Net pay determination by artificial neural network: Case study on Iranian offshore oil fields

Masoudi, Pedram; Arbab, Bita; Mohammadrezaei, Hossein

doi:10.1016/j.petrol.2014.07.007

Cited by 15 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are numerous further uses (Alajmi and Ertekin, 2007; Artun et al, 2011a, 2011b; Bansal et al, 2013; Enab and Ertekin, 2014; Enyioha and Ertekin, 2014; Parada and Ertekin, 2012; Ramgulam et al, 2007; Siripatrachai et al, 2014; Sun and Ertekin, 2012, 2017, 2018). Such work also includes Masoudi et al (2012) and Masoudi et al (2014).…”

Section: Methodsmentioning

confidence: 99%

An artificial-intelligence-based petrophysical property predictor for compositional volatile oil reservoir using three-phase production data

Zhang,

Ertekin,

et al. 2024

Energy Exploration & Exploitation

View full text Add to dashboard Cite

When considering multiphase flow scenarios, the interpretation of petrophysical properties poses significant challenges for production forecasts and reservoir modeling. The findings of the numerical modeling were therefore subject to uncertainty because characteristics like relative permeability and capillary pressure curve were hardly ever bound by interpretations. The uncertainty may result in inaccurate predictions of reservoir performance and skewed perceptions of the reservoir. Due to the difficulty in directly interpreting such property from the available field data and the expensive cost of coring, analyses or experimental measurements to determine relative permeability and capillary pressure were infrequently carried out. Such a gap would be filled by a straightforward yet rigorous method. In this study, we develop production projections for a wide range of three-phase compositional volatile oil reservoirs. Then, we used an artificial neural network to figure out how petrophysical characteristics and production data relate to one another. The artificial neural network model was adjusted, and the final trained model was tested blindly to determine how well it predicted permeability, multiphase relative permeability, and capillary pressure data. For the testing scenarios, consistency is seen between the predicted values and the original ones, despite some mispredictions being present. To provide production projections that can be compared to those from the reservoir model that include the initial petrophysical characteristic, the anticipated properties are then propagated into reservoir models. The comparison findings show that for 65/59/34 out of 74 testing scenarios, the reservoir model with artificial neural network-predicted features can anticipate oil/gas/water output with < 20% inaccuracy. With the developed artificial neural network tool, the reservoir engineers can evaluate the three-phase relative permeability surface from rate-transient data conveniently improving the accuracy of the relative permeability data implemented by history matching or from core experiments which sometimes are extremely expensive. The findings of this study can help for a better understanding of the relationships between three-phase rate-transient data and the relative permeability surface as well as the horizontal/vertical permeability.

show abstract

Section: Methodsmentioning

confidence: 99%

An artificial-intelligence-based petrophysical property predictor for compositional volatile oil reservoir using three-phase production data

Zhang,

Ertekin,

et al. 2024

Energy Exploration & Exploitation

View full text Add to dashboard Cite

show abstract

“…In the same manner, Masoudi et al (2014a) used ANN to determine net pay zones within two Iranian offshore oil fields: a carbonate reservoir (Mishrif), and a sandy or clastic one (Burgan). It plainly appears in both geological environments that the application of ANN technique showed results nearly similar to well test results.…”

Section: (Iii) Net Pay Determination -Oil or Gas Recoverymentioning

confidence: 99%

Neural network applications to reservoirs: Physics-based models and data models

Aı̈fa

2014

Journal of Petroleum Science and Engineering

View full text Add to dashboard Cite

show abstract

“…However, the time cost and economic cost of using such formation test apparatuses are extremely high; therefore, considering the economics, this is difficult to promote on a large scale in the whole work area. Subsequently, with the development of mathematical geological methods and machine learning methods, such as fuzzy classifier fusion, Bayesian theory, and artificial neural networks, etc., these methods have also been gradually applied to determine the lower limits of the effective reservoir properties [21][22][23]. As a powerful classification tool, it is not only suitable for high-dimensional data but also simple, easy to understand, and capable of processing a large amount of data in a short time.…”

Section: Introductionmentioning

confidence: 99%

A Method for Judging the Effectiveness of Complex Tight Gas Reservoirs Based on Geophysical Logging Data and Using the L Block of the Ordos Basin as a Case Study

2023

View full text Add to dashboard Cite

As an important unconventional oil and gas resource, the tight gas reservoir faces many technical challenges due to its low porosity, low permeability, and strong heterogeneity. Among them, the accurate definition of effective reservoirs and ineffective reservoirs in tight gas reservoirs directly affects the formulation and adjustment of subsequent development plans. This paper proposes a reservoir effectiveness identification method based on double factors based on conventional geophysical logging data and core experimental data. The double factors considered are based on the logging response and physical parameters of the reservoir. The identification factor F1 is obtained by using the difference in the logging response values of the natural gamma logging curve, compensated density logging curve, and acoustic time difference logging curve in different reservoirs combined with mathematical operation, and the identification factor F2 is calculated by using porosity parameters combined with Archie’s formula. The validity of the reservoir can be judged by the intersection of the above double factors. This method is applied to the Shihezi Formation in the L block, and the applicability of the double factors is compared and analyzed using the traditional method. The results show that the method has strong applicability in tight gas reservoirs and that the accuracy rate reaches 96%. Compared with the direct use of the porosity lower limit method, the accuracy of the judgment is significantly improved, and the calculation is simple, easy to implement, and unaffected by mud invasion. For study areas with different geological backgrounds, the process of this method can also be used to determine the effectiveness of the reservoir. The reservoir effectiveness identification method proposed in this paper has practical engineering significance and lays a solid foundation for subsequent fluid property identification, production calculation, and development plan formulation and adjustment.

show abstract

Net pay determination by artificial neural network: Case study on Iranian offshore oil fields

Cited by 15 publications

References 14 publications

An artificial-intelligence-based petrophysical property predictor for compositional volatile oil reservoir using three-phase production data

An artificial-intelligence-based petrophysical property predictor for compositional volatile oil reservoir using three-phase production data

Neural network applications to reservoirs: Physics-based models and data models

A Method for Judging the Effectiveness of Complex Tight Gas Reservoirs Based on Geophysical Logging Data and Using the L Block of the Ordos Basin as a Case Study

Contact Info

Product

Resources

About