Smoke and Heat Exhaust Ventilation Systems (SHEVS) are technical solutions used to limit the consequences of a fire in a building, related to the spread of the smoke. Currently, the design methodologies follow an assumption, that if a system constantly exhausts amount of the smoke estimated with a calculation method based on an onerous steady-state fire scenario, the building should be safe. This paper introduces a different approach to optimal system design, based on the transient characteristic of the fire. The main idea is that a system can adapt to the momentary density of the removed smoke, to benefit from the thermal expansion of gases in the fire. As the fire grows, the temperature does rise, and the pressure within installation falls down, and so do the forces acting on ducts, dampers and other elements. This behaviour is observed in high-temperature furnace tests of exhaust fans, during which the change of pressure and power-supply requirement in changing temperature can be measured. The author presents results of 9 high-temperature tests that are a proof of the concept. The practical implementation of the idea presented in this paper could mean, that a system designed with an existing methodology for ambient conditions, could work with a higher capacity in high-temperature, without additional strain on the elements of the system. This shift in thinking allows using higher capacity systems, in place of currently used, by artificially limiting the capacity of an oversized fan in ambient conditions and increasing it, following the measurements of the temperature of the exhausted air or pressure in the shafts. Beside the theoretical introduction to a new concept, the paper presents results of 8 numerical analysis (CFD) of airflow in an enclosed car park during a fire, performed in ANSYS (R) Fluent (R) solver. The author created an User Defined Function (UDF) to automate the transient change of the fan boundary condition, dependant on the exhausted smoke temperature - following the assumptions of the adaptive solution presented in the paper. Four CFD analysis were performed for each traditional and new solution, and their results were compared in with qualitative and quantitative approach. Results of the CFD analysis show a possible gain of 25-41% of system capacity, using the same ductwork and reaching the same design goal, as contemporary SHEVS. The pressure within the ductwork and at fans is almost constant in the adaptive analysis. The paper is closed with a discussion of legal aspects, possible limitations in the design and the further research necessary to establish the new method of the design. (C) 2017 Elsevier Ltd. All rights reserved.