Electrical Submersible Pumps reliability and run-life analysis has been extensively studied since its development. Current machine learning algorithms allow to correlate operational conditions to ESP run-life in order to generate predictions for active and new wells. Four machine learning models are compared to a linear proportional hazards model, used as a baseline for comparison purposes. Proper accuracy metrics for survival analysis problems are calculated on run-life predictions vs. actual values over training and validation data subsets. Results demonstrate that the baseline model is able to produce more consistent predictions with a slight reduction in its accuracy, compared to current machine learning models for small datasets. This study demonstrates that the quality of the date and it pre-processing supports the current shift from model-centric to data-centric approach to machine and deep learning problems.
{"title":"Model Comparison for Esp Run-Life Prediction: Classic Statistics Vs. Machine Learning","authors":"Alejandro Celemín, Diego Estupiñan, Ricardo Nieto","doi":"10.2118/206028-ms","DOIUrl":"https://doi.org/10.2118/206028-ms","url":null,"abstract":"\u0000 Electrical Submersible Pumps reliability and run-life analysis has been extensively studied since its development. Current machine learning algorithms allow to correlate operational conditions to ESP run-life in order to generate predictions for active and new wells. Four machine learning models are compared to a linear proportional hazards model, used as a baseline for comparison purposes. Proper accuracy metrics for survival analysis problems are calculated on run-life predictions vs. actual values over training and validation data subsets. Results demonstrate that the baseline model is able to produce more consistent predictions with a slight reduction in its accuracy, compared to current machine learning models for small datasets. This study demonstrates that the quality of the date and it pre-processing supports the current shift from model-centric to data-centric approach to machine and deep learning problems.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"53 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86308133","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. Pirrone, Satria Andrianata, S. Moriggi, G. Galli, S. Riva
Conventional downhole dynamic characterization is based on data from standard production logging tool (PLT) strings. Such method is not a feasible option in long horizontal drains, deep water scenarios, subsea clusters, pump-assisted wells and in presence of asphaltenes/solids deposition, mainly due to high costs and risk of tools stuck. In this respect, intrawell chemical tracers (ICT) can represent a valid and unobtrusive monitoring alternative. This paper deals with a new production allocation interpretation model of tracer concentration behavior that can overcome the limitation of standard PLT analyses in challenging environments. ICT are installed along the well completion and are characterized by a unique oil and/or water tracer signature at each selected production interval. Tracer concentration is obtained by dedicated analyses performed for each fluid sample taken at surface during transient production. Next, tracer concentration behavior over time is interpreted, for each producing interval, by means of an ad-hoc one-dimensional partial differential equation model with proper initial and boundary conditions, which describes tracer dispersion and advection profiles in such transient conditions. The full time-dependent analytical solutions are then utilized to obtain the final production allocation. The methodology has been developed and validated using data from a dozen of tracer campaigns. The approach is here presented through a selected case study, where a parallel acquisition of standard PLT and ICT data has been carried out in an offshore well. The aim was to understand if ICT could be used in substitution of the more impacting PLT for the future development wells in the field. At target, the well completion consists of a perforated production liner with tubing. The latter, which is slotted in front of the perforations, includes oil and water tracer systems. The straightforward PLT interpretation shows a clear dynamic well behavior with an oil production profile in line with the expectations from petrophysical information. Then, after a short shut-in period, the ICT-based production allocation has been performed in transient conditions with a very good match with the available outcomes from PLT: in fact, the maximum observed difference in the relative production rates is 5%. In addition, the full analytical solution of the ICT model has been fundamental to completely characterize some complex tracer concentration behaviors over time, corresponding to non-simultaneous activation of the different producing intervals. Given the consistency of the independent PLT and ICT interpretations, the monitoring campaign for the following years has been planned based on ICT only, with consequent impact on risk and cost mitigations. Although the added value of ICT is relatively well known, the successful description of the tracer signals through the full mathematical model is a novel topic and it can open the way for even more advanced applic
{"title":"Full Analytical Modeling Of Intrawell Chemical Tracer Concentration For Robust Production Allocation In Challenging Environments","authors":"M. Pirrone, Satria Andrianata, S. Moriggi, G. Galli, S. Riva","doi":"10.2118/206245-ms","DOIUrl":"https://doi.org/10.2118/206245-ms","url":null,"abstract":"\u0000 Conventional downhole dynamic characterization is based on data from standard production logging tool (PLT) strings. Such method is not a feasible option in long horizontal drains, deep water scenarios, subsea clusters, pump-assisted wells and in presence of asphaltenes/solids deposition, mainly due to high costs and risk of tools stuck. In this respect, intrawell chemical tracers (ICT) can represent a valid and unobtrusive monitoring alternative. This paper deals with a new production allocation interpretation model of tracer concentration behavior that can overcome the limitation of standard PLT analyses in challenging environments.\u0000 ICT are installed along the well completion and are characterized by a unique oil and/or water tracer signature at each selected production interval. Tracer concentration is obtained by dedicated analyses performed for each fluid sample taken at surface during transient production. Next, tracer concentration behavior over time is interpreted, for each producing interval, by means of an ad-hoc one-dimensional partial differential equation model with proper initial and boundary conditions, which describes tracer dispersion and advection profiles in such transient conditions. The full time-dependent analytical solutions are then utilized to obtain the final production allocation. The methodology has been developed and validated using data from a dozen of tracer campaigns.\u0000 The approach is here presented through a selected case study, where a parallel acquisition of standard PLT and ICT data has been carried out in an offshore well. The aim was to understand if ICT could be used in substitution of the more impacting PLT for the future development wells in the field. At target, the well completion consists of a perforated production liner with tubing. The latter, which is slotted in front of the perforations, includes oil and water tracer systems. The straightforward PLT interpretation shows a clear dynamic well behavior with an oil production profile in line with the expectations from petrophysical information. Then, after a short shut-in period, the ICT-based production allocation has been performed in transient conditions with a very good match with the available outcomes from PLT: in fact, the maximum observed difference in the relative production rates is 5%. In addition, the full analytical solution of the ICT model has been fundamental to completely characterize some complex tracer concentration behaviors over time, corresponding to non-simultaneous activation of the different producing intervals. Given the consistency of the independent PLT and ICT interpretations, the monitoring campaign for the following years has been planned based on ICT only, with consequent impact on risk and cost mitigations.\u0000 Although the added value of ICT is relatively well known, the successful description of the tracer signals through the full mathematical model is a novel topic and it can open the way for even more advanced applic","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"90323983","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Hydrocarbon production is commonly associated as the dispersed flow of two and more immiscible phases starting from porous media to surface facilities. In the dispersed flow, one phase is usually dispersed into another dominating phase in terms of droplets. Accurate prediction of the droplet size distribution of a dispersed phase is critical in characterizing complex flow behavior in pipe flows. In the first part of this paper, we provide the analyses of open-source experimental data on the maximum droplet size in gas-liquid annular flow and evaluate the existing theoretical models and suggest an improvement based on the experimental data analyses to predict the maximum droplet size of the entrained liquid droplets in gas-liquid annular flow. In the second part of this paper, we cover the experimental results from the open-source literature data and in-house experimental data to give the general understanding on droplet formation concepts and evaluate the existing predictive models and present a new modeling approach to determine a maximum stable droplet size of the dispersed phase in the liquid-liquid dispersed flow under turbulent flow conditions.
{"title":"Modeling Maximum Droplet Size In Gas-Liquid Annular Flow and Liquid–Liquid Dispersed Flow","authors":"Kanat Karatayev, Yilin Fan","doi":"10.2118/206081-ms","DOIUrl":"https://doi.org/10.2118/206081-ms","url":null,"abstract":"\u0000 Hydrocarbon production is commonly associated as the dispersed flow of two and more immiscible phases starting from porous media to surface facilities. In the dispersed flow, one phase is usually dispersed into another dominating phase in terms of droplets. Accurate prediction of the droplet size distribution of a dispersed phase is critical in characterizing complex flow behavior in pipe flows. In the first part of this paper, we provide the analyses of open-source experimental data on the maximum droplet size in gas-liquid annular flow and evaluate the existing theoretical models and suggest an improvement based on the experimental data analyses to predict the maximum droplet size of the entrained liquid droplets in gas-liquid annular flow. In the second part of this paper, we cover the experimental results from the open-source literature data and in-house experimental data to give the general understanding on droplet formation concepts and evaluate the existing predictive models and present a new modeling approach to determine a maximum stable droplet size of the dispersed phase in the liquid-liquid dispersed flow under turbulent flow conditions.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"61 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74959222","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. Meng, L. Frash, J. Carey, Wenfeng Li, N. Welch, Hongtao Zhang
Accurate characterization of oilwell cement mechanical properties is a prerequisite for maintaining long-term wellbore integrity. The drawback of the most widely used technique is unable to measure the mechanical property under in situ curing environment. We developed a high pressure and high temperature vessel that can hydrate cement under downhole conditions and directly measure its elastic modulus and Poisson's ratio at any interested time point without cooling or depressurization. The equipment has been validated by using water and a reasonable bulk modulus of 2.37 GPa was captured. Neat Class G cement was hydrated in this equipment for seven days under axial stress of 40 MPa, and an in situ measurement in the elastic range shows elastic modulus of 37.3 GPa and Poisson's ratio of 0.15. After that, the specimen was taken out from the vessel, and setted up in the triaxial compression platform. Under a similar confining pressure condition, elastic modulus was 23.6 GPa and Possion's ratio was 0.26. We also measured the properties of cement with the same batch of the slurry but cured under ambient conditions. The elastic modulus was 1.63 GPa, and Poisson's ratio was 0.085. Therefore, we found that the curing condition is significant to cement mechanical property, and the traditional cooling or depressurization method could provide mechanical properties that were quite different (50% difference) from the in situ measurement.
{"title":"Measurement of Cement in Situ Stresses and Mechanical Properties Without Cooling or Depressurization","authors":"M. Meng, L. Frash, J. Carey, Wenfeng Li, N. Welch, Hongtao Zhang","doi":"10.2118/206139-ms","DOIUrl":"https://doi.org/10.2118/206139-ms","url":null,"abstract":"\u0000 Accurate characterization of oilwell cement mechanical properties is a prerequisite for maintaining long-term wellbore integrity. The drawback of the most widely used technique is unable to measure the mechanical property under in situ curing environment. We developed a high pressure and high temperature vessel that can hydrate cement under downhole conditions and directly measure its elastic modulus and Poisson's ratio at any interested time point without cooling or depressurization. The equipment has been validated by using water and a reasonable bulk modulus of 2.37 GPa was captured. Neat Class G cement was hydrated in this equipment for seven days under axial stress of 40 MPa, and an in situ measurement in the elastic range shows elastic modulus of 37.3 GPa and Poisson's ratio of 0.15. After that, the specimen was taken out from the vessel, and setted up in the triaxial compression platform. Under a similar confining pressure condition, elastic modulus was 23.6 GPa and Possion's ratio was 0.26. We also measured the properties of cement with the same batch of the slurry but cured under ambient conditions. The elastic modulus was 1.63 GPa, and Poisson's ratio was 0.085. Therefore, we found that the curing condition is significant to cement mechanical property, and the traditional cooling or depressurization method could provide mechanical properties that were quite different (50% difference) from the in situ measurement.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"26 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74385285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Sidewall coring is a cost-effective process to complement conventional fullbore coring. Because sidewall cores target exact depth points, verification of the sidewall core recovery depth is required. We present an automated, fast workflow to perform the depth verification using borehole images, thereby providing consistent results. An application example using a typical dataset is used to showcase the workflow. A novel automated approach based on image analysis techniques and Bayesian statistical analysis is developed to verify sidewall core recovery depth using borehole image logs. A complete workflow is presented covering: 1) utilization of reference logs, e.g., gamma ray, to correct image log depth using cross correlation and/or dynamic time warping, 2) automated identification of sidewall core cavity in borehole image log using the circle Hough transform, and 3) estimation of confidence in the identification using Bayesian statistics and specialized metrics. The workflow is applied on a typical dataset containing tens of sidewall core cavities with varying quality. Results are comparable to the manual interpretation from an experienced engineer. A number of observations are made. First, the use of reference logs to correct the image log allows for determining the exact well logs values where the sidewall core was sampled, which is then compared to the initial target well logs values. This increases the confidence that the target lithofacies was sampled as planned. Second, the circle Hough Transform is suitable for this problem because it provides stable solutions for partially imaged sidewall core cavities typical in pad-based borehole images. Third, the use of Bayesian statistics and specialized metrics for the problem, such as average and standard deviation borehole image intensity in the cavity, provides customizability to work with multiple types of borehole images and with varying initial depth guess uncertainties. Overall, the use of fast and automated methodology for depth verification opens up avenues for near real-time combined sidewall coring, imaging, and verification workflows. The novelty in this study lies in using a combination of image processing techniques and statistical analysis to automate an established manual workflow. The automated workflow provides consistent results in minutes rather than hours. Results also incorporate a confidence index estimation.
{"title":"Automated Verification of Sidewall Core Recovery Depth using Borehole Image Logs","authors":"M. A. Ibrahim, V. Torlov, M. Mezghani","doi":"10.2118/206145-ms","DOIUrl":"https://doi.org/10.2118/206145-ms","url":null,"abstract":"\u0000 Sidewall coring is a cost-effective process to complement conventional fullbore coring. Because sidewall cores target exact depth points, verification of the sidewall core recovery depth is required. We present an automated, fast workflow to perform the depth verification using borehole images, thereby providing consistent results. An application example using a typical dataset is used to showcase the workflow. A novel automated approach based on image analysis techniques and Bayesian statistical analysis is developed to verify sidewall core recovery depth using borehole image logs. A complete workflow is presented covering: 1) utilization of reference logs, e.g., gamma ray, to correct image log depth using cross correlation and/or dynamic time warping, 2) automated identification of sidewall core cavity in borehole image log using the circle Hough transform, and 3) estimation of confidence in the identification using Bayesian statistics and specialized metrics. The workflow is applied on a typical dataset containing tens of sidewall core cavities with varying quality. Results are comparable to the manual interpretation from an experienced engineer. A number of observations are made. First, the use of reference logs to correct the image log allows for determining the exact well logs values where the sidewall core was sampled, which is then compared to the initial target well logs values. This increases the confidence that the target lithofacies was sampled as planned. Second, the circle Hough Transform is suitable for this problem because it provides stable solutions for partially imaged sidewall core cavities typical in pad-based borehole images. Third, the use of Bayesian statistics and specialized metrics for the problem, such as average and standard deviation borehole image intensity in the cavity, provides customizability to work with multiple types of borehole images and with varying initial depth guess uncertainties. Overall, the use of fast and automated methodology for depth verification opens up avenues for near real-time combined sidewall coring, imaging, and verification workflows. The novelty in this study lies in using a combination of image processing techniques and statistical analysis to automate an established manual workflow. The automated workflow provides consistent results in minutes rather than hours. Results also incorporate a confidence index estimation.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"31 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"74411570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Xupeng He, Weiwei Zhu, R. Santoso, M. AlSinan, H. Kwak, H. Hoteit
Geologic CO2 Sequestration (GCS) is a promising engineering technology to reduce global greenhouse emissions. Real-time forecasting of CO2 leakage rates is an essential aspect of large-scale GCS deployment. This work introduces a data-driven, physics-featuring surrogate model based on deep-learning technique for CO2 leakage rate forecasting. The workflow for the development of data-driven, physics-featuring surrogate model includes three steps: 1) Datasets Generation: We first identify uncertainty parameters that affect the objective of interests (i.e., CO2 leakage rates). For the identified uncertainty parameters, various realizations are then generated based on Latin Hypercube Sampling (LHS). High-fidelity simulations based on a two-phase black-oil solver within MRST are performed to generate the objective functions. Datasets including inputs (i.e., the uncertainty parameters) and outputs (CO2 leakage rates) are collected. 2) Surrogate Development: In this step, a time-series surrogate model using long short-term memory (LSTM) is constructed to map the nonlinear relationship between these uncertainty parameters as inputs and CO2 leakage rates as outputs. We perform Bayesian optimization to automate the tuning of hyperparameters and network architecture instead of the traditional trial-error tuning process. 3) Uncertainty Analysis: This step aims to perform Monte Carlo (MC) simulations using the successfully trained surrogate model to explore uncertainty propagation. The sampled realizations are collected in the form of distributions from which the probabilistic forecast of percentiles, P10, P50, and P50, are evaluated. We propose a data-driven, physics-featuring surrogate model based on LSTM for CO2 leakage rate forecasting. We demonstrate its performance in terms of accuracy and efficiency by comparing it with ground-truth solutions. The proposed deep-learning workflow shows promising potential and could be readily implemented in commercial-scale GCS for real-time monitoring applications.
{"title":"CO2 Leakage Rate Forecasting Using Optimized Deep Learning","authors":"Xupeng He, Weiwei Zhu, R. Santoso, M. AlSinan, H. Kwak, H. Hoteit","doi":"10.2118/206222-ms","DOIUrl":"https://doi.org/10.2118/206222-ms","url":null,"abstract":"\u0000 Geologic CO2 Sequestration (GCS) is a promising engineering technology to reduce global greenhouse emissions. Real-time forecasting of CO2 leakage rates is an essential aspect of large-scale GCS deployment. This work introduces a data-driven, physics-featuring surrogate model based on deep-learning technique for CO2 leakage rate forecasting. The workflow for the development of data-driven, physics-featuring surrogate model includes three steps: 1) Datasets Generation: We first identify uncertainty parameters that affect the objective of interests (i.e., CO2 leakage rates). For the identified uncertainty parameters, various realizations are then generated based on Latin Hypercube Sampling (LHS). High-fidelity simulations based on a two-phase black-oil solver within MRST are performed to generate the objective functions. Datasets including inputs (i.e., the uncertainty parameters) and outputs (CO2 leakage rates) are collected. 2) Surrogate Development: In this step, a time-series surrogate model using long short-term memory (LSTM) is constructed to map the nonlinear relationship between these uncertainty parameters as inputs and CO2 leakage rates as outputs. We perform Bayesian optimization to automate the tuning of hyperparameters and network architecture instead of the traditional trial-error tuning process. 3) Uncertainty Analysis: This step aims to perform Monte Carlo (MC) simulations using the successfully trained surrogate model to explore uncertainty propagation. The sampled realizations are collected in the form of distributions from which the probabilistic forecast of percentiles, P10, P50, and P50, are evaluated. We propose a data-driven, physics-featuring surrogate model based on LSTM for CO2 leakage rate forecasting. We demonstrate its performance in terms of accuracy and efficiency by comparing it with ground-truth solutions. The proposed deep-learning workflow shows promising potential and could be readily implemented in commercial-scale GCS for real-time monitoring applications.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"39 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76002126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
New water treatment facilities in the Gulf of Mexico include a seawater Sulfate Removal Unit (SRU) to mitigate reservoir souring and scaling. The general industry sulfate target for offshore SRU is usually 20 mg/L or even 40 mg/L; however, some facilities may require <10 mg/L of sulfate in injection water, which makes water quality monitoring more critical and challenging. Current industrial practice relies on only pressure drop and a constant cleaning interval frequency to perform SRU maintenance which may result in reduced membrane life due to frequency cleaning or severe membrane fouling without the capability to predict fouling based on process conditions. The machine learning techniques applied will fill the gap and deliver a prediction model based on both simulation and real-time field data. This model will track and monitor the system key performance indicators (KPIs) including pressure, membrane fouling factor (FF), permeate sulfate concentration etc. The monitoring and prediction of these KPIs provide estimates on when the next maintenance procedure is required, track membrane system status for troubleshooting and actions, and optimize membrane performance by tuning operation conditions.
{"title":"Offshore Water Treatment KPIs Using Machine Learning Techniques","authors":"L. Flores, Martin Morles, Cheng Chen","doi":"10.2118/206173-ms","DOIUrl":"https://doi.org/10.2118/206173-ms","url":null,"abstract":"\u0000 New water treatment facilities in the Gulf of Mexico include a seawater Sulfate Removal Unit (SRU) to mitigate reservoir souring and scaling. The general industry sulfate target for offshore SRU is usually 20 mg/L or even 40 mg/L; however, some facilities may require <10 mg/L of sulfate in injection water, which makes water quality monitoring more critical and challenging. Current industrial practice relies on only pressure drop and a constant cleaning interval frequency to perform SRU maintenance which may result in reduced membrane life due to frequency cleaning or severe membrane fouling without the capability to predict fouling based on process conditions. The machine learning techniques applied will fill the gap and deliver a prediction model based on both simulation and real-time field data. This model will track and monitor the system key performance indicators (KPIs) including pressure, membrane fouling factor (FF), permeate sulfate concentration etc. The monitoring and prediction of these KPIs provide estimates on when the next maintenance procedure is required, track membrane system status for troubleshooting and actions, and optimize membrane performance by tuning operation conditions.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"18 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79662207","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the process of deriving the reservoir petrophysical properties of a basin, identifying the pay capability of wells by interpreting various geological formations is key. Currently, this process is facilitated and preceded by well log correlation, which involves petrophysicists and geologists examining multiple raw log measurements for the well in question, indicating geological markers of formation changes and correlating them with those of neighboring wells. As it may seem, this activity of picking markers of a well is performed manually and the process of ‘examining’ may be highly subjective, thus, prone to inconsistencies. In our work, we propose to automate the well correlation workflow by using a Soft- Attention Convolutional Neural Network to predict well markers. The machine learning algorithm is supervised by examples of manual marker picks and their corresponding occurrence in logs such as gamma-ray, resistivity and density. Our experiments have shown that, specifically, the attention mechanism allows the Convolutional Neural Network to look at relevant features or patterns in the log measurements that suggest a change in formation, making the machine learning model highly precise.
{"title":"Automating Well Log Correlation Workflow Using Soft Attention Convolutional Neural Networks","authors":"A. Abubakar, Mandar Kulkarni, A. Kaul","doi":"10.2118/205985-ms","DOIUrl":"https://doi.org/10.2118/205985-ms","url":null,"abstract":"\u0000 In the process of deriving the reservoir petrophysical properties of a basin, identifying the pay capability of wells by interpreting various geological formations is key. Currently, this process is facilitated and preceded by well log correlation, which involves petrophysicists and geologists examining multiple raw log measurements for the well in question, indicating geological markers of formation changes and correlating them with those of neighboring wells. As it may seem, this activity of picking markers of a well is performed manually and the process of ‘examining’ may be highly subjective, thus, prone to inconsistencies. In our work, we propose to automate the well correlation workflow by using a Soft- Attention Convolutional Neural Network to predict well markers. The machine learning algorithm is supervised by examples of manual marker picks and their corresponding occurrence in logs such as gamma-ray, resistivity and density. Our experiments have shown that, specifically, the attention mechanism allows the Convolutional Neural Network to look at relevant features or patterns in the log measurements that suggest a change in formation, making the machine learning model highly precise.","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"1 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82990744","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
M. J. Ahsan, Shaikha Al-Turkey, N. Rane, F. Snasiri, A. Moustafa, H. Benyounes
The acquisition of mud gas data for well control and gathering of geological information is a common practice in oil and gas drilling. However, these data are scarcely used for reservoir evaluation as they are presumably considered as unreliable and non-representative of the formation content. Recent development in gas extraction from drilling mud and analyzing equipment has greatly improved the data quality. Combined with proper analysis and interpretation, these new datasets give valuable information in real-time lithological changes, hydrocarbons content, water contacts and vertical changes in fluid over a pay interval. Post completion, Mud logging data have been compared with PVT results and they have shown excellent correlation on the C1-C5 composition, confirming the consistency between gas readings and reservoir fluid composition. Having such information in real time has given the oil company the opportunity to optimize its operations regarding formation evaluation, e.g downhole sampling, wireline logging or testing programs. Formation fluid is usually obtained during well tests, either by running downhole tools into the well or by collecting the fluid at surface. Therefore, its composition remains unknown until the arrival of the PVT well test results. This case intends to use mud gas information collected while drilling to predict information about the reservoir fluid composition in near real time. To achieve this goal we compared mud gas data collected while drilling with reservoir fluid compositional results. Pressure volume temperature (PVT) analysis is the process of determining the fluid behaviors and properties of oil and gas samples from existing wells. The reason any oil and gas company decides to drill a well is to turn the project into an oil-producing asset. But the value of the oil extracted from a single well is not the same as the value of the oil produced from another. The makeup of the oil, which can be determined from the compositional analysis, is an important piece of the equation that determines how profitable the play will be. The compositional analysis will determine just how much of each type of petroleum product can be produced from a single barrel of oil from that wells. Formation samples were obtained from offset wells in the Marrat Formation. These datasets gave valuable indications on fluid properties and phase behavior in the reservoir and provided strong base for reservoir engineering analysis, simulation and surface facilities design. The comparison of the gas data to PVT results gives a good match for reservoir fluid finger print, early acquisition of this data will help for decision enhancement for field development.
{"title":"Advanced Gas While Drilling GWD Comparison with Pressure Volume Temperature PVT Analysis to Obtain Information About the Reservoir Fluid Composition, a Case Study from East Kuwait Jurassic Reservoir","authors":"M. J. Ahsan, Shaikha Al-Turkey, N. Rane, F. Snasiri, A. Moustafa, H. Benyounes","doi":"10.2118/206296-ms","DOIUrl":"https://doi.org/10.2118/206296-ms","url":null,"abstract":"\u0000 \u0000 \u0000 The acquisition of mud gas data for well control and gathering of geological information is a common practice in oil and gas drilling. However, these data are scarcely used for reservoir evaluation as they are presumably considered as unreliable and non-representative of the formation content.\u0000 Recent development in gas extraction from drilling mud and analyzing equipment has greatly improved the data quality. Combined with proper analysis and interpretation, these new datasets give valuable information in real-time lithological changes, hydrocarbons content, water contacts and vertical changes in fluid over a pay interval.\u0000 \u0000 \u0000 \u0000 Post completion, Mud logging data have been compared with PVT results and they have shown excellent correlation on the C1-C5 composition, confirming the consistency between gas readings and reservoir fluid composition. Having such information in real time has given the oil company the opportunity to optimize its operations regarding formation evaluation, e.g downhole sampling, wireline logging or testing programs.\u0000 Formation fluid is usually obtained during well tests, either by running downhole tools into the well or by collecting the fluid at surface. Therefore, its composition remains unknown until the arrival of the PVT well test results. This case intends to use mud gas information collected while drilling to predict information about the reservoir fluid composition in near real time. To achieve this goal we compared mud gas data collected while drilling with reservoir fluid compositional results.\u0000 Pressure volume temperature (PVT) analysis is the process of determining the fluid behaviors and properties of oil and gas samples from existing wells.\u0000 \u0000 \u0000 \u0000 The reason any oil and gas company decides to drill a well is to turn the project into an oil-producing asset. But the value of the oil extracted from a single well is not the same as the value of the oil produced from another. The makeup of the oil, which can be determined from the compositional analysis, is an important piece of the equation that determines how profitable the play will be. The compositional analysis will determine just how much of each type of petroleum product can be produced from a single barrel of oil from that wells.\u0000 \u0000 \u0000 \u0000 Formation samples were obtained from offset wells in the Marrat Formation. These datasets gave valuable indications on fluid properties and phase behavior in the reservoir and provided strong base for reservoir engineering analysis, simulation and surface facilities design.\u0000 The comparison of the gas data to PVT results gives a good match for reservoir fluid finger print, early acquisition of this data will help for decision enhancement for field development.\u0000","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"46 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87362217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Rasha Al-Muraikhi, Nami Al-Mutairi, Karim Ousdidene, C. Magnier, Sachin Sharma, H. Benyounes
As the pursuit of oil and gas in Middle East Jurassic carbonates reservoirs grows, it is increasingly evident that horizontal wellbore placement, or targeting, plays a first-order role in the production capability of a well. Indeed, the percentage of a wellbore "in target" is a common metric used when evaluating the causes for good or poor production from any particular well. The most common process used for geosteering a horizontal wellbore into a chosen target is the correlation of logging-while-drilling (LWD) total gamma-ray (GR) to a vertical pilot-hole GR log or offset wells GR logs. However, limitations inherent to this procedure can reduce the ability to effectively use LWD GR data due to 4 ½" slim hole diameter and mud telemetry issues, the non-descript signal from LWD tools due to high pressure and high temperature and the possibility of lost signal from LWD tools. In addition, the thickness of MRW-F11 targeted reservoir is limited to plus or minus 22 ft and low GR contrast from bed to bed might lead to loss of directional control in the target MRW-F11. To accurately geosteer a well, Geochemical analyses of drilled cuttings are proposed to assist well placement. The analyses performed were elemental data derived from energy-dispersive X-ray fluorescence (ED-XRF) and mineralogical quantitative content acquired from the direct measurement from energy-dispersive X-ray Diffraction (ED-XRD). The Elemental and mineralogy data were acquired from drilling cuttings taken at ten feet intervals, from two offsets wells. The mineral and elemental data were used to build a chemo-stratigraphic profile and zonation of the sedimentary section. Chemo-stratigraphic zones are defined as having multiple elements and keys ratios (where possible) which illustrate distinct changes in chemical and mineralogical composition profiles from one zone to another. These zones were correlated over reasonable distances (at a minimum the length of the horizontal wellbore) and can be readily identifiable in cuttings. Using these criteria chemo-stratigraphic zonation's have been constructed in the Middle Marrat formation going from MRW-F1 toward MRW-F11 layer. Well site ED-XRF and ED-XRD data were used in conjunction with LWD Gamma Ray to geosteer at approximately 22 feet thin zone which resides at the base of an approximately 100 ft thick reservoir carbonate section of the main MRW-F11 reservoir. The LWD GR Signal was 45 ft behind the bit while all XRF and XRD data were at plus or minus 5 feet while sliding at plus or minus 10 ft in rotary mode and with a controlled slow rate of penetration (ROP) of 10 ft/hr. Geochemical rock analyses (GEAR) using XRF & XRD chemical analyses was the unique reference for approximately 500 ft interval to geosteer the well when LWD lost the signal, wiper trip was cancelled which considerably reduced drilling costs. Well site XRF and XRD data was successfully applied to geosteer the well, determine the position of the wellbore in zones
{"title":"Using XRF Elemental Data and XRD Direct Measured Mineralogy for an Accurate Wellbore Placement and Geosteering through Carbonates Reservoirs* Drilled Within 04 ½\" Slim Hole: A Case Study from a Jurassic Middle Marrat Carbonates Reservoir-Kuwait","authors":"Rasha Al-Muraikhi, Nami Al-Mutairi, Karim Ousdidene, C. Magnier, Sachin Sharma, H. Benyounes","doi":"10.2118/206328-ms","DOIUrl":"https://doi.org/10.2118/206328-ms","url":null,"abstract":"\u0000 As the pursuit of oil and gas in Middle East Jurassic carbonates reservoirs grows, it is increasingly evident that horizontal wellbore placement, or targeting, plays a first-order role in the production capability of a well. Indeed, the percentage of a wellbore \"in target\" is a common metric used when evaluating the causes for good or poor production from any particular well. The most common process used for geosteering a horizontal wellbore into a chosen target is the correlation of logging-while-drilling (LWD) total gamma-ray (GR) to a vertical pilot-hole GR log or offset wells GR logs. However, limitations inherent to this procedure can reduce the ability to effectively use LWD GR data due to 4 ½\" slim hole diameter and mud telemetry issues, the non-descript signal from LWD tools due to high pressure and high temperature and the possibility of lost signal from LWD tools. In addition, the thickness of MRW-F11 targeted reservoir is limited to plus or minus 22 ft and low GR contrast from bed to bed might lead to loss of directional control in the target MRW-F11.\u0000 To accurately geosteer a well, Geochemical analyses of drilled cuttings are proposed to assist well placement. The analyses performed were elemental data derived from energy-dispersive X-ray fluorescence (ED-XRF) and mineralogical quantitative content acquired from the direct measurement from energy-dispersive X-ray Diffraction (ED-XRD). The Elemental and mineralogy data were acquired from drilling cuttings taken at ten feet intervals, from two offsets wells.\u0000 The mineral and elemental data were used to build a chemo-stratigraphic profile and zonation of the sedimentary section. Chemo-stratigraphic zones are defined as having multiple elements and keys ratios (where possible) which illustrate distinct changes in chemical and mineralogical composition profiles from one zone to another. These zones were correlated over reasonable distances (at a minimum the length of the horizontal wellbore) and can be readily identifiable in cuttings. Using these criteria chemo-stratigraphic zonation's have been constructed in the Middle Marrat formation going from MRW-F1 toward MRW-F11 layer.\u0000 Well site ED-XRF and ED-XRD data were used in conjunction with LWD Gamma Ray to geosteer at approximately 22 feet thin zone which resides at the base of an approximately 100 ft thick reservoir carbonate section of the main MRW-F11 reservoir. The LWD GR Signal was 45 ft behind the bit while all XRF and XRD data were at plus or minus 5 feet while sliding at plus or minus 10 ft in rotary mode and with a controlled slow rate of penetration (ROP) of 10 ft/hr.\u0000 Geochemical rock analyses (GEAR) using XRF & XRD chemical analyses was the unique reference for approximately 500 ft interval to geosteer the well when LWD lost the signal, wiper trip was cancelled which considerably reduced drilling costs.\u0000 Well site XRF and XRD data was successfully applied to geosteer the well, determine the position of the wellbore in zones ","PeriodicalId":10928,"journal":{"name":"Day 2 Wed, September 22, 2021","volume":"12 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2021-09-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"87748653","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：