The project’s first phase successfully transformed the complex HVAC network generation problem into a constrained quadratic model (CQM), potentially efficiently solvable by D-Wave’s quantum-classical hybrid solvers. This marked a clear step away from traditional computational methods to a more innovative, quantum-classical hybrid approach.
The implementation and experimentation phase saw the CQM translated into Python code and processed by D-Wave’s hybrid solvers. These solvers, utilizing using both classical and quantum devices, identified superior HVAC network designs, outperforming the existing data-driven method in significant ways. Notably, the new approach yielded solutions more quickly with shorter duct lengths and fewer construction elements, such as elbows. In addition, subject matter experts from VINCI Energies visually and manually inspected these solutions, confirming their superiority in terms of quality. Importantly, these results were based on one representative building plan and diverse HVAC systems, setting the stage for broader application in the industry.
