Performance of integrated R744 packs part 2 : ejectors performance, a comparison of data-driven model from onsite measurements with rom model predictions
Peer reviewed, Journal article
MetadataShow full item record
Original version10th IIR Conference on Compressors and Refrigerants
Rising concerns for climate change impacts along with the new legislation aimed at lowering emissions indicates the inevitable transition in the cooling/heating industry towards a more energy-efficient solution with minimal environmental impact. Integrated refrigeration, air condition and heat recovery solutions by CO2 (R744) packs are efficient solutions to high energy demanding building (supermarkets and hotels). Proven performance enhancement of Multi Ejector SolutionTM makes the R744 systems more energy efficient, especially in warm climates, compared to the most conventional synthetic refrigerant systems in food retail applications. Pilot installation in the frame of MultiPack, an EU funded project(Horizon 2020), provided a wide range of data, offering the possibility of evaluating the real performance of each ejector group: high pressure (HP), low pressure (LP), and liquid ejector(LE), by running the system in different operating modes. Analysis of the data indicated performance improvement of a system with ejector, on average a 35% lower energy consumption compared to baseline parallel compression R744 system without ejectors. The availability of mass flow measurement from 5 Coriolis mass flow meters on this pilot, enabled the possibility of comparison of mass flow rate in different evaporation temperatures. Data-driven models were used to estimate the performance of ejectors (Entrainment ratio) using onsite data from various operating modes and compare them with the simulated performance from CFD and lab measurements. Careful data pre-processing allows the data-driven model to predict the ejector performance using compressor mass flow rate models trained based on no ejector mode data derived from onsite measurements and compare with models validated by highly accurate lab experiments.