas within shale gas reservoirs.

A New Method for Evaluating Fracability Based on Cluster Analysis and Well Logging Petrophysical Facies Division

The fracability of rock is a crucial parameter in determining the efficiency of fracturing in tight reservoirs. However, current methods for evaluating fracability have significant limitations. This paper proposes a new method for evaluating reservoir fracability based on cluster analysis of logging petrophysical facies. This method considers rock mechanics characteristics, petrophysical response characteristics, and reservoir physical parameters, among other factors. The Well X of the block was used as the research subject. Using the brittleness index and permeability evaluation index, the reservoir fracability of the target block is classified into five categories: high brittleness and low permeability rock, high brittleness and super low permeability rock, low brittleness and low permeability rock, low brittleness and super low permeability rock, and clay/shale. Geolog software is used to generate fracability prediction profiles for representative wells based on petrophysical facies. To further validate this evaluation method, the fracture construction data and microseismic monitoring interpretation results of Well Y in a tight block of an oil field were synthesized and analyzed to predict its fracability. The overall prediction accuracy reached 81.8%, indicating that the new fracability evaluation method has good application effect.

Application of volumetric methods in estimating geothermal potential based on reservoir physical parameters (a case study of geothermal area "Z")

Application of volumetric methods in estimating geothermal potential based on reservoir physical parameters (a case study of geothermal area "Z")

Research on Displacement Efficiency by Injecting CO2 in Shale Reservoirs Based on a Genetic Neural Network Model

Carbon dioxide injection can help solve two issues in shale reservoir production. Firstly, it can reduce carbon emissions while, secondly, improving unconventional reservoir recovery. There are many controlling factors for CO2 injection to enhance oil recovery in shale reservoirs, and the effect of field implementation varies greatly. The key to popularizing this extraction technology is determining the main controlling factors of CO2 displacement efficiency. Using CO2 shale displacement laboratory results, the grey correlation analysis method was used to determine the main controlling factors affecting core oil displacement efficiency, such as shale reservoir physical parameters (rock compressibility, porosity, median pore size, matrix permeability, TOC, and oil saturation) and engineering parameters (soaking time and injection pressure). The genetic algorithm (GA) was introduced to optimize the backpropagation (BP) neural network to construct the prediction model of the CO2 indoor displacement experiments in shale cores. The results showed that the injection pressure among the engineering parameters, the CO2 soaking time among the gas injection parameters, and the porosity among the shale physical parameters were the main controlling factors affecting the oil displacement efficiency. The prediction accuracy of the genetic neural network model improved, and the coefficient of determination (R2) reached 0.983. Compared with the conventional neural network model, the mean absolute error (MAE) was reduced by 30%, the root mean square error (RMSE) was reduced by 46%, and the R2 increased by 11%. Optimizing the learning and training of the prediction model significantly reduces the cost of laboratory experiments. The deep-learning model completed by training can intuitively show the degree of influence of input parameters on output parameters, providing a theoretical basis for the study of CO2 displacement mechanisms in shale reservoirs.

Research on pore structures of fine-grained carbonate reservoirs and their influence on waterflood development

Abstract Carbonate reservoir has complex pore structures. At present, the influence of pore structure on water flooding mechanism of carbonate reservoirs is insufficient. In this article, a systematic workflow was designed in combination with scanning electron microscope, particle size, physical properties, and water flooding experiments to study the effect of pore structure on water flooding mechanism of fine-grained carbonate rocks. Due to the small particle size and strong heterogeneity, the acid fracturing operations, rather than hydraulic fracturing alone, are necessary to achieve increased production and reservoir reconstruction of carbonate reservoirs. Through this study, the mathematical model of reservoir physical parameters (permeability and porosity) was proposed, and the accuracy of the model was verified by comparing the simulated oil recovery with the experimental results. According to the comparison results, the experimental results are consistent with the simulation results, and their oil recovery efficiency is 33.62 and 31.87%, respectively. Finally, the effect of injection rate on oil production was discussed. It is shown that with the increase in injection rate, the output of displaced oil increases significantly. The cumulative oil production increases from 62.5 to 256.31 mL when the injection rate increases from 100 to 400 mL/min. The findings of this study can help for better understanding of the influencing factors and mechanisms of the development efficiency of carbonate reservoirs.

“Double sweet spot” identification method via cluster analysis of petrophysical properties from well logg data: A case study of a tight sandstone reservoir

In recent years, interest in unconventional resource plays has grown substantially due to horizontal drilling and hydraulic fracturing techniques. However, the high cost of drilling, choosing the right locations for perforation intervals during hydraulic fracturing is a crucial issue. Therefore, identifying so called “sweet spot” with high potential for hydrocarbon production plays an important role in oil and gas industry. In this study a new “double sweet spot” identification method using cluster analysis of well-logged petrophysical properties of facies is presented. Firstly, an engineering sweet spot identification method based on geophysical responses parameters (brittleness index and permeability) is introduced to classify the target zone. Triaxial compression tests are performed to obtain the static Young's modulus and Poisson's ratio of the rock samples. To calculate dynamic Young's modulus and Poisson's ratio, a novel model for determining interval transit time of S-wave is proposed based on the multiple linear regression theory according to well logging data. By performing the cluster analysis, an intersection diagram of permeability and calculated normalized Rickman's brittleness is built and then engineering sweet spots in target zone are characterized based on well-logged petrophysical properties of facies. By plotting the intersection diagram of porosity and oil saturation based on the cluster analysis theory, geological sweet spots in target zone were characterized and four types of rock petrophysical facies are divided: Sandstone A (So＞48%, 8＜φ＜19%); Sandstone B (31＜So＜48%, 6＜φ＜18%); Sandstone C (18＜So＜31%, 8＜φ＜20%) and Sandstone D (0＜So＜18%, 8＜φ＜17%). Secondly, target formation classification is carried out using a proposed geological sweet spot identification method based on reservoir physical parameters (oil saturation and porosity). The relationship between core electrical resistivity and water saturation is determined by traditional Archie's formula and then an oil saturation prediction model in target reservoir is established. Based on the experimental porosity and permeability measurements of 47 core samples and well logging data, a porosity interpretation model is proposed. By plotting the intersection diagram of porosity and oil saturation based on the cluster analysis theory, geological sweet spots in target formation are characterized and four types of rock petrophysical facies are given: Sandstone A (K＞0.1mD, BI＞38); Sandstone B (K＞0.1mD, 33＜BI＜38); Sandstone C (K＞0.1mD, BI＜33) and Sandstone D (0.01＜K＜0.1mD, 10＜BI＜60). The integrated “double sweet spot” identification results of Well 1 in Daqing tight sandstone oil reservoir are given. Rock AA is the most favorable type of petrophysical rock facies and rock DD is the worst favorable type of petrophysical rock facies. Finally, an integrated “double sweet spot” identification method proposed in this study was applied to screen proper interval candidates during refracturing operation in a tight sandstone oil reservoir in China. After refracturing operation a gradually increase of average oil production in well group was found from 1.10 t/d to 5.63 t/d. This novel “double sweet spot” identification method has applicability in other tight oil reservoirs around the world.

Establishment of sample pool and prediction of saturation field in complex fault block reservoir based on deep learning algorithm

The remaining oil prediction of complex fault block reservoir is the key to guide the well deployment in the late high water cut stage. We formed the construction methods of 20,000 numerical simulation forward models, including the dynamic and static attributes of the reservoir, such as structural depth, reservoir thickness, permeability, porosity, injection and production fluid volume, etc. The saturation field distribution of each model is obtained through numerical simulation, so as to form the saturation field characteristic data sample database. Through the deep convolutional generative adversative neural network model, the data of sample pool was trained, in which 70% of the data was randomly selected as the training set and 30% of the data as the test set. Finally, the saturation field prediction method that could be applied to an actual block was established. The result showed that: the new method did not need to carry out the numerical simulation research on the actual region, the remaining oil saturation distribution at different time step that obtained by the deep convolutional generative adversative neural network model only need to input the reservoir physical parameters, well location coordinates, injection production and other information of the actual block, and the prediction accuracy of the method was more than 90%. The test showed that the deep convolutional generative adversative neural network model had good generalization ability, and the model could be widely used in the prediction of residual oil saturation field of complex fault block reservoirs, which greatly improved the efficiency of reservoir research.

Porosity prediction from pre-stack seismic data via a data-driven approach

Porosity estimation plays an important role in geophysical exploration and reservoir development, which is a highly complicated and extremely challenging problem. Usually, the model-driven inversion for reservoir physical parameters consists of the rock-physics modelling and the probability statistical inversion. However, the rock-physics modelling usually is intricate and the efficiency of inversion algorithm is low. Besides, the relationship between seismic elastic properties and petrophysical parameters is highly non-linear. Therefore, the deep learning algorithm is introduced for two important aspects, rock-physics modelling and reservoir parameters estimation. We build two neural network models, the PorNet model and the BlstmNet model. The former is based on classical convolutional neural networks (CNNs), and the latter is based on the variant network of traditional recurrent neural networks (RNNs). CNNs extract and compress the spatial characteristics between input data and labeled data using multi-layer convolutions, while the bidirectional long short-term memory (BLSTM) neural network conducts porosity prediction by processing the internal correlation of the labeled data. First, we conduct the experiment on synthetic data set for evaluating the performance of these two neural network models on the prediction accuracy and the convergence speed. The experimental results show that the BlstmNet model provide a more precise prediction result and the PorNet model has a more stable convergence process. To further evaluate the generalization ability for two neural network models, an example for porosity prediction is carried out on oil field data. The prediction results demonstrate that two neural network models have ability to estimate reservoir properties with high accuracy and lower artificial cost. The data-driven inversion of reservoir physical parameters has certain feasibility and application prospects.

Prediction of reservoir physical parameters based on adaptive step size stochastic inversion

Reservoir physical parameter prediction is the ultimate goal of the quantitative seismic interpretation. Rock physics inversion is a method to obtain reservoir physical parameters from seismic elastic properties and the accuracy of predicted physical parameters depends largely on the inversion algorithm. We proposed an adaptive step size stochastic algorithm. In the algorithm, the step size of the parameters to be inverted are not fixed, but adaptive. Firstly, the parameters to be inverted are sampled according to their prior probability density distribution, and the error of each sampling point is calculated. Then select the sampling points with small errors and calculate their probability density distribution, and then encrypt the sampling point according to the new probability density distribution. Repeat the sampling several times, and finally get the optimal solution to the parameters to be inverted. Compared with the conventional given step size algorithm, the adaptive step size algorithm is more accurate. The prediction of reservoir physical parameters based on adaptive step size stochastic inversion is applied to sandstone reservoir, obtained reservoir physical parameters, such as porosity, water saturation, clay content, and their uncertainty distribution.

Study on the Lower Limits of Physical Parameters for Heavy Oil Reservoirs during the In Situ Combustion Process.

As an effective enhanced oil recovery (EOR) method in the late stage of the steam injection process, in situ combustion (ISC) has been verified by more and more researchers owing to its high recovery efficiency and low operation cost. However, the ISC process is prone to bringing about an unfavorable result with a low sweep efficiency and even an extinction process due to the complexity and uncertainty in oil and water distribution as well as the heterogeneity of the actual reservoir. In this work, combustion tube experiments combined with numerical simulations by CMG-STARS were conducted to study the influence of a reservoir's physical parameters on the stability of ISC. It was found that the sustainability of the combustion process is sensitive to permeability and oil saturation, the lower limits of which are determined as 0.2 μm2 and 0.2, respectively. Then, a formula was preliminarily proposed as 0.25 ≤ So/K ≤ 0.8, which could be used as a reference of screening criteria in designing the ISC process.

Numerical investigation of natural gas hydrate production performance via a more realistic three-dimensional model

Numerical simulation plays a crucial role in the prediction of natural gas hydrate production performance. However, most existing models are two-dimensional or three-dimensional with idealized geometries and uniform parameter assignments, which cannot depict the effects of stratigraphic undulation and spatial variability of reservoir physical parameters on gas production. Therefore, a convenient method to convert the image information (more accessible) into parameter attribute values (required for model construction) was proposed in this study. Using the converted data of reservoir depth, thickness, and porosity, a more realistic three-dimensional model was innovatively constructed. Then, the influences of reservoir fluctuations and spatial variability of physical parameters on production performance were quantitatively analyzed. It was found that placing the production well in an elevated area can facilitate gas production. Specifically, Well 1 (located in the highland) had a 34.1% higher normalized gas production rate (i.e., production rate per unit well length) and a 14.9% lower normalized water production rate than Well 3 (located in the flat area) in the free gas layer. In addition to reservoir fluctuations, the exploitation efficiency of the gas hydrate-bearing layer was also affected by the thickness. The spatial variability of hydrate saturation and that of gas saturation in the study area were not very prominent, and the gas production rate obtained by the heterogeneous scheme was approximately 10% different from that of the homogeneous scheme. However, although the spatial variability of porosity was also not great (no more than 2%), when the cubic law was used to calculate the corresponding permeability, the gas production rate obtained by the heterogeneous scheme was nearly 20% different from that of the homogenous scheme. This study demonstrates the need to use a more realistic three-dimensional model for gas hydrate production performance prediction and is expected to provide an important reference for well location selection.

Parameter Sensitivity Analysis of the Hydraulic Fracture Growth Geometry in a Deep Shale Oil Formation: An Experimental Study

The depth of shale oil of Fengcheng Formation in Mahu of Junggar Basin, China, is 4500-5000 m. The horizontal principal stress difference of deep shale reservoir is high, which makes it difficult to form complex fractures during fracturing reconstruction. In order to fully understand the law of hydraulic fracture propagation in the formation during fracturing construction, the anisotropy characteristics and basic reservoir physical parameters (mineral composition and rock strength parameters) of rock were obtained through mineral composition test and indoor rock mechanics test (Brazil splitting test), and it was found that the heterogeneity was strong. The true triaxial fracturing simulation experimental system is used to carry out experimental research on full-diameter core rock samples, and the propagation patterns of hydraulic fractures under the influence of different geological factors (in situ stress difference and natural fractures) and engineering factors (pumping rate and fracturing fluid viscosity) are compared and analyzed. The results show that the in situ stress is the most important factor affecting fracture propagation, which determines the direction and shape of fracture propagation. The natural weak surface (lamina/bedding and natural fractures, etc.) in shale reservoir is an important reason for complex fractures. The nature of the weak plane, occurrence, and in situ stress jointly determine whether the fracture can extend through the weak plane. With the increase of pumping rate (18 mL/min to 30 mL/min), the ability of hydraulic fractures to penetrate layers is continuously enhanced. The horizontal principal stress difference of deep shale reservoir is high, and the low viscosity fracturing fluid (10 mPa·s) tends to activate the horizontal bedding, while the high viscosity fracturing fluid (80 mPa·s) tends to directly penetrate the bedding to form the vertical main fracture. Therefore, the fracturing technology of alternating injection of prehigh viscosity fracturing fluid and postlow viscosity fracturing fluid can be adopted to maximize the complexity of fracturing fractures in deep shale reservoirs. The research results are designed to provide theoretical guidance for prediction of hydraulic fracturing fracture propagation in shale reservoir and have certain reference significance for field construction.

Two-stage semi-supervised learning inversion for reservoir physical parameters

Reservoir physical parameters are crucial indexes for reservoir description. Deep learning is a data-driven algorithm, which may avoid some issues in conventional deterministic and stochastic inversion methods, such as cumulative and approximate errors. We develop a two-stage semi-supervised learning inversion method to predict porosity, shale content, and water saturation. Firstly, an inversion network is designed to predict physical parameters from pre-stack seismic data and initial models. The Spatio-temporal characteristics of seismic data and the low-frequency structural information of the initial model can be utilized efficiently by the network. In addition, considering the high production cost of label data in exploration geophysics, we construct a semi-supervised learning inversion workflow to relieve the dependence on the label in training. The rock physics model and seismic convolution forward model are invoked as a forward process in the workflow. Finally, we propose a two-stage inversion strategy to deal with the inversion of physical parameters with different data characteristics. The proposed method is applied to a field survey successfully. A high-resolution and high-accuracy inversion result can be obtained with our method, and potential locations of the gas reservoir can be indicated by prediction results. Meanwhile, the comparisons of the two stages demonstrate that illusions in predicted water saturation can be eliminated effectively.

Reservoir Parameter Prediction Based on the Neural Random Forest Model

Porosity and saturation are the basis for describing reservoir properties and formation characteristics. The traditional, empirical, and formulaic methods are unable to accurately capture the nonlinear mapping relationship between log data and reservoir physical parameters. To solve this problem, in this study, a novel hybrid model (NRF) combining neural network (NN) and random forest (RF) was proposed based on well logging data to predict the porosity and saturation of shale gas reservoirs. The database includes six horizontal wells, and the input logs include borehole diameter, neutron, density, gamma-ray, and acoustic and deep investigate double lateral resistivity log. The porosity and saturation were chosen as outputs. The NRF model with independent and joint training was designed to extract key features from well log data and physical parameters. It provides a promising method for forecasting the porosity and saturation with R2above 0.94 and 0.82 separately. Compared with baseline models (NN and RF), the NRF model with joint training obtains the unsurpassed performance to predict porosity with R2above 0.95, which is 1.1% higher than that of the NRF model with independent training, 3.9% higher than RF, and superiorly greater than NN. For the prediction of saturation, the NRF model with joint training is still superior to other algorithms, with R2above 0.84, which is 2.1% higher than that of the NRF model with independent training and 7.0% higher than RF. Furthermore, the NRF model has a similar data distribution with measured porosity and saturation, which demonstrates the NRF model can achieve greater stability. It was proven that the proposed NRF model can capture the complex relationship between the logging data and physical parameters more accurately, and can serve as an economical and reliable alternative tool to give a reliable prediction.