Automatic Optimisation of Infinite Variable Control Valves

Abstract

TX 75083-3836, U.S.A., fax 01-972-952-9435. AbstractThis paper discusses the application of an optimisation technique to manage a field consisting of wells with Intelligent Well Technology (IWT) completions that allow the operator to reconfigure well architecture without well intervention. There is currently no generally accepted, efficient, optimisation technique available to optimise the IWT operation so as to maximise well production. The Sequential Quadratic Programming (SQP) candidate technique has previously proved its value in small and large-scale optimisation problems involving large-scale (gas-lifted) well production networks. SQP's ability to solve this new nonlinear problem in a relatively reasonable time will be explained in this paper.In the previous publications 1-4 we used a reservoir simulator controlled by a manual optimisation procedure to develop a value-statement for IWT. This method has proved it could increase value during the plateau period, but efficient optimisation of the decline period was too demanding in terms of engineering time & computer power. This later requires an automatic optimisation tool to be cost effective.The Reservoir simulator has been linked to a (SQP based) Network optimiser so as to set the well and zone so that the production objectives are maximized, subject to the normal constraints (e.g. pressures, flow rates, water cuts etc. frequently imposed on the individual zonal or well production).The cumulative production increase achieved with manual optimisation (compared to the field level control) was doubled once automatic optimisation with zone-level control was implemented for the real-field based example studied previously. Automatic optimisation did significantly improve tail-end production, but delivered its greatest value by extending the plateau period.The results show that the use of automatic optimisation is valuable when building a "value statement" for the implementation of IWT within a particular field. Automatic Production optimisation (using conventional well) in reservoir simulationProduction and Injection wells are usually connected together by a surface network contain pipes, chokes, pumps, etc. A number of studies have been performed in order to set the flow rate per well or per group of wells, different tools have been used to meet constraints inherent to this network of these facilities. Reservoir modellers may optimise rate allocation to optimally determine well rate setting so that network objectives and constraints are simultaneously satisfied.Literature examples previously reported in this area include:1. Davidson and Beckner 5 used SQP to set well rates in a facility network of a reservoir simulator. SQP was used to meet the objectives of maximizing the oil production subject to constraints on pressures, flow rates and stream compositions. The approach was tested on black oil and compositional examples. The examples show how the method has been formulated, individual well rates set and how infeasible conditions can be handled.2. Lo et al. 6 ...

Successful Application of a Robust Link To Automatically Optimise Reservoir Management of a Real Field

“…The oscillation of the automatically optimised results is a persistent challenge that has been identified by several authors 2,16,20 . In our, case, the oscillation was partially attributed to the well bottomhole reference depth mismatch between the models which was rectified later, and partially due to the none convergence of the network solver causing optimiser fluctuation between choking alternatives in an attempt to identify the best possible solution.…”

“…The S-Field reservoir model was thus transferred from the old reservoir simulator (Supplier 2) to the new reservoir simulation software (Supplier 3) to take advantage of this new opportunity 2 . As expected, this shift produced different values in term of the Oil-Originally-In-Place and the cumulative oil production due to the differences in the calculation methods employed by the two reservoir simulators.…”

“…However, not only are there differences in the degree of coupling between the individual software programs, but these links usually place high demands in terms of computing power, network architecture and, frequently, manual intervention of the engineer. In this paper we will illustrate a successful application of a robust and efficient linking tool developed to couple a commercial reservoir simulator with a surface network simulator and Sequential Quadratic Programming (SQP) optimiser 1,2 . The S-Field (a case study based on redevelopment of a real field with an Advanced or Intelligent Well development strategy) has been previously studied indepth.…”

“…The S-Field (a case study based on redevelopment of a real field with an Advanced or Intelligent Well development strategy) has been previously studied indepth. It has been found to be a very useful case history to illustrate the potential advantages of implementing an Advanced Well development scheme 2,3 . We have been able to show how the SQP optimiser can be used to increase the recovery compared to manual optimisation.…”