Fig 24- - uploaded by Jinding Zhang
Content may be subject to copyright.
Source publication
Although researchers have applied many methods to history matching, such as Monte Carlo methods, ensemble-based methods, and optimization algorithms, history matching fractured reservoirs is still challenging. The key challenges are effectively representing the fracture network and coping with large amounts of reservoir-model parameters. With incre...
Similar publications
![FIGURE 1 | Schematic of RAE architecture (figure modified from [5]).](publication/353196511/figure/fig1/AS:1044943023910912@1626145239350/Schematic-of-RAE-architecture-figure-modified-from-5_Q320.jpg)
Data-space inversion (DSI) is a data assimilation procedure that directly generates posterior flow predictions, for time series of interest, without calibrating model parameters. No forward flow simulation is performed in the data assimilation process. DSI instead uses the prior data generated by performing O ( 1000 ) simulations on prior geomodel...
Citations
... AI technology has achieved significant research outcomes in recent years, with various AI-based predictive models widely recognized for their ability to handle nonlinear problems [23]. AI technology has been applied in all stages of oil and gas exploration and development, including reservoir modeling and characterization [24][25][26], drilling and production optimization [27,28], and enhanced recovery scheme selection [29]. In the drilling fluid industry, combining numerical simulation and AI technologies with the expertise of drilling fluid specialists can significantly improve the design and optimization of drilling fluids [30][31][32][33]. ...
Drilling fluid is pivotal for efficient drilling. However, the gelation performance of drilling fluids is influenced by various complex factors, and traditional methods are inefficient and costly. Artificial intelligence and numerical simulation technologies have become transformative tools in various disciplines. This work reviews the application of four artificial intelligence techniques—expert systems, artificial neural networks (ANNs), support vector machines (SVMs), and genetic algorithms—and three numerical simulation techniques—computational fluid dynamics (CFD) simulations, molecular dynamics (MD) simulations, and Monte Carlo simulations—in drilling fluid design and performance optimization. It analyzes the current issues in these studies, pointing out that challenges in applying these two technologies to drilling fluid gelation performance research include difficulties in obtaining field data and overly idealized model assumptions. From the literature review, it can be estimated that 52.0% of the papers are related to ANNs. Leakage issues are the primary concern for practitioners studying drilling fluid gelation performance, accounting for over 17% of research in this area. Based on this, and in conjunction with the technical requirements of drilling fluids and the development needs of drilling intelligence theory, three development directions are proposed: (1) Emphasize feature engineering and data preprocessing to explore the application potential of interpretable artificial intelligence. (2) Establish channels for open access to data or large-scale oil and gas field databases. (3) Conduct in-depth numerical simulation research focusing on the microscopic details of the spatial network structure of drilling fluids, reducing or even eliminating data dependence.
... However, unlike conventional reservoirs, fractured reservoirs are composed of fractures with discrete and irregular characteristics, which is difficult to be processed by conventional DL methods. Zhang et al. [9] employed a deep sparse autoencoder model to develop the re-parameterization method for the inverse modeling of fractured reservoirs. In this approach, a deep artificial neural network (ANN) model was utilized to capture the features of the complete geometry parameters of the fracture network. ...
Inverse modeling can estimate uncertain parameters in subsurface reservoirs and provide reliable numerical models for reservoir development and management. The traditional simulation-based inversion method usually requires numerous numerical simulations, which is time-consuming. Recently, deep learning-based surrogate models have been widely studied as an alternative to numerical simulation, which can significantly improve the solving efficiency of inversion. However, for reservoirs with complex fracture distribution, constructing the surrogate model of numerical simulation presents a significant challenge. In this work, we present a deep graph learning-based surrogate model for inverse modeling of fractured reservoirs. Specifically, the proposed surrogate model integrates the graph attention mechanisms to extract features of fracture network in reservoirs. The graph learning can retain the discrete characteristics and structural information of the fracture network. The extracted features are subsequently integrated with a multi-layer recurrent neural network model to predict the production dynamics of wells. A surrogate-based inverse modeling workflow is then developed by combining the surrogate model with the differential evolutionary algorithm. Numerical studies performed on a synthetic naturally fractured reservoir model with multi-scale fractures illustrate the performance of the proposed methods. The results demonstrate that the proposed surrogate model exhibits promising generalization performance of production prediction. Compared with tens of thousands of numerical simulations required by the simulation-based inverse modeling method, the proposed surrogate-based method only requires 1000 to 1500 numerical simulations, and the solution efficiency can be improved by ten times.
... Zhang et al. [19] proposed a new production optimization framework combining advanced deep reinforcement learning techniques. Zhang et al. [20] proposed a new multi-scale fracture network characterization method and a strong dimensionality reduction method based on a model parameter autoencoder. Zhang et al. [21] established a deep learning model based on FNO for three types of PDE control problems of two-dimensional subsurface oil-water two-phase flow. ...
The development of shale gas reservoirs often involves the utilization of horizontal well segmental multi-stage fracturing techniques. However, these reservoirs face challenges, such as rapid initial wellhead pressure and production decline, leading to extended periods of low-pressure production. To address these issues and enhance the production during the low-pressure stage, pressurized mining is considered as an effective measure. Determining the appropriate pressurization target and method for the shale gas wells is of great practical significance for ensuring stable production in shale gas fields. This study takes into account the current development status of shale gas fields and proposes a three-stage pressurization process. The process involves primary supercharging at the center station of the block, secondary supercharging at the gas collecting station, and the introduction of a small booster device located behind the platform separator and in front of the outbound valve group. By incorporating a compressor, the wellhead pressure can be reduced to 0.4 MPa, resulting in a daily output of 12,000 to 14,000 cubic meters from the platform. Using a critical liquid-carrying model for shale gas horizontal wells, this study demonstrates that reducing the wellhead pressure decreases the critical flow of liquid, thereby facilitating the discharge of the accumulated fluid from the gas well. Additionally, the formation pressure of shale gas wells is estimated using the mass balance method. This study calculates the cumulative production of different IPR curves based on the formation pressure. It develops a dynamic production decline model for gas outlet wells and establishes a relationship between the pressure depletion of gas reservoirs and the cumulative gas production before and after pressurization of H10−2 and H10−3 wells. The final estimated ultimate recovery of two wells is calculated. In conclusion, the implementation of multi-stage pressurization, as proposed in this study, effectively enhances the production of, and holds practical significance for, stable development of shale gas fields.
... The evaporation loss of oil is large, and the oil loss rate is as high as 1.5~3% [5]. The study of oil and gas loss in storage tanks is conducive to discovering the loss problems in the production process of crude oil [6][7][8][9][10][11][12][13], and carrying out the corresponding transformations is conducive to promoting low-carbon energy saving and improving the economic benefits in oilfields [14,15]. ...
The process of storing oil depots and combined station tanks is affected by factors such as process technology, equipment, and management methods. Inevitably, some heavy hydrocarbon components will condense. According to the available literature, the existing detection methods are not enough to accurately measure the component composition so that the proportion of heavy hydrocarbon substances in the lost gas is reduced. In this paper, by inventing a homogeneous reduction device, the lost gas in the entire laboratory process was kept in a homogeneous state so that the gas components were well-retained. Using the homogeneous reduction method and a traditional inspection method, gas chromatography was performed on a standard gas and the on-site lost gas, respectively. The standard gas measurement results show that the mean deviations of the homogeneous reduction method and the traditional test method were −3.45% and −11.62%, respectively, and the reduction degree reached 96.55% with the homogeneous reduction method. The results of the on-site gas loss measurements show that the proportions of most hydrocarbon substances in each lost gas increase to varying degrees after using the homogeneous reduction technology. Therefore, it is proved that these components can be better preserved using the homogeneous reduction method. It can effectively avoid the condensation of components, which is of great significance to the study of oil and gas loss.
... Ma, Zhang [10] developed a multiscale parameterization method on fractured reservoir inversion based on data-driven EA. Zhang, Zhang [52] integrated deep sparse auto-encoder and ES with multiple data assimilation for naturally fractured reservoirs. Ringel, Jalali [11] introduced a flexible three-dimensional DFN inversion structure by hydraulic tomography with MCMC method. ...
... To estimate the parameters of the fracture network, parameterization technique is a necessary step to characterize the complex fracture field effectively. Existing parameterization methods can be classified into transform-based methods [56], equivalent methods [57], geometrical methods of discrete fractures [52] and fractal methods [8]. The geometry of a 2D fracture can be parameterized in two different ways, namely coordinate parameterization and angle/radius parameterization ( Fig. 1), with the former to be easier to determine and has less constraints [58]. ...
The characterization of fracture networks is challenging for enhanced geothermal systems, yet is crucial for the understanding of the thermal distributions, and the behaviors of flow field and solute transport. A novel inverse modeling framework is proposed for the estimation of the fracture networks. The hierarchical parameterization method is adopted in this work. For a small number of large fractures, each fracture is characterized by fracture length, azimuth and coordination of the fracture center. For dense small fractures, fracture density and fractal dimension are utilized to characterize the fracture networks. Moreover, we adopt variational auto-encoder and generative adversarial network (VAE-GAN) and fuse the GAN objective with prior constraint information to capture the distribution of the parameters of complex fracture networks and to satisfy the prior knowledge of
fracture fields, thereby mapping the high-dimensional complex parameter distribution into low-dimensional continuous parameter field. Afterwards, relying on the Bayesian framework, ensemble smoother is adopted based on the collected data from hydraulic tomography to reduce the uncertainty of the fracture distribution.
Two numerical cases with different complexity are used to test the performance of the proposed framework. The results show that the proposed algorithm can estimate effectively the distribution of the fracture fields.
... These generative models can transform non-Gaussian geological parameters into low-dimensional latent variables with the Gaussian distribution. These deep-learning-based parameterization methods can be combined with, e.g., ES and ESMDA to obtain reasonable posterior estimates of the geological models with complex characteristics (Chang et al., 2017;Laloy et al., 2017Laloy et al., , 2018Ma et al., 2021;Mo et al., 2020;Zhang et al., 2021). However, these parameterization methods often require a large number of prior geological models (tens of thousands) as samples for neural network training. ...
Inverse modeling can provide a reliable geological model for subsurface flow numerical simulation, which is a challenging issue that requires calibration of the uncertain parameters of the geological model to establish an acceptable match between simulation data and observation data. The general inverse modeling method needs to iteratively adjust the uncertain parameters, which is a difficult and time‐consuming high‐dimensional sampling problem. To address this problem, we propose a deep‐learning‐based inverse modeling method called pix2pixGAN‐DSI. In this method, the deep‐learning‐based image‐to‐image generative adversarial network (pix2pixGAN) is constructed to directly predict the posterior parameter fields from the posterior dynamic responses obtained by the data‐space inversion (DSI) method. This inverse modeling method does not need to iteratively adjust the uncertain parameters, which improves computational efficiency. The effectiveness of the proposed method is verified through a Gaussian model case and two non‐Gaussian channelized model cases. Through the analysis of posterior realizations, matching and forecast of production data, and uncertainty quantification, the results show that the proposed method can obtain reasonable estimates without iteration and parameterization.
... We show the loss of both training data and validating data in the training mode and further analyze the contribution of inner modules embedded in the network based on the loss curve [42][43][44]. When it comes to application, we do a prediction of saturation and pressure field in future 9 timesteps from a random reservoir state and show the trend of the error between the predicted field and real ones. ...
The well pattern and boundary shape of reservoirs determine the distribution of the remaining oil distribution to a large extent, especially for small-scale reservoir blocks. However, it is difficult to replicate experiences from other reservoirs directly to predict the remaining oil distribution because of the variety of irregular boundary shapes and corresponding well patterns. Meanwhile, the regular well pattern can hardly suit irregular boundary shapes. In this paper, we propose a well placement method for undeveloped irregular reservoirs and a multi-step prediction framework to predict both oil saturation and pressure fields for any reservoir shape and well pattern. To boost the physical information of input characteristics, a feature amplification approach based on physical formulae is initially presented. Then, 3D convolution technology is employed for the first time in 3D reservoir prediction to increase the spatial information in the vertical direction of the reservoir in the input. Moreover, to complete the two-field prediction, the concept of multi-task learning is adopted for the first time, improving the rationality of the forecast. Through the loss-based ablation test, we found that the operation we adopt will increase the accuracy of prediction to some extent. By testing on both manually designed and real irregular-shape reservoirs, our method is proven to be an accurate and fast oil saturation prediction method with its prediction loss less than 0.01 and calculation time less than 10 s in the future one year.
... At the final stage, the SAE model is employed for FER. During this case, it can execute the hyperbolic tangent activation function to encode and decode [21]. Primarily, the bias vector and weight matrix were allocated arbitrary values. ...
Facial expression recognition (FER) remains a hot research area among computer vision researchers and still becomes a challenge because of high intra-class variations. Conventional techniques for this problem depend on hand-crafted features, namely, LBP, SIFT, and HOG, along with that a classifier trained on a database of videos or images. Many execute perform well on image datasets captured in a controlled condition; however not perform well in the more challenging dataset, which has partial faces and image variation. Recently, many studies presented an endwise structure for facial expression recognition by utilizing DL methods. Therefore, this study develops an earthworm optimization with an improved SqueezeNet-based FER (EWOISN-FER) model. The presented EWOISN-FER model primarily applies the contrast-limited adaptive histogram equalization (CLAHE) technique as a pre-processing step. In addition, the improved SqueezeNet model is exploited to derive an optimal set of feature vectors , and the hyperparameter tuning process is performed by the stochastic gradient boosting (SGB) model. Finally, EWO with sparse autoencoder (SAE) is employed for the FER process, and the EWO algorithm appropriately chooses the SAE parameters. A wide-ranging experimental analysis is carried out to examine the performance of the proposed model. The experimental outcomes indicate the supremacy of the presented EWOISN-FER technique. 1 Introduction Facial expression recognition (FER) plays a pivotal role in the artificial intelligence (AI) period. Following the emotional information of humans, machines could offer personalized services. Several applications, namely customer satisfaction, virtual reality, personalized recommendations, and many more, rely upon an effective and dependable way of recognizing facial expressions [1]. This topic has grabbed the attention of several research scholars for years. Still, it becomes a challenging topic as expression features change significantly with environments, head poses, and discrepancies in the different individuals involved [2]. Technologies for transmission have conventionally been formulated based on the senses that serve as the main part of human communication. Specifically, AI voice recognition technology utilizing AI speakers and sense of hearing was commercialized due to the enhancements in AI technology [3]. By using this technology that identifies voice and language, there are AI robots that could communicate thoroughly with real life for handling the daily lists of individuals. But sensory acceptance was needed to communicate more accurately [4]. Thus, the most essential technology was a vision sensor, since vision becomes a large part of human perception in many communications.
... Ma, Zhang [8] developed a multiscale parameterization method on fractured reservoir inversion based on data-driven EA. Zhang, Zhang [51] integrated deep sparse auto-encoder and ES with multiple data assimilation for naturally fractured reservoirs. Ringel, Jalali [9] introduced a flexible three-dimensional DFN inversion structure by hydraulic tomography with MCMC method. ...
... To estimate the parameters of the fracture network, parameterization technique is a necessary step to characterize the complex fracture field effectively. Existing parameterization methods can be classified into transform-based methods [55], equivalent methods [56], geometrical methods of discrete fractures [51] and fractal methods [6]. The geometry of a 2D fracture can be parameterized in two different ways, namely coordinate parameterization and angle/radius parameterization ( Figure 1), with the former to be easier to determine and has less constraints [57]. ...
The distribution of fracture network is crucial to characterize the behaviors of flow field and solute transport, especially for enhanced geothermal systems, as fractures provide preferential flow paths. However, estimating the parameters of the fracture networks and quantifying their uncertainties based on observing data is a nontrivial task because inverse modeling of fractured model is strongly nonlinear and non-Gaussian distributed. To address this issue, a novel inverse modeling framework is proposed for the estimation of the fracture networks. The hierarchical parameterization method is adopted in this work. For a small number of large fractures, each fracture is characterized by fracture length, azimuth and coordination of the fracture center. For dense small fractures, fracture density and fractal dimension are utilized to characterize the fracture networks. Moreover, we adopt variational auto-encoder and generative adversarial network (VAE-GAN) and fuse the GAN objective with prior constraint information to capture the distribution of the parameters of complex fracture networks and to satisfy the prior knowledge of fracture fields, thereby mapping the high-dimensional complex parameter distribution into low-dimensional continuous parameter field. Afterwards, relying on the Bayesian framework, ensemble smoother is adopted based on the collected data from hydraulic tomography to reduce the uncertainty of the fracture distribution. Two numerical cases with different complexity are used to test the performance of the proposed framework. The results show that the proposed algorithm can estimate effectively the distribution of the fracture fields after providing sufficient prior constraint information and hydraulic head measurements.
... In the early stages of COVID-19, regions that border the pulmonary arteries can be seen to have a ground glass appearance; however, this kind of symptom is difficult to diagnose visually (K. Zhang et al., 2021). Clinical studies also report diffuse airspace or asymmetric patchy opacities as other frequent COVID-19 symptoms . ...
Applying Deep Learning (DL) in radiological images (i.e., chest X-Rays) is emerging because of the necessity of having accurate and fast COVID-19 detectors. Deep Convolutional Neural Networks (DCNN) have been typically used as robust COVID-19 positive case detectors in these approaches. Such DCCNs tend to utilize Gradient Descent-Based (GDB) algorithms as the last fully-connected layers’ trainers. Although GDB training algorithms have simple structures and fast convergence rates for cases with large training samples, they suffer from the manual tuning of numerous parameters, getting stuck in local minima, large training samples set requirements, and inherently sequential procedures. It is exceedingly challenging to parallelize them with Graphics Processing Units (GPU). Consequently, the Chimp Optimization Algorithm (ChOA) is presented for training the DCNN's fully connected layers in light of the scarcity of a big COVID-19 training dataset and for the purpose of developing a fast COVID-19 detector with the capability of parallel implementation. Following that, two publicly accessible datasets termed COVID-Xray-5k and COVIDetectioNet are used to benchmark the proposed detector known as DCCN-Chimp. In order to make a fair comparison, two structures are proposed: i-6c-2s-12c-2s and i-8c-2s-16c-2s, all of which have had their hyperparameters fine-tuned. The outcomes are evaluated in comparison to standard DCNN, Hybrid DCNN plus Genetic Algorithm (DCNN-GA), and Matched Subspace classifier with Adaptive Dictionaries (MSAD). Due to the large variation in results, we employ a weighted average of the ensemble of ten trained DCNN-ChOA, with the validation accuracy of the weights being used to determine the final weights. The validation accuracy for the mixed ensemble DCNN-ChOA is 99.11%. LeNet-5 DCNN's ensemble detection accuracy on COVID-19 is 84.58%. Comparatively, the suggested DCNN-ChOA yields over 99.11% accurate detection with a false alarm rate of less than 0.89 %. The outcomes show that the DCCN-Chimp can deliver noticeably superior results than the comparable detectors. The Class Activation Map (CAM) is another tool used in this study to identify probable COVID-19-infected areas. Results show that highlighted regions are completely connected with clinical outcomes, which has been verified by experts.
