Evolving Smart Model to Predict the Combustion Front Velocity for In Situ Combustion

Abstract

To determine the breakthrough time of the combustion front in the in situ combustion process for heavy oil recovery processes, no records have been reported in previous literature to date. In this work, the developed model was inspired by a new intelligent method called the “least‐squares support vector machine” (LSSVM) to specify the combustion front velocity in heavy oil recovery process. The proposed approach is applied to the experimental data from Iranian oil fields and reported data from the literature has been incorporated to develop and test this model. The estimated outcomes from the LSSVM approach are compared to the aforementioned actual in situ combustion data. By comparing the results obtained from suggested method with the relevant experimental ones it is clear that the LSSVM approach predicts the combustion front velocity with reasonable degree of precision. It worth mentioning that the LSSVM contains no conceptual errors, such as over‐fitting, which is an issue for artificial neural networks. The results of this study could couple with the industrial reservoir simulation software for heavy oil reservoirs to select the proper production method or achieve related goals.

Development of robust model to estimate gas–oil interfacial tension using least square support vector machine: Experimental and modeling study

“…Recently, Ahmadi and colleagues made huge amounts of efforts to perform various intelligent based approaches for specifying the challenging issues in oil and gas industries [10][11][12][13][14][15][16][17][18][19][20][21]. For example, Ahmadi et al, (2013) applied hybrid approach to specify permeability of the petroleum reservoir with routine conventional petrophysical logs [20].…”

Connectionist intelligent model estimates output power and torque of stirling engine

“…Nevertheless, they might propel to the random initialization of the networks and alteration of the terminating criteria during optimization of the ANN parameters [39][40][41][42][43]. The aforementioned features may decline the employing of the ANN methods for external estimations which indicates external inputs excluding in training, optimization, and test methods of treatment of the relevant networks.…”

“…The support vector machine (SVM) is an effective approach evolved from the machine-learning community concept [40][41][42][43][44][45]. A SVM is a device for number of related supervised learning ways that analyze data and identify patterns, applied for regression analysis.…”

“…A SVM is a device for number of related supervised learning ways that analyze data and identify patterns, applied for regression analysis. The SVM is assumed as a non-probabilistic binary linear classifier [40][41][42][43][44][45].…”

Phase equilibrium modeling of semi-clathrate hydrates of seven commonly gases in the presence of TBAB ionic liquid promoter based on a low parameter connectionist technique

“…Moreover, the aforementioned optimization approaches can be coupled with least square SVM [46][47][48][49][50][51][52][53][54][55][56][57][58].…”