The object of this research is the thermal and energy characteristics of medium-pressure boilers at the thermal power plant of a metallurgical plant operating on a common steam header. The fuel used is a mixture of blast furnace and natural gases. The volume fraction of blast furnace gas is 80–95 %. The characteristics of blast-furnace gas are not constant: the elementary composition, humidity and dust content change significantly.The mathematical model developed for the purpose of comprehensive optimization of the modes of joint operation of boilers requires adaptation of the energy characteristics obtained by calculation to real operating conditions. Adaptation of the calculated energy characteristics was carried out individually on the basis of thermal balance tests of boilers and taking into account the design and operational features of each boiler. During the tests, the fraction of the convective component of heat transfer in the furnace for the considered boiler units was determined using a non-stationary heat meter. The absolute value of the heat flux density was determined according to the theory of non-stationary regular heating of the sensitive element (SE). It was found that as the fuel burns out, the level of heat fluxes decreases, and with an increase in the thermal fraction of blast furnace gas, the absolute value of local falling fluxes decreases. The average share of the convective component is 15–20 %. It is proposed to adapt to the real operating conditions of the superheater and tail heating surfaces using empirical correction factors obtained during processing of the results of balance tests.Taking into account the proportion of convective heat flux in the furnace makes it possible to more accurately determine the parameters of the heat carriers along the steam-gas path of the boiler, and, therefore, it is possible to obtain adapted energy characteristics of the boilers. In comparison with analytical calculated dependences, the proposed method provides an individual approach to the operating modes of boilers. Thanks to this, it is possible to reduce the error in determining the optimal solutions to less than 1 %.