A stochastic optimization model for short-term production of offshore oil platforms with satellite wells using gas lift
- PDF / 2,206,791 Bytes
- 26 Pages / 439.37 x 666.142 pts Page_size
- 96 Downloads / 188 Views
A stochastic optimization model for short‑term production of offshore oil platforms with satellite wells using gas lift Carlos Gamboa1 · Thuener Silva1 · Davi Valladão1 · Bernardo K. Pagnoncelli2 · Tito Homem‑de‑Mello2 · Bruno Vieira3 · Alex Teixeira3 Received: 14 December 2019 / Accepted: 10 February 2020 © Sociedad de Estadística e Investigación Operativa 2020
Abstract Continuous gas lift is a popular method to enhance productivity in offshore oil platforms. We propose a steady-state two-stage stochastic programming model to maximize production, where the first-stage injection level determines the production potential, while recourse actions ensure capacity and platform constraints for each uncertainty realization. In particular, we develop a concave approximation of the performance curve that incorporates uncertainty in the water cut (WC) and gas–oil ratio (GOR). We generate WC and GOR realizations using a two-step data-driven approach: we extrapolate the trends using a 𝓁1-filter, and bootstrap historical deviations to generate future realizations of WC and GOR. We present numerical results for the sample average approximation of the problem and assess the solution quality using standard techniques in the literature. Our numerical results suggest that taking uncertainty into account in the problem can lead to considerable gains. Keywords Oil production optimization · Continuous gas lift · Two-stage stochastic programming · Sample average approximation Mathematics Subject Classification 97M50 · 90C15 · 62M10
1 Introduction Producing oil is a hard endeavor that involves several disciplines, such as automation, reservoir engineering, optimization methods, among others. The challenges are even more significant in offshore fields, which represent most of the production in * Carlos Gamboa [email protected]‑rio.br 1
Industrial Engineering Department, PUC-Rio, Rio de Janeiro, Brazil
2
School of Business, Universidad Adolfo Ibáñez, Santiago, Chile
3
CENPES, Petrobras, Rio de Janeiro, Brazil
13
Vol.:(0123456789)
C. Gamboa et al.
Fig. 1 Life cycle optimization, adapted from Foss and Jensen (1981)
some countries such as Brazil. A common practice in the oil industry is to decompose the problem into different layers, which represent different timescales, as shown in Fig. 1, adapted from Foss and Jensen (1981). The focus of this work is on short-term optimization. As asserted in Foss et al. (2018), despite the dynamics involved in the process, most problems can be solved using static models. It is only during well start-ups and shut-ins, or in shale-gas wells that we observe a dynamics in which abrupt changes occur. Many offshore oil platforms make use of gas lift to reduce the density of the fluid and enhance productivity. Under the steady-state assumption, previous works estimate the production response to gas injection via numerical fluid simulator and interpolation techniques. The existing approaches aim to obtain the injection level that maximizes oil production by embedding a deterministic perf
Data Loading...
