Integrating neural networks and fuzzy logic for improved reservoir property prediction and prospect ranking

Aminzadeh, Fred; Brouwer, Friso

doi:10.1190/1.2369863

Cited by 12 publications

(7 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is stated that much of the uncertainty in many situations is due to imprecision and subjectivity rather than underlying randomness, Aminzadeh et al, [5]. Fuzzy logic is an appropriate tool to deal with uncertainty of this type, inherent in most physical or natural systems.…”

Section: Fuzzy Logicmentioning

confidence: 99%

See 1 more Smart Citation

Identification of Horizontal Well Placement Using Fuzzy Logic

Popa

2013

SPE Annual Technical Conference and Exhibition

View full text Add to dashboard Cite

Estimated to have over 3.5 billion barrels of original oil in place, the 114 year old Kern River Oil Field was a sleeping giant until the late 1960's when steamflooding began and oil production ramped up from 19,000 to over 142,000 BOPD by the mid 1980's. Kern River Field is now in decline, but still produces over 70,000 BOPD, making it one of the largest onshore oil producers in the continental United States. In 2007 the field's cumulative production reached two billion barrels of heavy oil. With the introduction of horizontal wells in 2007, the production decline curve that had been at approximately 6%, began to flatten and today the decline rate is at 2%. Since the discovery of the field, over 19,000 wells have been drilled with the largest producers, horizontal wells, drilled in the last 6 years.Though the oil production is in decline, new opportunities are continuing to be identified and exploited using a broad array of reservoir management tools. The primary tool utilized for both tactical and strategic planning is a full field 3D reservoir model. The static lithologic framework for this full field model uses over 12,500 normalized resistivity logs. Temperature, steam and oil saturation data from a network of over 650 surveillance wells are periodically incorporated into this model to provide current reservoir conditions. Project teams first identify their opportunities in the full field model using a filtering criteria. For example, horizontal infill wells seeking "hot bottom oil" within partially depleted reservoir intervals can be developed around geo-body criteria including property cut-off values for gas saturation, oil saturation, temperature, resistivity and thickness.The filtering criterion was useful in identifying well locations; however, the pool of candidates significantly declined due to the exclusive nature of the hard cut-off approach. The lack of candidates threatened the continued horizontal drilling program. A Fuzzy Logic approach was considered to deal with the imprecise information and aggregate multiple variables with suitable membership functions. The Fuzzy Inference System (FIS) using the Mamdani model incorporated knowledge rules developed by the asset team. Once developed, the system was run for each grid block of the nine productive formations of the field, generating areas of interest for identification and placement of horizontal wells.The output of the system allowed generation of "quality maps" which revealed target areas with substantial potential of horizontal drilling. Moreover, some of these target areas were not even considered in previous searches. In a first pass, approximately 600 new locations were identified.The paper presents a new approach for identification and placement of horizontal wells using fuzzy logic. In addition to the large number of wells identified, the methodology focused the asset team on high potential areas, as well as offered an objective approach to deal with imprecise and uncertain reservoir information. Due to its success in Kern River, the tech...

show abstract

Section: Fuzzy Logicmentioning

confidence: 99%

“…To point out the fundamental differences between classical logic and fuzzy logic, we consider the following example, Aminzadeh, et al, [5]. A petro-physicist may consider sandstone with porosity of less than 3 as a low porosity.…”

Section: Fuzzy Logicmentioning

confidence: 99%

Identification of Horizontal Well Placement Using Fuzzy Logic

Popa

2013

SPE Annual Technical Conference and Exhibition

View full text Add to dashboard Cite

show abstract

“…The use of the GA for the estimation of parameters such as porosity and recovery rate can be counted as another application in this field . In addition, ANN has been used to determine the physical properties of petroleum through the correlations between these parameters and other known variables such as temperature, pressure, and normal boiling point . Mohaghegh studied the application of several AI techniques in oil and gas reservoirs characterization.…”

Section: Introductionmentioning

confidence: 99%

“…[16] In addition, ANN has been used to determine the physical properties of petroleum through the correlations between these parameters and other known variables such as temperature, pressure, and normal boiling point. [17] Mohaghegh [18] studied the application of several AI techniques in oil and gas reservoirs characterization. Zendehboudi et al [19] focused on the applications of hybrid models (HMs) in chemical processes, oil and gas processes, and applied energy systems.…”

Section: Introductionmentioning

confidence: 99%

Development of a data‐driven fuzzy screening model for enhanced oil recovery methods using an adaptive weighting system

2019

View full text Add to dashboard Cite

Screening of enhanced oil recovery (EOR) methods is a main prerequisite for EOR planning and design. In this study, an integrated data‐driven screening model (DDSM) is developed to improve EOR screening using the combined capabilities of the fuzzy expert approach (FEA) and support vector regression (SVR) techniques. In this study, EOR field data from the past 40 years were reviewed to generate an updated and reliable EOR criteria table as a basis to construct a fuzzy screening model. The DDSM was evaluated to determine the quantitative screening and ranking of EOR methods using seven field datasets, including the fast forecasting of the nominated EOR methods. In order to improve screening performance, a fuzzy model was integrated using 4 SVR models to predict the adaptive weights of the screening parameter for decision making. The SVR models can predict the recovery factor (RF) of EOR methods including gas, chemical, steam, and combustion to calculate the adaptive effective weight of the screening parameters. The SVR models were trained with datasets generated from simulations of the EOR process. The absolute average error (AAE) of the SVR models from the simulation varied within the range of 0.078–0.095 for the prediction of the RF. The DDSM results were compatible to the data published in other literature. In addition, the developed model can provide comparable results to common screening software. The results showed improvements due to the adaptive weighting system on the EOR methods’ screening for the studied reservoirs relative to the fuzzy engine with constant weights. The presented integrated model can guide the screening process to select the efficient EOR method in practical applications.

show abstract

“…Applications of these hybrid systems can be found in petroleum engineering (Mohaghegh 2000), particularly in the areas of reservoir lithology identification and property prediction (Zhou et al 1993;Lim 2005;Aminzadeh & Brouwer 2006), reservoir management (Nikravesh et al 1998;Alimonti & Falcone 2004;Popa & Cassidy 2012), and optimization of well operations design (Mohaghegh & Reeves 2000;Murillo et al 2009;Attia et al 2013).…”

Section: Introductionmentioning

confidence: 99%

An Integrated Application of Cluster Analysis and Artificial Neural Networks for SAGD Recovery Performance Prediction in Heterogeneous Reservoirs

Amirian

Leung

Zanon

et al. 2014

Day 1 Tue, June 10, 2014

View full text Add to dashboard Cite

Evaluation of steam-assisted gravity drainage (SAGD) performance that involves detailed compositional simulations is usually deterministic, cumbersome, expensive (manpower and time consuming), and not quite suitable for practical decision making and forecasting, particularly when dealing with highdimensional data space consisting of large number of operational and geological parameters. Data-driven modeling techniques, which entail comprehensive data analysis and implementation of machine learning methods for system forecast, provide an attractive alternative.In this paper, artificial neural network (ANN) is employed to predict SAGD production in heterogeneous reservoirs, an important application that is lacking in existing literature. Numerical flow simulations are performed to construct a training data set consists of various attributes describing characteristics associated with reservoir heterogeneities and other relevant operating parameters. Empirical Arps decline parameters are tested successfully for parameterization of cumulative production profile and considered as outputs of the ANN models. Sensitivity studies on network configurations are also investigated. Principal components analysis (PCA) is performed to reduce the dimensionality of the input vector, improve prediction quality, and limit over-fitting. In a case study, reservoirs with distinct heterogeneity distributions are fed to the model. It is shown that robustness and accuracy of the prediction capability are greatly enhanced when cluster analysis are performed to identify internal data structures and groupings prior to ANN modeling. Both deterministic and fuzzy-based clustering techniques are compared, and separate ANN model is constructed for each cluster. The model is then verified using a testing data set (cases that have not been used during the training stage).The proposed approach can be integrated directly into most existing reservoir management routines. In addition, incorporating techniques for dimensionality reduction and clustering with ANN demonstrates the viability of this approach for analyzing large field data set. Given that quantitative ranking of operating areas, robust forecasting, and optimization of heavy oil recovery processes are major challenges faced by the industry, the proposed research highlights the significant potential of applying effective data-driven modeling approaches in analyzing other solvent-additive steam injection projects.

show abstract

Integrating neural networks and fuzzy logic for improved reservoir property prediction and prospect ranking

Cited by 12 publications

References 3 publications

Identification of Horizontal Well Placement Using Fuzzy Logic

Identification of Horizontal Well Placement Using Fuzzy Logic

Development of a data‐driven fuzzy screening model for enhanced oil recovery methods using an adaptive weighting system

An Integrated Application of Cluster Analysis and Artificial Neural Networks for SAGD Recovery Performance Prediction in Heterogeneous Reservoirs

Contact Info

Product

Resources

About