go to top scroll for more

Projects: Projects for Investigator
Reference Number	FT0555
Title	AFM83 Investigation & application of artificial intelligence techniques for optimum defrosting of refrigerated displays
Status	Completed
Energy Categories	Energy Efficiency(Residential and commercial) 100%;
Research Types	Applied Research and Development 100%
Science and Technology Fields	PHYSICAL SCIENCES AND MATHEMATICS (Applied Mathematics) 50%; ENGINEERING AND TECHNOLOGY (Mechanical, Aeronautical and Manufacturing Engineering) 50%;
UKERC Cross Cutting Characterisation	Not Cross-cutting 100%
Principal Investigator	Professor SA Tassou No email address given Sch of Engineering and Design Brunel University
Award Type	Standard
Funding Source	DEFRA
Start Date	01 September 1998
End Date	01 February 2002
Duration	41 months
Total Grant Value	£129,934
Industrial Sectors	Food and Drink
Region	London
Programme	DEFRA Resource Efficient and Resilient Food Chain
Investigators	Principal Investigator	Professor SA Tassou , Sch of Engineering and Design, Brunel University (99.998%)
	Other Investigator	Project Contact , Safeway Stores Plc (0.001%) Project Contact , Elm Ltd (0.001%)
Web Site
Objectives	To investigate the process of frost formation on display cabinet evaporator coils for both dairy/meat and frozen food products and develop artificial intelligence techniques for the optimum defrost initiation and termination, and sequencing of the defrost process of the cabinets.
Abstract	Frost forms on evaporator coils by water vapour in the air condensing and freezing when the surface temperature of the coil falls below 0ºC. Significant frost accumulation deteriorates the coil performance and thereby the refrigerating capacity of the evaporator. The overall aim of the project was to investigate the process of frost formation on display cabinet evaporator coils, both for dairy/meat and frozen food products and develop artificial intelligence techniques for the optimum defrost initiation and termination of remote and integral cabinets and the sequencing of the defrost process for remote cabinets. Specific objectives included: 1. quantifying the energy requirements for both electric and cool gas defrost by monitoring the energy consumption, store and cabinet conditions in a number of supermarkets around the country. 2. identifying and establishing, through monitoring, the effects of frosting and defrosting on the display cabinet and product temperatures 3. investigating experimentally in an environment test chamber at Brunel University the factors influencing frost formation in both dairy/meat and frozen food display cabinets 4. developing a demanddefrost method based on Artificial Intelligence (AI) techniques to provide optimum defrost initiation and termination, optimum cabinet defrost sequencing and display cabinet diagnostics. Optimisation was based on the minimisation of energy consumption and product temperature fluctuations during the frosting/defrosting process for both electric and cool gas defrost. To identify suitable controlparameters for the artificial intelligence (AI) controller, tests were carried out at controlled conditions in an environmental chamber in the laboratory, and field measurements in two supermarkets-one equipped with electric defrost and the other with cool-gas defrost. From the field measurements, which involved collecting and analysing environmental data and quantities of condensate collected after defrost, the environment within a supermarket was characterised. Tests in the laboratory were used to establish the performance of four different types of cabinet over a wide range of conditions and to monitor in detail the influence of frost formation on their performance characteristics. The data collected from the stored showed that, for open display cabinets, the store humidity is the primary parameter influencing the rate of frost accumulation. From the laboratory investigations it was concluded that the system deterioration caused by frost accumulation was best informed by the evaporator coil ‘air off’ velocity as an indirect indication of frost formation. Measurement of evaporator coil ‘air-off’ velocity as an indirect indication of frost formation could be agoodparameter to use in a demand defrost controller. Other parameters that influenced the rate of frost formation included space temperature and relative humidity, and air and coil temperatures at various positions in the cabinet. The data fusion technique was employed in the AI-based demand defrost controller that has been developed to predict the deterioration of the evaporator coil ‘air-off’ velocity, based on measurements of space temperature and humidity, refrigerant evaporating temperature, coil ‘air on’ and ‘air off’ temperatures and cooling time between defrosts. Predictions were performed using Multi-Layered Perceptron (MLP) artificial neural network (ANN) models which were trained and validated using laboratory test data for each cabinet. Inorder to implement thedeveloped control strategy, a software-based controller was developed using JAVA. The software-based controller is very flexible and has monitoring, prediction, control and fault detection and diagnosis (FDD) capabilities. Software implementation of the defrost control strategy was carried out on the serve-over delicatessen cabinet, the single air curtain cabinet and the frozen food cabinet in the laboratory. Also the implementation of the new defrost controller was successfully carried out by Brunel University with the help of Safeway Stores on an integral chilled cabinet at their Acton supermarket. The software code developed is now in the process of being implemented on a Honeywell-Elm controller, initially for a stand-alone cabinet, followed by implementation on multiple cabinetsin a store environment.
Publications	(none)
Final Report	(none)
Added to Database	29/10/09