Performance optimization of a CO²
thermally driven heat pump application
for domestic heating
based on modeling and control

Abstract

This PhD thesis focused on developing advanced modeling and control strategies for a CO₂-based Thermal Compressor Heat Pump (TCHP), a thermally driven system proposed as an energy-efficient alternative to conventional gas boilers. The work first investigated the thermal compressor (TC) through a review of Stirling-type machines, an experimental campaign on a single-stage heat pump, and a finite-volume model that quantified exergy destruction and identified optimal operating conditions. Because this detailed model is too slow for dynamic simulations, several data-driven models were developed to provide fast and reliable predictions. The analysis was then extended to the complete three-stage TCHP cycle, combining experimental measurements with a dynamic FV model and reduced-order recurrent neural network representations of the system. Building on these tools, a model predictive control (MPC) strategy was designed to enhance the overall performance of the TCHP. The approach and its benefits were validated using both component-level and system-level experimental data.

Jury members  

• Prof. Pascal STOUFFS, Université De Pau Et Des Pays De L'Adour, France, President
  under joint supervision with
• Dr. Pascal DUFOUR, Université Claude Bernard Lyon 1, France, Co-promotor
• Prof. Nadri, Madiha;  Université Claude Bernard. Lyon 1, France, Co-promotor
• Prof. Vincent LEMORT, Université De Liège, Belgium, Co-promotor
• Prof. Emilio NAVARRO-PERIS, Université Polytechnique de Valence, Spain
• Prof. Melaz TAYAKOUT-FAYOLLE, Université Claude Bernard Lyon 1, France
• Prof. Ali ZEMMOUCHE, Université De Lorraine, France
• Prof. Gregoire LEONARD, Université De Liège, Belgium
• Dr. Laurent BAKO, Université De Lille, France

Videoconference available on Teams