Search citation statements
Paper Sections
Citation Types
Year Published
Publication Types
Relationship
Authors
Journals
Normally, the optimization of hydraulic fracturing performance is limited to pre-job modeling and analytics. A design is determined for a particular well or project and applied without significant change during the course of the stimulation. Performance results are collected during the job and then analyzed after the fact, with the primary purpose of designing for the next project. Significant design improvements can be made by evaluating stage performance in real-time as the well is being stimulated. Unfortunately, real-time analytics are difficult because the immense of volume, variety, and velocity of the available data. The typical frac fleet captures metered data from as many as one hundred measurement points simultaneously on a second-by-second basis. This means that for a single stage, the comma-separated values (CSV) files containing the recorded channels often include over one million discrete data points. Utilizing these large files (approximately 5 MB) with typical off-the-shelf software can be time-consuming. The manual process of file acquisition by analytical staff alone can often exceed the time available between stages. While these files are an invaluable resource, they are often left untouched until long after a job is completed, if they are ever used at all. Cloud-based analytics greatly shorten the acquisition and utilization timeline, making near real-time analysis possible. While the challenges involved in utilizing "big data"; for actionable analytics are frequently discussed, the technology and approaches described in this paper are relatively new to the field of real-time stage management. This paper introduces a novel and highly effective approach in the field of hydraulic fracturing optimization. The history of CSV analysis is presented along with examples of specific types of beneficial stage analytics.
Normally, the optimization of hydraulic fracturing performance is limited to pre-job modeling and analytics. A design is determined for a particular well or project and applied without significant change during the course of the stimulation. Performance results are collected during the job and then analyzed after the fact, with the primary purpose of designing for the next project. Significant design improvements can be made by evaluating stage performance in real-time as the well is being stimulated. Unfortunately, real-time analytics are difficult because the immense of volume, variety, and velocity of the available data. The typical frac fleet captures metered data from as many as one hundred measurement points simultaneously on a second-by-second basis. This means that for a single stage, the comma-separated values (CSV) files containing the recorded channels often include over one million discrete data points. Utilizing these large files (approximately 5 MB) with typical off-the-shelf software can be time-consuming. The manual process of file acquisition by analytical staff alone can often exceed the time available between stages. While these files are an invaluable resource, they are often left untouched until long after a job is completed, if they are ever used at all. Cloud-based analytics greatly shorten the acquisition and utilization timeline, making near real-time analysis possible. While the challenges involved in utilizing "big data"; for actionable analytics are frequently discussed, the technology and approaches described in this paper are relatively new to the field of real-time stage management. This paper introduces a novel and highly effective approach in the field of hydraulic fracturing optimization. The history of CSV analysis is presented along with examples of specific types of beneficial stage analytics.
This paper explores a holistic approach to characterize trouble stages by applying automated event recognition of abnormal pressure increases and associating those events to formation and operational causes. This analysis of pressure increases provides insight into the potential causes of operational difficulties, and the related diagnostics can suggest improvements to future pump schedules. Improving how stages are pumped is profitable both in the short-term (reducing wasted fluid and chemicals, and other remediation measures) and in the long-term (increased well productivity). Quantifying how design decisions ultimately affect operations can help decrease the frequency of operational problems and help realize these gains. In this study, the identification of problematic frac stages was initially performed manually (stage-by-stage) using a cloud-based hydraulic fracture data application. During this process, the team recognized that the problem stages had their own characteristic pressure signature - a sudden unexplained pressure increase in the absence of rate changes. A machine learning algorithm was then developed to automatically identify this type of signature. Additionally, previously published machine learning algorithms were used to recognize other operational events of interest, e.g., when proppant reaches the perforations. Then by combining the various events and creating short search windows around each abnormal pressure increase, it is possible to find concurrent operations that may be associated with the observed pressure behavior. A subsequent statistical analysis revealed that abnormal pressure increases often coincided with changes in proppant concentration in problem stages (stages with abnormal treating pressure behavior). This behavior may be due to near-wellbore effects caused by the changing fluid flow dynamics. Furthermore, it was observed that treating pressures that behaved contrary to hydrostatic pressure effects may be useful in identifying when injectivity is lost and provide an early signal for screen outs. Through this holistic approach, we were able to identify trouble stages and discern some diagnostics for automated detection of abnormal treating pressure increases. The team was able to identify areas within the stages that were inefficiently pumped, resulting in cost-savings through optimization of proppant and friction reducer (FR) loadings while maintaining a level of caution to prevent screen outs. Finally, the automated detection of pressure anomalies offers a pathway to the real-time prediction and avoidance of operational difficulties such as pressure outs and screen outs.
The paper presents a practical tool for hydraulic fracturing efficiency evaluation. The tool is based on a data-driven approach that helps in interpreting real-time data. Based on the hydraulic fracturing (HF) job monitoring, statistic metrics and key performance indicators (KPIs) are generated to be valuable input for further designs and identification of potential savings in operation. Machine learning (ML) algorithms are proposed to reduce the tedious work of completion engineers by automatically classifying each treatment schedule's timestamp and assigning the stage label. For operation stages classification Support vector machines and neural networks algorithms are used. These models are trained and evaluated on real-time treatment datasets. After automatic stage recognition, relevant statistic parameters are calculated, enabling advanced data analytics. Detailed analysis of historical data allows to identify the areas for improvements and set new best practices. The first research objective was to gather data from various companies and structure them under the same template to conserve the most critical information gained during the hydraulic fracturing job. Afterwards, the data are preprocessed and labelled by using signal processing routines that significantly decrease the labelling time. The labels or classes are used to define different stages that can be distinguished during the treatment. Finally, the goal is to decrease the necessary time for data labelling. Therefore, two multiclass classification models (Support Vector Machines (SVM) and Neural Network (NN)) are built and evaluated. Based on evaluation metrics, both models resulted in high accuracy and reliable results. However, the SVM model resulted in slightly higher accuracy and an F1 score. The key value of these models is that they provide a computational method to extract a pumping schedule from hydraulic fracturing time-series data automatically. Also, these models allow conducting post-job analysis and choosing the proper pump schedule for a future HF treatment based on previous experience. This past-job analysis could contribute to the effectiveness of future operations by utilizing the materials and fluids more efficiently.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.