Demonstration of Mealy machine approach in simulated floor heating control

Demonstration of Mealy machine approach in simulated floor heating control

Floor heating control is demanding due to the thermal lag caused by massive structures of the floor. A new floor heating control (Fig. 1) was demonstrated by using Mealy machine approach (Fig. 2) in modelling the thermal behavior of the heating system and the indoor environment. Test runs show that the state machine control damps down disturbances of indoor temperature more effectively than the conventional method.Näyttökuva3

Results

The state machine approach can be characterized as a data-driven or model based control. Contrary to an ordinary parametric model, the state machine model consist of collected and modified data, created in the memory of the control system. Before using the model, identification is needed, which is based on measurements taken from the heating system, indoor and outdoor environments.

The control method was demonstrated in a simulation environment, consisting of a small office building equipped with a water circulated floor heating in a concrete slab. Inner and outer loads as well as occupation of the building was scheduled for each day and week of the year.

The state machine control was compared to the conventional method, where the temperature of the circulating heating fluid is linearly dependent of the outdoor temperature. The results show that stability of the indoor temperature was better and energy consumption decreased when the new model was used. The effect of thermal lag, caused by massive structures, can be decreased as well, but this needs a weather forecast and correct timing of control actions. Cooling of the building before the start of solar radiation decreases rising of indoor temperature and lowers energy consumption.

Näyttökuva4

Effects

The new techniques makes the modelling of heating process flexible. The effect of each input variable can be properly designed and details of the model is easy to modify even after the identification. Using the new approach, identification becomes a robust procedure. Probably, data-driven control will be soon applied in real building automation and heating control systems.

Future Opportunities

The method was successfully demonstrated in a simulated environment. The approach and its details still needs some development before applied in practice.

RYM Program

Indoor Environment

Participants

Aalto University

Contact Person

More info

e-mail: jouko.pakanen@aalto.fi

SHARE