Polydome

General description

The Polydome platform comprises a single-zone large conference building on the EPFL campus accounting for 400 sq. meters. The building is characterized by a relatively lightweight structure with a limited glass surface. During both heating and cooling season, Polydome is served by a ~20Kwp air-to-air Heat Pump (HP).

Hardware description

Indoor Measurements

The conference room has been equipped with three Aoetec Z-Wave Multisensors that monitor indoor temperature, presence, humidity, and lightning. All measuremnets are sent through the low-power protocol, Z-Wave, to the central database.

Weather Forecast & Record

Weather information are retrieved from an external website, called Wunderground, by means of a standard API. Obtained data comprise both recorded measurements as well as future predictions for both outside temperature and solar global irradiance.

Actuation

The overall energy consumption of the building is composed of standard equipment, lights, and the HVAC system. While the first two elements can be predicted, in our study, we focus solely on the latter as it represents the only controllable element.

The HVAC system is characterized by an air-to-air HP used for both cooling and heating purposes. The HP is characterized by the presence of a ventilation unit, and two SCROLL compressors in an 'uneven' tandem configuration. While the ventilation unit is always active, the internal proprietary controller of the HP determines, depending on the current conditions, whether to activate one or the other compressor. In some extreme situation, both compressors can be activated accounting for a peak power of roughly 12kW.

To measure the instantaneous power consumption of the HP we installed an EMU Professional TCP/IP smart meter which allows us to monitor both active and reactive power at a second granularity. All the measurements are sent through standard TCP/IP communication to the central database.

Database

All data are sent to a central node which takes care of storing them in a dedicated database. The database is implemented in InfluxDB which is an open-source time series database developed by InfluxData. Finally, for visualization and monitoring purposes, we utilize the Grafana platform.

System Peculiarities:

The Polydome platform represents a perfect candidate for testing our control method. In fact, it displays all the main challenges that one encounters when attempting to control a building.

The HVAC system displays both non-linear and switched dynamics,
Limited control authority is possible as only a few internal parameters can be observed and/or modified,
The building is highly occupied one with an unpredictable schedule.

Results:

Normal operations

https://snapshot.raintank.io/dashboard/snapshot/Ry4CDQudwhu1eY9iXT4EA7AqXWfGmBI0

https://snapshot.raintank.io/dashboard/snapshot/3Qy1cRSPoOwzC2zRFx1ubbEkvpfEoWkX

Identification experiments

https://snapshot.raintank.io/dashboard/snapshot/Ykhtitx7aHYp9dbeQuDzFKTOQeF0Mgu7

https://snapshot.raintank.io/dashboard/snapshot/k2IROdcG8o7DemFgSi8mDUJNQ4yn0D33

References:

Google Sites

Report abuse