Output details
15 - General Engineering
Coventry University
A Learning Feed-Forward Current Controller for Linear
Reciprocating Vapor Compressors
This paper presents research for a Carbon Trust Project (Low cost energy efficient compressors for refrigeration, £197,258). Industrial collaborators included Magnet Applications Ltd, Höganäs AB, Indesit Company UK and Danfoss Compressors GmbH (Poul-Erik Hansen, R&D Director, p.e.hansen@danfoss.de). An advanced neural network based control algorithm was successfully applied to this industrial application. The compressor prototype was tested on Danfoss R&D laboratory’s standard compressor test Calorimeter, and the testing results shown improved performance compared with conventional control algorithm. Part of the project research outcome was incorporated by Danfoss in several patents (US20070040456, US20070041855, DE102005038780B4, and CN1916409).