Improved predictions of wellhead choke liquid critical-flow rates: Modelling based on hybrid neural network training learning based optimization

Choubineh, Abouzar; Ghorbani, Hamzeh; Wood, David A.; Moosavi, Seyedeh Robab; Khalafi, Elias; Sadatshojaei, Erfan

doi:10.1016/j.fuel.2017.06.131

Cited by 96 publications

(30 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Artificial intelligence algorithms have been extensively applied for predicting flow rates through wellhead chokes (Elhaj et al, 2015;Choubineh et al, 2017). For the dataset evaluated here, there are 180 (M = 180) data records and four variables (N = 4) defining the system, with the dependent variable being the fifth variable.…”

Section: Dataset 1 (Prediction Of Loss Circulation While Drilling)mentioning

confidence: 99%

“…Other machine learning techniques including support vector machine (SVM) (Cortes and Vapnik, 1995), least squares support vector machine (LSSVM) (Suykens and Vandewalle, 1999), and fuzzy logic combinations with ANN such as Adaptive Neuro-Fuzzy Inference Systems (ANFIS) (Sugeno and Kang, 1988;Jang, 1993) of transparency has not inhibited these algorithms from being widely and successfully applied, often hybridized with various optimization algorithms, to improve predictions from various oil, gas and other industrial systems (Li et al, 2013;Meng and Zhao, 2015;Zamani et al, 2015;Choubineh et al, 2017;Yavari et al, 2018). Indeed, there is a growing body of research in the petroleum sector that enters dataset records with multiple variables into opaquely coded soft-computing machine learning and neural network functions (MatLab, in particular), blindly accepting the results in many cases without even checking the veracity of the input data values.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms

Wood

2018

Adv. Geo-Energy Res.

Self Cite

View full text Add to dashboard Cite

It is now commonplace to deploy neural networks and machine-learning algorithms to provide predictions derived from complex systems with multiple underlying variables. This is particularly useful where direct measurements for the key variables are limited in number and/or difficult to obtain. There are many petroleum systems that fit this description. Whereas artificial intelligence algorithms offer effective solutions to predicting these difficult-to-measure dependent variables they often fail to provide insight to the underlying systems and the relationships between the variables used to derive their predictions are obscure. To the user such systems often represent "black boxe". The novel transparent open box (TOB) learning network algorithm described here overcomes these limitations by clearly revealing its intermediate calculations and the weightings applied to its independent variables in deriving its predictions. The steps involved in building and optimizing the TOB network are described in detail. For small to mid-sized datasets the TOB network can be deployed using spreadsheet formulas and standard optimizers; for larger systems coded algorithms and customised optimizers are easy to configure. Hybrid applications combining spreadsheet benefits (e.g. Microsoft Excel Solver) with algorithm code are also effective. The TOB learning network is applied to three petroleum datasets and demonstrates both its learning capabilities and the insight to the modelled systems that it is able to provide. TOB is not proposed as a replacement for neural networks and machine learning algorithms, but as a complementary tool; it can serve as a performance benchmark for some of the less transparent algorithms.

show abstract

Section: Dataset 1 (Prediction Of Loss Circulation While Drilling)mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms

Wood

2018

Adv. Geo-Energy Res.

Self Cite

View full text Add to dashboard Cite

show abstract

“…used a hybrid model of an artificial network with a teaching learning based optimization to predict oil flowrate in chokes. 22 R 2 of their method is equal to 0.981. Although these methods can predict flowrate with more accuracy, they cannot be employed easily, therefore developing a new correlation with flexible form is vital for oil and gas fields.…”

Section: Introductionmentioning

confidence: 99%

New empirical correlation for oil flowrate prediction through chokes

Fuladgar¹,

Vatani

2020

JSCS

View full text Add to dashboard Cite

Accurate prediction of choke restricted multiphase flowrate is always interesting in oil industry in the absence of flowmeters for field production monitoring and management purposes. Also, in a producing well, choke is installed for controlling the production rate. In this research, new correlation is developed for predicting the multiphase flowrate through choke specified for one of the Iranian southern oilfields. For this purpose, 166 sets of flowrate measurement data, was gathered and filtered to 142 data sets, at the quality control step. This data was regressed to a linearized modified correlation between the measured flowrate, the wellhead pressure choke size and the producing gas liquid ratio to find the best set of correlation parameters. The resulted correlation was evaluated by determining its average relative error and root mean square error by the excluded set of test data. The evaluation indicated that new correlation could significantly improve the accuracy of flowrate predictions in contrast to previous prominent correlations.

show abstract

“…indicating more accurate results than the previously published correlations. More recently, Choubineh et al . developed a hybrid version of training learning‐based optimization and ANNs for estimating choke liquid rate in critical state.…”

Section: Introductionmentioning

confidence: 99%

“…Moreover, a least-square SVM model with five input parameters were developed by Baghban et al [2] indicating more accurate results than the previously published correlations. More recently, Choubineh et al [38] developed a hybrid version of training learning-based optimization and ANNs for estimating choke liquid rate in critical state. For enhancing their model performance, the authors have considered the impact of three extra input variables including temperature and oil and gas-specific densities in their model in addition to the choke size, gas-to-liquid ratio, and wellhead pressure.…”

Section: Introductionmentioning

confidence: 99%

Toward gene expression programming for accurate prognostication of the critical oil flow rate through the choke: correlation development

Rostami

Ebadi

2017

Asia-Pacific J Chem Eng

View full text Add to dashboard Cite

In this study, gene expression programming (GEP) was employed to develop a new equation for calculating choke flow rate as a function of fluid properties and wellhead choke characteristics. For this reason, a comprehensive databank was adopted from the petroleum industry, and then, the databank was split into three groups of training set for model construction (70% of database), test set for model evaluation (15% of database), and validation set for avoiding over‐fitting problem (15% of database). For assessing of GEP model, numerous statistical and graphical tools were applied, and then, it was compared with widely applied literature correlations. Consequently, the proposed tool in this study gives the best fit with the actual flow rate data with the average absolute relative deviation of 11.0872% and determination coefficient (R2) of 0.9695. Additionally, cumulative frequency plot demonstrates that the GEP model has the highest frequency over the database. The outcomes of the GEP model were also compared with the outputs of the artificial neural network (ANN) revealing the higher precision of the ANN than the proposed GEP model in this study. At the end, it is worthwhile that the suggested model can be used as reliable tool for choke performance modeling. © 2017 Curtin University and John Wiley & Sons, Ltd.

show abstract

Improved predictions of wellhead choke liquid critical-flow rates: Modelling based on hybrid neural network training learning based optimization

Cited by 96 publications

References 48 publications

A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms

A transparent Open-Box learning network provides insight to complex systems and a performance benchmark for more-opaque machine learning algorithms

New empirical correlation for oil flowrate prediction through chokes

Toward gene expression programming for accurate prognostication of the critical oil flow rate through the choke: correlation development

Contact Info

Product

Resources

About