Applications of hybrid models in chemical, petroleum, and energy systems: A systematic review

Mathematical modeling and simulation methods are important tools in studying various processes in science and engineering. In the current review, we focus on the applications of hybrid models (HMs) in chemical processes, oil and gas processes, and applied energy systems. These processes suffer from complexities that demand advanced mathematical tools to be employed for various purposes such as process development, identification, simulation and modeling, optimization, control, classification, clustering, forecasting, and monitoring. Beside the lack of adequate knowledge about the processes related to the chemical and energy systems, there are other mathematical complexities such as non-linearity, large and multi-scale, long dynamics, uncertainties, and high dimensionality. The HMs can provide a practical solution to such complex processes. The arrangements of different black-box models with themselves or with white-box models are alternatively used to reduce the model complexities. The hybrid gray-box models are of significant importance as they combine the physical significance and generalization capabilities of the white-box models with the complexity reduction capability of the black-box models to facilitate/enhance the modeling strategy, while a desired precision is targeted. Such a hybridized model enables the physically-meaningful computation for the time-demanding applications. In this paper, we review different sub-models, hybridization strategies, structural designs, screening criteria, and new directions in hybrid modeling, with focus on the corresponding applications in chemical, petroleum, and energy systems.