Forecasting the Post-Fracturing Response of Oil Wells in a Tight Reservoir

Joo, Jhi In; Wu, Dongrui; Mendel, Jerry M.; Bugacov, Alejandro

doi:10.2118/121394-ms

Cited by 3 publications

(1 citation statement)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, since emotion evaluations are subjective and subtle, usually multiple assessors (e.g., 6-17 in the VAM corpus [2], and at least 110 in IADS-2 [3]) are needed for each utterance, which is very time-consuming and labor-intensive. Another example is 180-day post-fracturing cumulative oil production prediction in enhanced oil recovery in the oil and gas industry [4]. The inputs (fracturing parameters of an oil well, such as its location, length of perforations, number of zones/holes, volumes of injected slurry/water/sand, etc.)…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Pool-Based Active Learning for Linear Regression

Liu,

2020

Preprint

Self Cite

View full text Add to dashboard Cite

In many real-world machine learning applications, unlabeled data can be easily obtained, but it is very timeconsuming and/or expensive to label them. So, it is desirable to be able to select the optimal samples to label, so that a good machine learning model can be trained from a minimum amount of labeled data. Active learning (AL) has been widely used for this purpose. However, most existing AL approaches are supervised: they train an initial model from a small amount of labeled samples, query new samples based on the model, and then update the model iteratively. Few of them have considered the completely unsupervised AL problem, i.e., starting from zero, how to optimally select the very first few samples to label, without knowing any label information at all. This problem is very challenging, as no label information can be utilized. This paper studies unsupervised pool-based AL for linear regression problems. We propose a novel AL approach that considers simultaneously the informativeness, representativeness, and diversity, three essential criteria in AL. Extensive experiments on 14 datasets from various application domains, using three different linear regression models (ridge regression, LASSO, and linear support vector regression), demonstrated the effectiveness of our proposed approach.

show abstract

Section: Introductionmentioning

confidence: 99%

Unsupervised Pool-Based Active Learning for Linear Regression

Liu,

2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

A Predictive Model for Improving the Efficiency of Frac Jobs

Korjani

Mendel

Ershaghi

2015

SPE Western Regional Meeting

View full text Add to dashboard Cite

This paper presents a new method for forecasting post-fracturing responses in a tight oil reservoir using historical hydraulic fracturing data. The methodology is based on a nonlinear regression method called Variable Structure Regression (VSR). Data for 80 frac jobs were used to calibrate and predict responses. Information per fracturing job included feet of perforation, number of perforations per stage, number of stages, pad volume, slurry volume and sand volume. Information for a set of hydraulically fractured vertical wells was used to test the proposed technique. The VSR method establishes the optimal nonlinear dependencies of variables to predict the cumulative liquid production after fracturing. This method determines the relationships between fracturing parameters and post-fracturing productions in the form of linguistic terms, thereby providing a physical understanding of the process. The exact mathematical structure of these linguistic terms and the number of terms are established simultaneously, thereby freeing the end user from time-consuming trial and error studies. Meanwhile, the knowledge of domain experts can be preserved as qualitative expert knowledge and may be combined with quantitative data. A Monte Carlo simulation of 5-fold cross-validation method on fracture optimization data set shows the practical accuracy of the VSR model on forecasting post-fracturing performance. The VSR method achieved satisfactory prediction accuracy as well as statistical performance given the relatively small training data size and the complexity of the problem. The final result shows increase prediction accuracy for all the wells that outperforms other existing methods.

show abstract

CiSoft and Smart Oilfield Technologies

Ershaghi

Paul

Hauser³

et al. 2016

All Days

View full text Add to dashboard Cite

With the ever increasing complexity of upstream oilfield operations, it has long been recognized that there is a serious need for capturing, communicating, analyzing, and using real-time information for better decision making, improving safety and reducing the operational cost of the oilfield operations. During the last two decades, the use of information technology for advanced services such as wireless, voice, and video communication plus advanced algorithms for data cleansing, data mining, data-driven analytics have changed the landscape for many international and national oil companies. These types of enhancements can lead to an increase in efficiencies when all stakeholders in an organization can access networked applications and services anywhere at any time, Figure 1. Although the need for smart oil field technology was long recognized, the development of smart oil field tools did not keep pace with the increased complexity or the availability of large amounts of data. Even when these tools became available, there was a reticence to immediately utilize them. Managing the required organizational and cultural changes required to effectively deploy and apply smart oilfield technologies has been a challenge. But the key to success has been upper management resolve to utilize increased flows of information and work processes redesign to reduce operating costs and increase capital efficiency. Development and deployment of advanced technologies require a blend of information science expertise and domain knowledge in oilfield operations. Forming and successfully managing a strategic alliance between industry and universities can be an effective solution to forming a hub for critical technology development, and also preparing digital oilfield engineers to expedite deployment. This paper presents an analysis of a more than a decade of advances in smart oilfield technologies with examples of tools and techniques developed at USC Center for Interactive Smart Oilfield Technologies (CiSoft).). This globally unique strategic alliance (www.cisoft.usc.edu) was initiated in 2003 as an extended education and research partnership with the Chevron Corporation to build new capabilities, academic degree programs, and technology advancements for the first generation of digital oilfields. Through the unique collaborative efforts of USC researchers from information technology and petroleum engineering, smart working solutions have been developed to advance the cause of oilfield safety, efficiency, and enhanced economics using intelligent algorithms and tools. The success of the Center has been the development of innovative workflow processes addressing how typical human intensive oilfield monitoring and control can be transformed into solutions that can reduce failures, substantially enhance raw and interpreted data access, and help in engineering decision making with innovative machine learning algorithms. The alliance has included efforts of 180 Ph.D. students, producing 160 publications, over 50 inventions and 40 patents in a decade time. Many of these publications are source materials that can be used for development of organized courses and training programs on smart oilfield technologies. Topics discussed include the working model for this alliance, research management process, an educational component and a new dimension of the Center to commercialize certain technologies through service companies. Technologies already patented are discussed in the technical papers listed in the reference section. Focus on many projects has been on preventing failures and for lowering operational cost. Best practices and lessons learned from formation and operation of CiSoft can serve as effective ways for establishing more similar strategic partnerships between the industry and academia.

show abstract

Forecasting the Post-Fracturing Response of Oil Wells in a Tight Reservoir

Cited by 3 publications

References 8 publications

Unsupervised Pool-Based Active Learning for Linear Regression

Unsupervised Pool-Based Active Learning for Linear Regression

A Predictive Model for Improving the Efficiency of Frac Jobs

CiSoft and Smart Oilfield Technologies

Contact Info

Product

Resources

About