Dynamic optimization of heat exchanger using entropy generation minimization
Résumé
In order to reduce fossil fuel consumption for heat and cold generation, different low
temperature heat sources have been considered for heat pump operations such as sewage
systems, sea water, ground water and drinking water. Raw water systems can also be
used as a heat source for heat pump, heat sink for chiller, or for direct cooling. These systems
represents a high thermal potential for some clients, according to their activity. In this study,
the thermal potential of a raw water system (made up of 5000 km of piping and carrying 200
million m3 annually in the south of France) is investigated. To this day, this source has not
been used. The main problem is the optimisation of the equipment sizing according to the
temporal variability of water flow, temperature and heat (or cold) demand. The dynamic
behaviour of heat exchanger between network and user is the key parameter to enhance the
performance of the system.
For the raw water source, temperature, flow, and heat demand are highly time dependent.
The method is based on the minimization of entropy generation in the heat exchangers
between the water pipes and the users. The optimisation of the sizing and the location indicate
where to install direct heat exchange systems. In order to calculate entropy generation in the
heat exchanger, a dynamic model of a simple heat exchanger (concentric tube) is developed.
It determines the pressure and temperature profile along the heat exchanger according to time.
Temperature is calculated for water and for the heat exchanger partition wall. By integrating
the terms of entropy generation due to temperature difference and to pressure drop along the
heat exchanger, the value of total entropy generation is obtained under transient operation.
Total entropy generation is used as an objective function to optimize heat exchanger geometry
(length and diameters) and control mechanism.
The method has been used to model direct cooling of a datacenter using a raw water
network as a heat sink. Optimization is done considering a year of operation; seasonal and
daily fluctuations for water flow and temperature are considered. Results show that entropy
gain is important when the heat exchanger optimal size is chosen. The use of the raw water
network connected to reversible heat pump for building heating and cooling have also been
studied and achieves high gain compared to air-source heat pump.