Combined Physics-Data Driven Modeling for Design and Operation Optimization of an Energy Concept

Abstract

The industrial sector accounts for a huge amount of energy- and process-related CO2 emissions. One decarbonization strategy is to build an energy concept which provides electricity and heat for industrial processes using combination of different renewable energy sources such as photovoltaic, wind turbine, and solar thermal collector system combined with energy conversion power-to-heat components such as heat pump, electric boiler etc. The challenge for the industries is the economic aspect of the decarbonization, as industries require a cost-efficient solution. The total cost for an industrial energy concept includes investment and operating costs. This complex problem of minimizing cost and emission requires two major tasks: (I) modeling of components and (II) multi-objective coupled design and operation optimization of the energy concept. The optimal design and capacity of the components and optimal system operation depend majorly on the modeling of the components. The modeling of the components is either physics-driven or data-driven. The corresponding multi-objective coupled optimization is a complex problem with a large number of variables and constrains involved. This paper shows different types of physics- and data-driven modeling of energy components for the multi-objective coupled optimization for minimizing cost and emission of an industrial process as a case study. The optimization problem is solved as single-level problem and bi-level problem with different combinations of physics- and data-driven models. Different modeling techniques and their influence on the optimization are compared in terms of computational effort, solution accuracy and optimal capacity of components. The results show that the combination of physics and data-driven models have computational time reduction up to 37% with high accuracy compared to complete physics-driven models for the considered case study. Specific combination of physics-driven and polynomial regression models show the best trade-off between computational speed and accuracy.