The efficiency of heat transfer in air-cooled heat exchangers of various industrial facilities depends on the flow rate of the coolant, its inlet temperature and ambient temperature. These parameters are transient and depend both on the features of the technological process and on weather conditions. One option for a compact design of heat exchangers is the use of close-packed coils with a small bending radius. In this case, heat transfer in the complex geometry of the annular space cannot be described by simple one-dimensional dependencies. To solve this problem, it is necessary to consider the three-dimensional spatial structure of the heat exchange surface. Since the size of the grid elements will be several orders of magnitude less than the size of the facility, the size of the computational grids for CFD modeling full-scale heat exchangers will be billions of finite volumes, and even on powerful supercomputers, the solution time will be about a month. One way to reduce computational costs is to use reduced order models, in which the computational domain is not modeled directly; instead, simplified models, such as a porous medium model, are used to describe it. However, such models require additional closing relations and coefficients that characterize the actual channel geometry. This paper presents a technique for creating a digital twin of a heat exchanger with small bend radius coils based on a porous medium model. The values of heat transfer coefficients and hydraulic resistance depend on the speed of air movement in the space between the coils. The calculated value of the thermal power obtained using the strengthened model was 529 kW, which corresponds to the passport data of 500 kW, with less than 6% deviation for the heat exchanger under study. This confirms the correctness of the calculation with accepted simplifications. The calculation time in this case was only a few minutes when using a personal computer. The developed numerical model allows for the resolution of performance characteristics based on the temperature of the cooled medium at the inlet, air temperature, and fan speed. Analyzing the different modes of turning on the cooling fans made it possible to determine the values of the thermal power when turning off the fans or reducing the number of revolutions.