The Leiden Pilot will be located in the Leiden City Hall. The municipality offers the Save Energy Project to use 12 office rooms for the experiment. In each room there will be 1 to 4 persons working at a desk. Personnel changes (moves) should be avoided during the period the pilot is running as we are measuring changes in personal behavior. Data on absence (e.g. part time workers, holidays, leave) will be required.

The light in the rooms can be controlled by two (sometimes four) switches both for two sets of fluorescent tubes at the same time. Rooms have venetian blinds that can be opened/closed manually.

The rooms have two (double-) windows that can be opened or closed by the employees and two (hot water) radiators for heating that the employees can adjust themselves. Most rooms have two radiators. One has a thermostatic control valve, one has a manual valve.

The ventilation system can not be adjusted. The rooms are ventilated by blowing outside air into the rooms via a central system at a constant rate. This ventilation system is rather old (dates from the 1940′s). In summer a few degrees cooling is obtained by ground water cooling (up to 40 m3/h is used). There is also humidification of the air (in winter).

We want to measure both the electricity usage and heating/cooling/humidification required in each room.

• Electricity is used by the computers, screens, lighting and other room-specific devices like coffee machines or printers. This measurement is straightforward.

• The energy supplied by the hot water radiators can be calculated from the radiator surface temperature and room temperature or by measuring the water flow and in/outlet temperatures.

• The amount of cooling/humidification energy can be estimated from the entrance air temperature/humidity, the outside air temperature/humidity, and the room temperature/humidity (exhaust air). For rooms at the ‘inside’ (i.e. the atrium) an additional measurement of the atrium air temperature will be required.

This initial setup measures the total energy usage of a room. The system should be simple, reliable and cost effective. The number of sensors should therefore be minimized.

Employees can influence the energy consumption of their room by:

• switching lights or electrical equipment on/off

• adjusting radiator control valve and/or thermostatic valve set point

• open/close windows

• open/close door(s)

• adjust venetian blinds

During the pilot we may decide, in consultation with the employees, to place extra sensors to see if extra information might help the user to save more energy. It is for example possible to measure whether a window is opened or closed using an extra sensor. In this way it is possible to communicate to the user of a room that the heating is on and a window is open simultaneously. Examples of sensors that can provide extra information to the users are:

• infrared sensor to indicate someone is in the room

• window open/closed sensor

• door open/closed sensor

• day light sensor: is there sufficient daylight to switch off the fluorescent tubes?

• venetian blinds sensor for amount of radiation entering room

A panel for easy, remote control of appliances may be provided.

In the initial setup we will install sensors for the computer screens, the computers, lighting, heating and temperature. If a room has some special equipment like a private coffee machine, then an extra sensor will be added to this room. We use sensors of the company Plugwise of Sassenheim. The data that is measured by the sensors will be collected by a Plugwise concentrator that can communicate with an AnyWi router using an open protocol.

The AnyWi router connects via Internet to the SAVE ENERGY infrastructure which is outside the scope of this Leiden Pilot project. This infrastructure consists of a Middleware Layer, where the Business Logic is applied to the measured data and several clients such as the Serious Game, Portal or ProcessVision that use the data from the pilot.

The data is processed and presented in close-to-real time to the users. This part of the system is still under development. In the meantime a direct simple feedback to the users will be provided e.g. on a local screen.

The following image shows a graphical representation of the Leiden Pilot Setup.

As we can see from the image we use the ISA middleware layer that allows many clients and the local gateway at the pilot to connect. The local gateway at the pilot, or in the image, the AnyWi Router, has the knowledge of and connections to the sensors in the pilot. The sensors can be obtained e.g. from Plugwise (electrical energy sensors) and ISTA (heating energy sensors). They are all seamlessly integrated in the pilot using an open XML-protocol. Plugwise can also provide a simple system for direct feedback to the employees.

The measured data will be stored locally and sent via Internet to the Save Energy Project system. Close-to-real time data will be available locally to participating employees and the local repository can be used for data analysis.

Before deciding on the experimental setup we will determine which parameters should be measured, how long/often measurements will be recorded and how the data will be analyzed. It is not trivial to get meaningful results and we have to think carefully about the setup of the measurements. Energy use not only depends on the behavior of occupants but varies strongly with e.g. outside temperature and time of year (daylight; sun shine) and work pattern (meetings, coffee breaks, etc) and personal behavior /type of work. Therefore there is a lot of 'noise' in this measurement. To be able to interpret the results and draw (valid) conclusions we will start with measuring the base-line use of all the rooms involved. A method for analysis of the data will be chosen. Subsequently we will measure the energy consumption in the participating rooms. Along the way we will analyze the data, determine how long the pilot should run and whether adjustments are required. A statistical analysis will provide information on the significance of the differences found.