Search

Modern production relies on data-based analytics for the prediction and optimization of production processes. Specialized data scientists perform tasks at companies and research institutions, dealing with real data from actual production environments. The roles of data preprocessing and data quality are crucial in data science, and an active research field deals with methodologies and technologies for this. While anecdotes and generalized surveys indicate preprocessing is the major operational task for data scientists, a detailed view of the subtasks and the domain of production data is missing. In this paper, we present a multi-stage survey on data science tasks in practice in the field of production. Using expert knowledge and insights, we found data preprocessing to be the major part of the tasks of data scientists. In detail, we found that tackling missing values, finding data point meanings, and synchronization of multiple time-series were often the most time-consuming preprocessing tasks.

Effective identification of anomalous data from production time series in the oilfield affects future analysis and forecasting. Such time series is often characterized by irregular time intervals due to uneven manual sampling, and missing values caused by incomplete measurements. Therefore, the identification task becomes more challenging. In this paper, an Attention-Embedded Time-Aware Imputation Network (ATIN) with two sub-networks is proposed for this task. First, Time-Aware Imputation LSTM (TI-LSTM) is designed for modeling irregular time intervals and incomplete measurements. It decays the long-term memory component as the producing well conditions may be varied during the water cut stage. Second, Attention-Embedding LSTM (ATEM) is designed to improve the effectiveness of anomaly detection. It focuses on the correlation between the last and historical measurements in a given sequence. Comparison experiments with several state-of-the-art methods, including mTAN, GRU-D, T-LSTM, ATTAIN, and BRITS are conducted. Results show that the proposed ATIN performs better in accuracy, F 1 -score, and area under curve (AUC). [ABSTRACT FROM AUTHOR]

Modern production relies on data-based analytics for the prediction and optimization of production processes. Specialized data scientists perform tasks at companies and research institutions, dealing with real data from actual production environments. The roles of data preprocessing and data quality are crucial in data science, and an active research field deals with methodologies and technologies for this. While anecdotes and generalized surveys indicate preprocessing is the major operational task for data scientists, a detailed view of the subtasks and the domain of production data is missing. In this paper, we present a multi-stage survey on data science tasks in practice in the field of production. Using expert knowledge and insights, we found data preprocessing to be the major part of the tasks of data scientists. In detail, we found that tackling missing values, finding data point meanings, and synchronization of multiple time-series were often the most time-consuming preprocessing tasks. [ABSTRACT FROM AUTHOR]

Abstract Natural fractures play an essential role in the characterization and modeling of hydrocarbon reservoirs. Modeling fractured reservoirs requires an understanding of fracture characteristics. Fractured zones can be detected by using seismic data, petrophysical logs, well tests, drilling mud loss history and core description. In this study, the feed-forward neural networks (FFNN), cascade feed forward neural networks (CFFN) and random forests (RF) were used to determine fracture density from petrophysical logs. The model performance was assessed using statistical measures including the root mean squared error (RMSE), coefficient of determination (R 2), mean absolute error (MAE), Kling Gupta efficiency (KGE) and Willmott’s index (WI). Conventional good logs and full-bore micro-resistivity imaging data were available from three drilled wells of the Mozduran reservoir, Khangiran gas field. According to the findings of this research, the FFNN model showed a higher KGE and WI, and a higher correlation coefficient (R 2) compared to the CFNN model. The CFNN model outperformed the FFNN model with lower neurons. The models' performance was also improved by increasing the number of neurons in the hidden layers from 8 to 35. The findings of this study demonstrate that the measured and FFNN calculated fracture intensity is in excellent agreement with image log results showing a correlation coefficient of 92%. The RF algorithm showed higher stability and robustness in predicting fracture intensity with a correlation coefficient of 93%. The results of this study can successfully be used as an aid in a more successful reservoir dynamic modeling and production data analysis.

Based on the fundamental processes and principles of ironmaking in a BF, an analysis was conducted on the production technical data from 18 large-scale 3200 m3 BFs in China in 2020. The study systematically investigated the influence of top pressure on reduction, air injection, furnace top, energy consumption, and pig iron composition within the BF. Practical operations in a newly No. 1 3200 m3 BF in a Chinese company demonstrated that increasing the top pressure not only effectively improved the smelting process, operational stability, and reduced furnace fluctuations but also led to a decrease in dust emissions, increased pig iron yield and quality, and significantly reduced fuel consumption and process energy consumption. Against the backdrop of dual carbon objectives, gaining a scientific understanding of the dynamic operation principles of the BF smelting process and continuously exploring low-carbon ironmaking methods have become the focus for professionals and researchers in the fields of BF ironmaking, production management, technology, and scientific research. [ABSTRACT FROM AUTHOR]

Natural fractures play an essential role in the characterization and modeling of hydrocarbon reservoirs. Modeling fractured reservoirs requires an understanding of fracture characteristics. Fractured zones can be detected by using seismic data, petrophysical logs, well tests, drilling mud loss history and core description. In this study, the feed-forward neural networks (FFNN), cascade feed forward neural networks (CFFN) and random forests (RF) were used to determine fracture density from petrophysical logs. The model performance was assessed using statistical measures including the root mean squared error (RMSE), coefficient of determination (R2), mean absolute error (MAE), Kling Gupta efficiency (KGE) and Willmott's index (WI). Conventional good logs and full-bore micro-resistivity imaging data were available from three drilled wells of the Mozduran reservoir, Khangiran gas field. According to the findings of this research, the FFNN model showed a higher KGE and WI, and a higher correlation coefficient (R2) compared to the CFNN model. The CFNN model outperformed the FFNN model with lower neurons. The models' performance was also improved by increasing the number of neurons in the hidden layers from 8 to 35. The findings of this study demonstrate that the measured and FFNN calculated fracture intensity is in excellent agreement with image log results showing a correlation coefficient of 92%. The RF algorithm showed higher stability and robustness in predicting fracture intensity with a correlation coefficient of 93%. The results of this study can successfully be used as an aid in a more successful reservoir dynamic modeling and production data analysis. [ABSTRACT FROM AUTHOR]

Abstract Enhancing oil recovery in reservoirs with light oil and high gas content relies on optimizing the miscible water alternating gas (WAG) injection profile. However, this can be costly and time-consuming due to computationally demanding compositional simulation models and numerous other well control variables. This study introduces WAGeq, a novel approach that expedites the convergence of the optimization algorithm for miscible water alternating gas (WAG) injection in carbonate reservoirs. The WAGeq leverages production data to create flexible solutions that maximize the net present value (NPV) of the field, while providing practical implementation of individual WAG profiles for each injector. The WAGeq utilizes an injection priority index to rank the wells and determine which should inject water or gas at each time interval. The index is built using a parametric equation that considers factors such as producer and injector relationship, water cut (W CUT), gas–oil ratio (GOR), and wells cumulative gas production, to induce desirable effects on production and WAG profile. To evaluate WAGeq’s effectiveness, two other approaches were compared: a benchmark solution named WAGbm, in which the injected fluid is optimized for each well over time, and a traditional baseline strategy with fixed 6-month WAG cycles. The procedures were applied to a synthetic simulation case (SEC1_2022) with characteristics of a Brazilian pre-salt carbonate field with karstic formations and high CO2 content. The WAGeq outperformed the baseline procedure, improving the NPV by 6.7% or 511 USD million. Moreover, WAGeq required fewer simulations (less than 350) than WAGbm (up to 2000), while delivering a slightly higher NPV. The terms of the equation were also found to be essential for producing a WAG profile with regular patterns on each injector, resulting in a more practical solution. In conclusion, WAGeq significantly reduces computational requirements while creating consistent patterns across injectors, which are crucial factors to consider when planning a practical WAG strategy.

Enhancing oil recovery in reservoirs with light oil and high gas content relies on optimizing the miscible water alternating gas (WAG) injection profile. However, this can be costly and time-consuming due to computationally demanding compositional simulation models and numerous other well control variables. This study introduces WAGeq, a novel approach that expedites the convergence of the optimization algorithm for miscible water alternating gas (WAG) injection in carbonate reservoirs. The WAGeq leverages production data to create flexible solutions that maximize the net present value (NPV) of the field, while providing practical implementation of individual WAG profiles for each injector. The WAGeq utilizes an injection priority index to rank the wells and determine which should inject water or gas at each time interval. The index is built using a parametric equation that considers factors such as producer and injector relationship, water cut (WCUT), gas–oil ratio (GOR), and wells cumulative gas production, to induce desirable effects on production and WAG profile. To evaluate WAGeq's effectiveness, two other approaches were compared: a benchmark solution named WAGbm, in which the injected fluid is optimized for each well over time, and a traditional baseline strategy with fixed 6-month WAG cycles. The procedures were applied to a synthetic simulation case (SEC1_2022) with characteristics of a Brazilian pre-salt carbonate field with karstic formations and high CO2 content. The WAGeq outperformed the baseline procedure, improving the NPV by 6.7% or 511 USD million. Moreover, WAGeq required fewer simulations (less than 350) than WAGbm (up to 2000), while delivering a slightly higher NPV. The terms of the equation were also found to be essential for producing a WAG profile with regular patterns on each injector, resulting in a more practical solution. In conclusion, WAGeq significantly reduces computational requirements while creating consistent patterns across injectors, which are crucial factors to consider when planning a practical WAG strategy. [ABSTRACT FROM AUTHOR]

This study explores using the speech of Najdi social media influencers in Snapchat and TikTok as a source of sociolinguistic production data. Saudis in general, Najdis in particular, post their daily life vlogs on these apps building a huge volume of recorded speech in the local dialects. This type of content has been examined to introduce the possibility of using it as a source of production data alongside/instead of the sociolinguistic interview. The speech of 18 social media influencers, representing three sub-dialects of Najd, has been transcribed and analyzed to test whether it is natural and not mostly formal. The variables have been ranked according to Labov's (1972) classification of speakers' awareness: stereotype, marker and indicator. It was found in this study that, similar to data elicited from the interview, social media influencers tend to use their local variants which are classified as markers and/or indicators, and they avoid using stereotypes. In some social settings, the social media influencers were even more spontaneous than they would be in an interview. The study suggests that the speech of social media influencers in Najd is an easily accessible, larger, and better-quality source of sociolinguistic production data. [ABSTRACT FROM AUTHOR]

This article presents a machine learning (ML) application to examine artificial lift (AL) selection, using only field production datasets from a Sudanese oil field. Five ML algorithms were used to develop a selection model, and the results demonstrated the ML capabilities in the optimum selection, with accuracy reaching 93%. Moreover, the predicted AL has a better production performance than the actual ones in the field. The research shows the significant production parameters to consider in AL type and size selection. The top six critical factors affecting AL selection are gas, cumulatively produced fluid, wellhead pressure, GOR, produced water, and the implemented EOR. This article contributes significantly to the literature and proposes a new and efficient approach to selecting the optimum AL to maximize oil production and profitability, reducing the analysis time and production losses associated with inconsistency in selection and frequent AL replacement. This study offers a universal model that can be applied to any oil field with different parameters and lifting methods. [ABSTRACT FROM AUTHOR]

The paper focuses on the use of augmented reality (AR) by industrial engineers, especially for determining the necessary competencies required for its use. Industrial engineers are not inherently programmers. Nevertheless, augmented reality is a modern trend in their field, especially in relation to the concept of Industry 4.0 and industry in general, where it has a higher potential than virtual reality. In the first part of this paper, we placed augmented reality and the competencies required for its use in the context of industrial engineering. Subsequently, we described our own methods of implementing an augmented reality industrial metadata visualization system, namely Help Lightning Fieldbit and Unity 3D, using the Vuforia extension. We chose the metadata used in the methods with regard to their environmental potential. In this part of the paper, we also described the chosen and applied testing methodology using a questionnaire survey. Subsequently, we described the results from the questionnaire surveys of both these approaches of implementing augmented reality methods. Finally, we evaluated the results and compared them with each other and with results from other authors. As the results show, the most important competencies for creating the described AR environments are analytical competencies. We draw conclusions from the collected data regarding the necessary competencies for the creation of AR scenes using these methods and their deployment in industry, including an outline for further research. [ABSTRACT FROM AUTHOR]

Carbonate reservoirs usually have a strong heterogeneity. The zones with relatively high permeability will form a channel through which fluids can easily flow. These channels are called thief zones. Thief zones have notable effects on oil or gas production, for example, high oil recovery rates at the early stage of the exploitation or an early water breakthrough during the later stage of water flooding development. Therefore, it is essential to have a precise identification of thief zones in carbonate reservoirs. In this research study, a simple approach to identify thief zones based on reservoir permeability gathered from well logging is developed. The thief zones are first identified at wells based on the lower limit value of the thief-zone permeability. This value is determined based on the dynamic production data, indicating that the thief zones identified by applying this criterion can reflect the product characteristics. Then, a zonal inter-well recognition method is adopted to identify the connectivity and distributions of thief zones in the regions far away from the well. This method is applied to identify thief zones for the Cretaceous Mishrif Formation in the H oilfield, Iraq. The reliability of the identification results is tested by the well-group injection test. The distributions of thief zones in the study region are discussed. In the study region, 12 members developed thief zones, while two members (i.e., MC1-3 and MC2-2) did not develop thief zones. Specifically, there are five members having a high level of thief-zone development. They are MB1-2C, MB2-1, MB2-2, MC2-3, and MC3-2. Comparing the distribution of thief zones with that of sedimentary microfacies, it is concluded that the thief-zone development is mainly controlled by the sedimentary microfacies and tends to occur in high-energy shoals.

In production data, missing values commonly appear for several reasons including changes in measurement and inspection items, sampling inspections, and unexpected process events. When applied to product failure prediction, the incompleteness of data should be properly addressed to avoid performance degradation in prediction models. Well-known approaches for missing data treatment, such as elimination and imputation, would not perform well under usual scenarios in production data, including high missing rate, systematic missing and class imbalance. To address these limitations, here we present a method for predictive modelling with missing data by considering the characteristics of production data. It builds multiple prediction models on different complete data subsets derived from the original data-set, each of which has different coverage of instances and input variables. These models are selectively used to make predictions for new instances with missing values. We demonstrate the effectiveness of the proposed method through a case study using actual data-sets from a home appliance manufacturer. [ABSTRACT FROM AUTHOR]

This study doesn't have clinical significance, although the company presents 31 cases of mesothelioma due to occupational past asbestos exposure. One case with environmental and another one with familial exposure. The exposures showed in a job/exposure/time matrix will allow an estimate of the dose response in the cohort. Objective: The aim of this study is the reconstruction of asbestos exposure in absence of environmental monitoring data, to use the results in a cohort study of railway rolling stock workers. Methods: To reconstruct past exposures, the production data (number of rolling stock and asbestos content) and working methods were reconstructed with former employees and company documentation, literature data, and author expertise. Results: The result of the work is a job/exposure matrix from 1956 to 1979, when sprayed asbestos was used as insulator of the metal bodies. Annual exposure estimate varies from 0.08 to 4.9 fb/mL depending on the specific jobs. Thirty-one mesotheliomas with occupational exposure, one with environmental and one with family exposures, were identified. Conclusions: The originality of the study consists on the use of company production data to establish frequency duration of asbestos exposure. [ABSTRACT FROM AUTHOR]

This article presents a machine learning (ML) application to examine artificial lift (AL) selection, using only field production datasets from a Sudanese oil field. Five ML algorithms were used to develop a selection model, and the results demonstrated the ML capabilities in the optimum selection, with accuracy reaching 93%. Moreover, the predicted AL has a better production performance than the actual ones in the field. The research shows the significant production parameters to consider in AL type and size selection. The top six critical factors affecting AL selection are gas, cumulatively produced fluid, wellhead pressure, GOR, produced water, and the implemented EOR. This article contributes significantly to the literature and proposes a new and efficient approach to selecting the optimum AL to maximize oil production and profitability, reducing the analysis time and production losses associated with inconsistency in selection and frequent AL replacement. This study offers a universal model that can be applied to any oil field with different parameters and lifting methods.

Industry 4.0 has fostered innovation in industries around the world. Manufacturing industries in particular are advancing towards smart manufacturing by integrating and applying relevant technologies. The output value of machine tools in Taiwan is among the top of the world and the central region is a key area for this industry chain, which supplies manufacturers in Taiwan and their international downstream customers. To support innovation in this industry, the current study used the Six Sigma quality indices for smaller-the-better, larger-the-better, and nominal-the-best quality characteristics to construct a fuzzy decision-making model. Based on this model, we propose a process quality fuzzy analysis chart (PQFAC) for process quality improvement. Our use of fuzzy decision values to replace lower confidence limits decreases the probability of misjudgment made by sampling errors. The proposed fuzzy model also offers a more accurate assessment of process improvement requirements. We provide a real-world example to demonstrate the applicability of the proposed approach. Machine tool manufacturers can apply the platform and proposed model to evaluate their process capabilities for the vital parts suppliers and downstream customers, determine optimal machine parameter settings for processes with inadequate accuracy or precision, establish more suitable machine repair and maintenance systems, and combine the improvement experiences of customers to create an improvement knowledge base. This will enhance product value and industry competitiveness for the entire machine tool industry chain. [ABSTRACT FROM AUTHOR]

Abstract Volume transformation technology has become a key technology for developing low-permeability/tight oil and gas reservoirs. Evaluating the post-fracturing effect is very important for the development plan formulation and fracturing plan evaluation. In this paper, the vicinity of the main fracture is divided into the main fracture zone and the secondary fracture zone. The main fracture with infinite conductivity and the branch fracture with increased permeability are used to describe the transformation area. Based on this physical model, a numerical model considering the nonlinear seepage characteristics of the reservoir, stress sensitivity, wellbore storage and skin effects was established. Based on this numerical model, a comprehensive evaluation method for the fracturing effect of volume modification of vertical wells based on well test data and production data was established and this method was applied to three typical vertical wells. The results show that conventional vertical fracturing vertical wells can only form a single primary fracture and the range of equivalent permeability increase is very small. Volume fracturing can form a fracture network composed of primary fractures and secondary fractures, and increase the equivalent permeability of the fracture network area. The fracture half-length, equivalent permeability and reconstruction area of the volume fracturing well are dynamically changing and gradually decrease with the increase in production time and the fracturing effect becomes weak until it fails.

In real-world production data, missing values often occur randomly or systematically with various missing patterns. Missing values need to be handled properly to build effective prediction models. This paper presents a novel method based on graph representation and graph neural networks for improving prediction in missing value conditions. To utilize the entire information of a training dataset without direct manipulation, all instances of the dataset are represented as graphs of varying sizes, in which nodes and edges represent the observed input variables and their pairwise relationships. Prediction models learn from the graph representations. These models can make predictions of unknown labels for new instances that have arbitrary missing patterns. The superiority of the proposed method was investigated on seven different product failure prediction tasks from a home appliance manufacturer. The proposed method outperformed all other methods in six of the seven tasks. [ABSTRACT FROM AUTHOR]

With the increasing requirements on machine tools such as higher complexity, individuality and digitization, their maintenance becomes also more complex, which increases time consumption and demands special knowledge. The maintenance engineer can be supported in this process using the Augmented Reality (AR) technology, mobile devices and suitable production data. This requires the integration and adaptation of the maintenance documentation as well as further production systems and their data such as the CAD system, product data management (PDM) system, manufacturing execution system (MES) and CNC. For this purpose, a method is developed that allows an efficient integration process of various production data into an AR supported maintenance documentation. This includes concepts for the integration process of different production systems and data into an AR maintenance system as well as the design of the AR maintenance system architecture. Basis is the analysis of the required production and AR systems, data and processes. The aim of the method is to reduce the high expenditure of the overall planning process and the necessary expertise of the maintenance planner of the AR technology. The developed method is verified with different maintenance scenarios, production systems and data as well as machine tools and AR devices. [ABSTRACT FROM AUTHOR]

Predictive Analytics is increasingly applied as a basis for decision-making in product and process optimization of manufacturing companies. In this context, Predictive Quality enables companies to make data-driven predictions of product quality, with data integration being one of the most significant challenges. This paper provides insights into exemplary applications of a flexible process-independent meta-model to integrate data for multi-step manufacturing processes. Three application use cases from different production domains are presented, demonstrating the meta-model's applicability for Predictive Quality applications in manufacturing companies without restrictions regarding the product or manufacturing process. [ABSTRACT FROM AUTHOR]