Application of artificial intelligence techniques in the petroleum industry: a review

A Comprehensive review of data-driven approaches for forecasting production from unconventional reservoirs: best practices and future directions

Prediction of well production from unconventional reservoirs is a complex problem given an incomplete understanding of physics despite large amounts of data. Recently, Data Analytics Techniques (DAT) have emerged as an effective approach for production forecasting for unconventional reservoirs. In some of these approaches, DAT are combined with physics-based models to capture the essential physical mechanisms of fluid flow in porous media, while leveraging the power of data-driven methods to account for uncertainties and heterogeneities. Here, we provide an overview of the applications and performance of DAT for production forecasting of unconventional reservoirs examining and comparing predictive models using different algorithms, validation benchmarks, input data, number of wells, and formation types. We also discuss the strengths and limitations of each model, as well as the challenges and opportunities for future research in this field. Our analysis shows that machine learning (ML) based models can achieve satisfactory performance in forecasting production from unconventional reservoirs. We measure the performance of the models using two dimensionless metrics: mean absolute percentage error (MAPE) and coefficient of determination (R2). The predicted and actual production data show a high degree of agreement, as most of the models have a low error rate and a strong correlation. Specifically, ~ 65% of the models have MAPE less than 20%, and more than 80% of the models have R2 higher than 0.6. Therefore, we expect that DAT can improve the reliability and robustness of production forecasting for unconventional resources. However, we also identify some areas for future improvement, such as developing new ML algorithms, combining DAT with physics-based models, and establishing multi-perspective approaches for comparing model performance.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10462-024-10865-5.

Applications of AI in oil and gas projects towards sustainable development: a systematic literature review

Oil and gas construction projects are critical for meeting global demand for fossil fuels, but they also present unique risks and challenges that require innovative construction approaches. Artificial Intelligence (AI) has emerged as a promising technology for tackling these challenges, and this study examines its applications for sustainable development in the oil and gas industry. Using a systematic literature review (SLR), this research evaluates research trends from 2011 to 2022. It provides a detailed analysis of how AI suits oil and gas construction. A total of 115 research articles were reviewed to identify original contributions, and the findings indicate a positive trend in AI research related to oil and gas construction projects, especially after 2016. The originality of this study lies in its comprehensive analysis of the latest research on AI applications in the oil and gas industry and its contribution to developing recommendations for improving the sustainability of oil and gas projects. This research's originality is in providing insight into the most promising AI applications and methodologies that can help drive sustainable development in the oil and gas industry.

A Data-Driven Intelligent System for Assistive Design of Interior Environments

This paper analyses the design of a healthy interior environment using big data intelligence. The application of big data intelligence in the design of healthy interior environments is necessary because the traditional interior design approaches consume a lot of energy and other problems. Benefited by its strong ability of computation and analytics, artificial intelligence can well improve a series of problems in the field of interior design. The proposal summarizes the sources, classifications, and expressions of behavioral data in interior spaces, carries out analysis and research on behavioral data from two aspects: display space and supermarket space, summarizes the interior methods based on behavioral data, and analyses the visualization application of behavioral data in different interior scenes, to explore the application value of behavioral data in interior design. In contrast to it is the unconscious behavioral response, the biggest characteristic of which is that it is regulated by the behavioral subject's physiological factors or habits of the behavior issuer. In this paper, we convert the layout recommendation problem of a space into a functional classification problem of segmented segments and household segments on a plane. The scene layout features are extracted by binary coding, the abstraction of the cross features between the vector segments is achieved by using a word embedding algorithm, the feature matrix is reduced in dimensionality, and finally, the segmentation network model and the layout network model are constructed, respectively, by using a bidirectional LSTM. The experiments show that the accuracy of the layout recommendation model in this paper is 98%, which can meet the demand for real-time online layouts.

Adsorption of benzene on soils under different influential factors: an experimental investigation, importance order and prediction using artificial neural network

The adsorption of benzene on soils is specifically associated with its migration and transformation. Although previous studies have proved that the adsorption of benzene is affected by various factors, studies simultaneously considering the effects of multiple factors are rare. This study aimed to identify the qualitative and quantitative relationships between multiple influential factors and the adsorption capacity of benzene (BC). Batch adsorption experiments considering different influential factors, including initial concentration (IC), pH, temperature (T), ion strength (IS) and organic matter content (OMC), were conducted in three kinds of soils collected in a chemical industry park. The correlation analysis between different influential factors and BC was carried out based on the experimental data. The artificial neural network (ANN) was applied to predict BC. The results showed that BC increased with the increase of T. As the pH increased, BCs on silty loam and loam increased, while that on sandy loam decreased. Besides, BCs on silty loam and loam raised with increasing OMC, while that on sandy loam remained unchanged. BCs on all three kinds of soils attained their peaks when IS was small and then become stable with an increase in IS. The sequence of correlation between BC and influential factors is listed as IC > OMC > T > IS > pH for silty loam, OMC > IC > T > IS > pH for loam and IC > T > IS > pH > OMC for sandy loam. ANN analysis showed satisfactory accuracy in predicting BC under different influential factors. These results help us understand the important factors affecting benzene adsorption and provide a tool to get the adsorption information easily in complex site conditions.

Residual Strength Assessment and Residual Life Prediction of Corroded Pipelines: A Decade Review

Prediction of residual strength and residual life of corrosion pipelines is the key to ensuring pipeline safety. Accurate assessment and prediction make it possible to prevent unnecessary accidents and casualties, and avoid the waste of resources caused by the large-scale replacement of pipelines. However, due to many factors affecting pipeline corrosion, it is difficult to achieve accurate predictions. This paper reviews the research on residual strength and residual life of pipelines in the past decade. Through careful reading, this paper compared several traditional evaluation methods horizontally, extracted 71 intelligent models, discussed the publishing time, the evaluation accuracy of traditional models, and the prediction accuracy of intelligent models, input variables, and output value. This paper’s main contributions and findings are as follows: (1) Comparing several traditional evaluation methods, PCORRC and DNV-RP-F101 perform well in evaluating low-strength pipelines, and DNV-RP-F101 has a better performance in evaluating medium–high strength pipelines. (2) In intelligent models, the most frequently used error indicators are mean square error, goodness of fit, mean absolute percentage error, root mean square error, and mean absolute error. Among them, mean absolute percentage error was in the range of 0.0123–0.1499. Goodness of fit was in the range of 0.619–0.999. (3) The size of the data set of different models and the data division ratio was counted. The proportion of the test data set was between 0.015 and 0.4. (4) The input variables and output value of predictions were summarized.

Artificial neural networks applied for predicting and explaining the education level of Twitter users

This paper provides a novel procedure to estimate the education level of social network (SN) users by leveraging artificial neural networks (ANN). Additionally, it provides a robust methodology to extract explanatory insights from ANN models. It also contributes to the study of socio-demographic phenomena by utilizing less explored data sources, such as social media. It proposes Twitter data as an alternative data source for in-depth social studies, and ANN for complex patterns recognition. Moreover, cutting edge technology, such as face recognition, on social media data are applied to explain the social characteristics of country-specific users. We use nine variables and three hidden layers of neurons to identify high-skilled users. The resulted model describes well the level of education by correctly estimating it with an accuracy of 95% on the training set and an accuracy of 92% on a testing set. Approximately 30% of the analyzed users are highly skilled and this share does not differ among the two genders. However, it tends to be lower among users younger than 30 years old.

Sustainability of operations through disruptive technologies in the petroleum supply chain

Purpose

Disruptive technologies can significantly contribute to the sustainability of operations in the petroleum supply chain. The present study aims to identify the prime sustainable dimensions and disruptive technologies implementation in the supply chain of the petroleum industry. The authors used content analysis in the literature and experts input to explore the sustainable dimensions and disruptive technologies in the supply chain.

Design/methodology/approach

This study used a hybrid method of hesitant fuzzy set and regret theory to identify the prominent sustainability dimensions and prominent disruptive technologies. This method emphasizes the decision-makers psychological characteristics under uncertain environments.

Findings

The result indicates that social responsibility, labor practices, safety and technical standards hold the most prominent sustainable dimensions in the petroleum supply chain. Further, the result also depicts that when consider an equal degree of regret and rejoice, artificial intelligence and big data could significantly enhance operations sustainability in the petroleum industry.

Research limitations/implications

This study considers only 11 sustainable dimensions and 43 sustainable factors, whereas other dimensions and factors could also be considered in future research. The research uses hesitant fussy set and regret set theory to identify the prominent sustainable dimensions and disruptive technologies, whereas other multiple-criteria decision-making (MCDM) techniques can also be used.

Originality/value

To the best of the authors’ knowledge, this is the first paper to explore the sustainable dimensions (environmental, social and economic) and disruptive technologies in the supply chain of the petroleum industry. This research intended to guide the practitioners, policymakers and academicians to emphasize their effort toward sustainable operations supply chain management.

Experimental investigation, binary modelling and artificial neural network prediction of surfactant adsorption for enhanced oil recovery application

Throughout the application of enhanced oil recovery (EOR), surfactant adsorption is considered the leading constraint on both the successful implementation and economic viability of the process. In this study, a comprehensive investigation on the adsorption behaviour of nonionic and anionic individual surfactants; namely, alkyl polyglucoside (APG) and alkyl ether carboxylate (AEC) was performed using static adsorption experiments, isotherm modelling using (Langmuir, Freundlich, Sips, and Temkin models), adsorption simulation using a state-of-the-art method, binary mixture prediction using the modified extended Langmuir (MEL) model, and artificial neural network (ANN) prediction. Static adsorption experiments revealed higher adsorption capacity of APG as compared to AEC, with sips being the most fitted model with R² (0.9915 and 0.9926, for APG and AEC respectively). It was indicated that both monolayer and multilayer adsorption took place in a heterogeneous adsorption system with non-uniform surfactant molecules distribution, which was in remarkable agreement with the simulation results. The (APG/AEC) binary mixture prediction depicted contradictory results to the experimental individual behaviour, showing that AEC had more affinity to adsorb in competition with APG for the adsorption sites on the rock surface. The adopted ANN model showed good agreement with the experimental data and the simulated adsorption values for APG and AEC showed a decreasing trend as temperature increases. Simulating the impact of binary surfactant adsorption can provide a tremendous advantage of demonstrating the binary system behaviour with less experimental data. The utilization of ANN for such prediction procedure can minimize the experimental time, operating cost and give feasible predictions compared to other computational methods. The integrated workflow followed in this study is quite innovative as it has not been employed before for surfactant adsorption studies.

A Deep-Learning-Based Oil-Well-Testing Stage Interpretation Model Integrating Multi-Feature Extraction Methods

The interpretation of well-testing data is a key means of decision-making support for oil and gas field development. However, conventional processing methods have many problems, such as the stochastic nature of the data, feature redundancies, the randomness of the initial weights or thresholds, and fluctuations in the generalization ability with slight changes in the network parameters. These result in a poor ability to characterize data features and a low generalization ability of the interpretation models. We propose a new integrated well-testing interpretation model based on a multi-feature extraction method and deep mutual information classifiers (MFE-DMIC). This model can avoid the low model classification accuracy caused by the simple training structures, lacking of redundancy elimination, and the non-optimal classifier configuration parameters. First, we obtained the initial features according to four classical feature extraction methods. Then, we eliminated feature redundancies using a deep belief network and united the maximum information coefficient method to achieve feature purification. Finally, we calculated the interpretation results using a hybrid particle swarm optimization–support vector machine classification system. We used 572 well-testing field samples, including five working stages, for model training and testing. The results show that the MFE-DMIC model had the highest total stage classification accuracy of 98.18% as well as the least of features (nine) compared with the classical feature extraction and classification methods and their combinations. The proposed model can reduce the efforts of oil analysts and allow accurate labeling of samples to be predicted.