Optimisation of gas-lifted system using nonlinear model predictive control.

• Ano de defesa: 2023
• Autor(a) principal: Adukwu, Ojonugwa
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: Casing-heading instability; Controle preditivo de modelo não linear; Controle tolerante a falhas; Fault-tolerant control; Gas lift; Instabilidade do cabeçote do revestimento; Model predictive control; Optimisation; Otimização; Sistema de elevação a gás
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/3/3139/tde-26042023-151430/

Gas-lifted system like every other artificial lift system is used when the natural energy for lifting crude oil from the reservoir into the downstream facilities becomes insufficient. This research focused on optimising crude oil recovery from gas-lifted oil well by using nonlinear model predictive control (NMPC). Two key approaches were used: (a) casingheading instability reduction/elimination and (b) fault-tolerant control in the system. At first a developed nonlinear model predictive controller (NMPC) was presented. The controller was tested on continuous stirred tank reactor (CSTR) using IPOPT solver in CasADi and fmincon optimizer in MATLAB. Finite horizon NMPC was selected and used to optimise the gas-lifted system. The controller stabilised the undisturbed system improving production by 5.63% compared to the open-loop operation when the system is in casing-heading instability. For the two input case, the steady state production, aided by the high input target, reached 12.25kg/s which is far more than 9.57 kg/s for the one input case. This controller showed a 3.76% improvement over PI controller for the same purpose. Estimation performances of three nonlinear filters were compared and Extended Kalman filter was selected to provide the estimated states of the system which were used for fault-tolerant control of the gas-lifted system. Passive FTC, altering control bound and altering control cost were used to implement the FTC problems. Passive FTC provided more robustness but small output change. Reducing the upper control bound ensured stability but production could decline. Increasing the controller cost that prioritised the input target increased production but it was prone to casing-heading instability. While the FTC scheme could reduce the downtime, the casing-heading instability removal increases the average oil production rate hence optimising the gas-lifted system.