Key points are not available for this paper at this time.
Abstract This work discusses optimizing the operations of a complex gas-oil separation plant (GOSP) network. The objective is to operate this complex network such that the required oil target is produced at minimum OPEX while leveraging the transfer capability between GOSPs. A mixed integer non-linear programming (MINLP) model is developed to optimize swing line production and processing equipment operation. This allows the systematic identification of optimal operating points, which optimizes operations. Consequently, it will result in further reduction in the processing cost and greenhouse gas emissions associated with power generation and flaring. Many GOSPs exist to process crude production from oil wells to separate the multi-phase produced fluids into oil, gas and water. These plants include equipment, which are highly energy-intensive such as high-pressure gas compressors and water injection pumps, which are used to boost separated fluids to their final destinations. In addition, transfer lines might exist between facilities allowing for shifting production or part of it from one plant to another. As a result, there is an opportunity to optimize crude oil distribution among plants to improve asset utilization and minimize power consumption. This is in addition to the benefits of reducing greenhouse gas emissions by an average of 10% based on calculations using an MINLP model and Aramco's best practices for optimizing crude oil handling operations. The paper proposes use of parameter generation techniques to improve the model's prediction using data analytics, thereby delivering a digitalized fit-for-purpose application. This results in minimizing energy consumption while maintaining the oil target without added investment (neither OPEX nor CAPEX). Consequently, it will result in further reduction in the processing cost and greenhouse gas emissions associated with power generation. This paper proposes a novel methodology to formulate and achieve a desired optimization solution. It also describes the level of fidelity used to model physical process equipment. This varies between use of detailed first-principles models in certain equipment to a more simplified representation elsewhere. This is done systematically based on the overall impact on the solution's accuracy and robustness.
Alhuraifi et al. (Mon,) studied this question.