Grey box modeling of supermarket refrigeration cabinets

Leerbeck, Kenneth; Bacher, Peder; Heerup, Christian; Madsen, Henrik

doi:10.1016/j.egyai.2022.100211

Cited by 6 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to fit model parameters from historical data, the continuous state space model (7), should be discretised using the sampling time of the sensors.…”

Section: Cabinet Temperaturementioning

confidence: 99%

“…This limits the application of the method to control of real-world systems. An alternative grey-box modelling approach is presented in [7], which uses a state-space model derived from thermodynamic heat and mass balances of individual cabinets and evaporators. Model parameters were estimated using a maximum likelihood estimation approach, based on 2.5 years of minutely telemetry from 6 chilled and 4 frozen units.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Data-Driven Modelling of Supermarket Refrigeration Systems for Model Predictive Control Applications

Bird

Acha

Escrivà

et al. 2023

36th International Conference on Efficiency, Cost, Optimization, Simulation and Environmental Impact of Energy Systems (ECOS 20

View full text Add to dashboard Cite

With uncertainty in energy markets, and the effects of climate change looming, reducing energy use and operational cost of existing building systems is more important than ever. To this end, this paper presents a grey-box modelling approach to characterise the behaviour of chilled and frozen and coldrooms using basic system specifications and measured data. An overall energy balance is used to devise a discrete state space model for each cabinet, characterised by unknown empirical parameters relating to heat capacity and heat transfer properties. Historical system data from a UK supermarket are used in combination with a genetic algorithm optimisation to determine the optimal empirical parameters for 10 display cases and 10 coldrooms. The resulting cabinet temperature predictions have a good level of accuracy, achieving a root-mean squared error (RMSE) of 0.37°C to 0.98°C. Overall this data-driven approach is effective and efficient in modelling refrigeration systems, and can be easily generalised to any system where historical data is available. Finally, the use of the proposed approach in cost minimisation or demand response application is presented.

show abstract

“…In order to fit model parameters from historical data, the continuous state space model (7), should be discretised using the sampling time of the sensors.…”

Section: Cabinet Temperaturementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%