Light switch behaviour: occupant behaviour stochastic models in office buildings

It is common knowledge that occupants’ behaviour on building control systems plays a significant role to achieve comfortable indoor environmental conditions. Moreover, different research studies have shown how occupants’ behaviour also has a huge influence on energy consumption. Consequently, since the building sector still consumes nearly half of the total amount of energy used in Europe and because occupants’ comfort should be one of the major aim of a building construction, this influential factor should be further investigated. Reliable information concerning occupants’ behaviours in a building could help to better evaluate building energy performances and design robustness, as well as, it could help supporting the development of occupants’ education to energy awareness. Concerning occupant behaviour related to indoor lighting systems, many studies have been made regarding occupants’ feelings and performances to certain visual stimuli due to different light systems. Nevertheless, occupants’ interactions with lighting control systems needs further investigation also because few models to predict switching operations have been implemented in energy simulation programmes. This study proposes probabilistic models to describe occupants’ switching on-off control over lighting. They have been developed using a multivariate logistic regression based on measurements of indoor climate parameters, outdoor environmental conditions and artificial lights “switch on/off” actions. Measurements were made over eleven months for three different office rooms. Two predictive light-switch behaviour models were inferred in relation to the number of actions carried out by the users (active or passive). The models are presented and critically discussed in this paper. The study extends the information on environmental parameters influencing occupants’ manual control of the lighting system in offices and energy consumption.