Performance analysis of deterministic, min-max and multi-stage NMPC applied to a subsea gas compression system

Abstract

In recent years, subsea technologies have become more reliable. New projects emerged and an increase of investment in the area is expected in the following years. However, subsea installations are becoming more complex and smart decision making controllers are necessary to reduce operational costs and increase process reliability. Besides, the presence of uncertainties makes the development of controllers that can handle operation in an uncertain environment imperative. A prominent subsea technology is the subsea compression, which has been recently delivered in the North Sea. Åsgard eld compression system design was based on topside design. Therefore, surge avoidance strategy is necessary in order to operate without major issues in presence of disturbances and uncertainties. If surge occurs, compression system operation is strongly a ected, leading even to compressors breakage. Thus, in this work, Nonlinear Model Predictive Controls (NMPC), such as deterministic, o ine min-max and multi-stage were employed to a subsea compression system to evaluate controllers performance and closed-loop robustness in an environment with unknown disturbances a ecting upstream pressure. For performance assessment, indicators that consider set-point tracking, constraint violation, gas production, energy consumption, and production e ciency were employed. Deterministic NMPC was the most e cient controller, but constraint violation was detected. Although o ine min-max operation managed to handle constraint violation, it proved to be overly conservative. Multi-stage NMPC controller was able to also handle constraint violation, while being less conservative than o ine min-max NMPC.