Learning predictive control for GHP [gas heat pump]

The GHP (gas heat pump) is the attractive air-conditioning system that realizes energy savings and low emissions, from an ecological viewpoint. However it has some difficult characteristics to be controlled, that is the system response is very slow and the dead time is large because of the large system. In addition, system dynamics are easily changed by the installation and the ambient conditions. Therefore it is difficult to adapt to these changes by conventional control methods. In this paper, the authors propose a learning predictive control scheme, which can control the variable system, identifying the system dynamics in an online process. The authors show the effectiveness of the proposed algorithm through the computer simulations and experiment