The threat of global warming is driving the increase of renewable energy in the electricity mix to reduce carbon emissions. However, renewable sources such as wind and solar energies introduce a fluctuating feed-in power to the grid due to their inherent variability, requiring compensation from other sources such as thermal power plants; this will result in a more dynamic operation of the latter. Monitoring dynamically operated power plants necessitates sophisticated methods and reliable digital twins. Suitable modeling strategy allowing simple calibration and good representation of the dynamic behavior is the major challenge of this work. The modeling methodology aimed as an outcome of this study will be used to build and calibrate heat recovery steam generator (HRSG) models based on a small set of sensor measurements, that can realistically be found in a power plant. Dynamic models of thermal power plants have been developed to ensure safe and reliable operation under fluctuating load conditions. Heat exchangers in HRSGs have a delayed response to load variation, primarily due to their high thermal inertia. Accurate models require extensive information on the system's design, which may not always be available for the modeler. This study assesses the impact of modeling assumptions on the simulation results, with the aim of minimizing the information requirements and simplifying the calibration process of heat exchangers models of a full industrial HRSG. First, fully detailed, pseudo three dimensional Modelica model of the heat exchangers are developed. This model is then simplified progressively. First, the impact of the tubes’ discretization in all directions and the averaging of flue gas temperature between heat exchangers on both steady-state and dynamic responses are evaluated. The time constants are computed to observe the response delay of a train of heat exchangers. Finally, a more simplified model is developed by assuming a lumped global heat transfer coefficient and a mass. A sensitivity analysis is conducted on the ratio between water and flue gas heat transfer coefficients. The results indicate that the discretization along the length of the tubes and the radial temperature distribution in the tube’s wall have the greatest impact on the response. The simplified model provides promising initial results, demonstrating a good compromise between simple calibration process and accuracy.