Asset optimization has recently become a crucial issue in Oil&Gas industry, considering oil price conjuncture and an increased awareness on environmental aspects. In this paper, an Artificial Intelligence technique is presented, which is able to manage big datasets to automatically match the production models and propose an operative solution that maximizes the production. The tool is based on a hybrid in-house developed AI technique, integrating deep neural networks, biogenetical algorithms, commercial simulators and real-time data. The workflow starts with the modeling of the production system through physics based commercial simulators. A sensitivity analysis identifies the critical variables, which are then randomly varied with a Sobol distribution, exploring the entire solution domain. With these data, a proxy model to the commercial software is generated: it consists in an artificial neural network able to replicate the field behavior. Finally, the AI tool fed by real-time data is used firstly to match the field behavior and, successively, to maximize the field production. The tool has been developed considering an offshore oil field. The typical uncertain parameters used in the matching phase are fluid characteristics, in particular densities and compositions, but also some pipeline physical characteristics. The typical optimization parameters could range from choke opening and gas lift to plant operating temperatures. The matching process is performed coupling the proxy model with a differential evolution algorithm whose fitness function is an error function, to be minimized, that represents the distance between the actual field production parameters and the modeled ones. The algorithm finds the solution by varying the uncertain parameters and keeping fixed the known optimization ones. Once the model matches with the production data, the set of uncertain parameters is defined and fixed and the process of optimization can start, by changing the optimization parameters. This is achieved with a second differential evolution algorithm working with the same proxy model. The fitness function in this phase is the oil production, which must be maximized, having as constraint the maximum gas flow rate treatable by the plant. This tool has proven to be able to solve the problem in less than a minute. In this paper, the substantial advantages of substituting the physical model with a proxy one have been shown. With the novel approach, the computational time has been reduced by three orders of magnitude with respect to a classic method. This is a breakthrough because it allows the matching and optimization procedures to be done on a daily basis, which results in having always tuned model and a powerful tool to monitor online the field production.

Hybrid artificial intelligence techniques for automatic simulation models matching with field data and constrained production optimization

Grimaccia F.;Mussetta M.;Niccolai A.
2020

Abstract

Asset optimization has recently become a crucial issue in Oil&Gas industry, considering oil price conjuncture and an increased awareness on environmental aspects. In this paper, an Artificial Intelligence technique is presented, which is able to manage big datasets to automatically match the production models and propose an operative solution that maximizes the production. The tool is based on a hybrid in-house developed AI technique, integrating deep neural networks, biogenetical algorithms, commercial simulators and real-time data. The workflow starts with the modeling of the production system through physics based commercial simulators. A sensitivity analysis identifies the critical variables, which are then randomly varied with a Sobol distribution, exploring the entire solution domain. With these data, a proxy model to the commercial software is generated: it consists in an artificial neural network able to replicate the field behavior. Finally, the AI tool fed by real-time data is used firstly to match the field behavior and, successively, to maximize the field production. The tool has been developed considering an offshore oil field. The typical uncertain parameters used in the matching phase are fluid characteristics, in particular densities and compositions, but also some pipeline physical characteristics. The typical optimization parameters could range from choke opening and gas lift to plant operating temperatures. The matching process is performed coupling the proxy model with a differential evolution algorithm whose fitness function is an error function, to be minimized, that represents the distance between the actual field production parameters and the modeled ones. The algorithm finds the solution by varying the uncertain parameters and keeping fixed the known optimization ones. Once the model matches with the production data, the set of uncertain parameters is defined and fixed and the process of optimization can start, by changing the optimization parameters. This is achieved with a second differential evolution algorithm working with the same proxy model. The fitness function in this phase is the oil production, which must be maximized, having as constraint the maximum gas flow rate treatable by the plant. This tool has proven to be able to solve the problem in less than a minute. In this paper, the substantial advantages of substituting the physical model with a proxy one have been shown. With the novel approach, the computational time has been reduced by three orders of magnitude with respect to a classic method. This is a breakthrough because it allows the matching and optimization procedures to be done on a daily basis, which results in having always tuned model and a powerful tool to monitor online the field production.
International Petroleum Technology Conference 2020, IPTC 2020
978-1-61399-675-1
Artificial Intelligence, flow rate, machine learning, Upstream Oil & Gas, differential evolution algorithm, neural network, objective function, optimization problem
File in questo prodotto:
File Dimensione Formato  
Giuliani_IPTC_draft_v2.pdf

Accesso riservato

Descrizione: DRAFT
: Post-Print (DRAFT o Author’s Accepted Manuscript-AAM)
Dimensione 1.07 MB
Formato Adobe PDF
1.07 MB Adobe PDF   Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11311/1165117
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact