Uncertainty in whole house monitoring

Monitoring energy and temperatures in dwellings is becoming commonplace due to the reduction in sensing costs. Measurements can be used for informing the occupants on their energy as well as developing better inputs for building performance simulation and verifying analysis. In a home monitoring environment making sense of this data can be difficult as the number of measurements increases; one of the key challenges for the homeowner and for organisations that collect and analysis energy data is understanding what can and cannot be ‘seen’ in the data. In building simulation, there is a growing interest in applying uncertainty to generate robust model predictions, however there is also a need to understand the uncertainties in measurements used. What is often missed in these analysis is an evaluation of the uncertainties in the measurements in relation to the intended analysis. This paper presents a set of typical domestic energy monitoring measurements that have recently been collected as part of a 4 year research project in the UK. Levels of uncertainty are evaluated and the consequences for typical metrics used in energy and comfort analysis are discussed.