Adding sub-hourly occupancy prediction, occupancy-sensing control and manual environmental controlto ESP-r

A wide range of events in buildings occur at sub-hourly frequencies. Notable examples include daylight-sensing control and manual adjustment of blinds and lights in response to sub-hourly illuminance variations. Such short-term changes can produce notable shifts in instantaneous solar and equipment loads, in turn affecting electrical energy demand. Although sub-hourly time discretization is currently available in several whole building energy simulation programs, it remains challenging to model this level of complexity as traditional hourly utilization profiles describing occupancy, lighting and equipment loads remain the basic input data model. This paper provides a background review of building energy simulation models of stochastic occupancy prediction, occupancy-sensing control and stochastic manual control, and outlines the current addition of a few investigated models to ESP-r. INTRODUCTION Sub-hourly time discretization presents an opportunity to reconsider several existing assumptions in building energy simulation. The use of hourly meteorological input data is one example. Walkenhorst et al. (2002) demonstrate that the predicted annual artificial lighting demand can be underestimated by up to 27% if daylighting simulations are based on 1-hour means instead of 1-min means of measured beam and diffuse irradiances. To this end, an adapted Skartveit and Olseth (1992) stochastic model, deriving short-time fluctuations from hourly time series irradiance data, is found in DAYSIM (Reinhart 2001), a RADIANCE-based (Ward 1994; Ward Larson and Shakespeare 1998) dynamic daylight simulation method. Similar work is described in Janak and Macdonald (1999). Analogous work on stochastic modelling of short time-step wind velocity fluctuations is presented by Marques da Silva and Saraiva (2002). One should expect to witness the predestined integration of a number of these models within larger whole building energy simulation programs in the near future. Sub-hourly time discretization may equally justify a rethinking of occupancy-related input data models. Although whole building energy simulation programs such as ESP-r (ESRU 1999; Clarke 2001) offer sub-hourly simulation time-steps, hourly utilization profiles of occupancy and related internal gains, such as lighting and equipment, remain the sole input data model; a solution passed down from the previous generation of hourly simulation programs. The purpose of this paper is to review building energy simulation modelling of stochastic occupancy prediction, occupancy-sensing control and stochastic manual control, and briefly describe the current addition of several models to ESP-r. BACKGROUND Commonly expressed as ratios of some maximum user-defined value, hourly utilization profiles are often adequate as average input data models for large thermal zones containing multiple spaces. Occupancy ratios of thermal zones describing offices environments may typically range from 30% to 90% of specified maximum loads when occupied, reflecting variable levels of absenteeism, while office equipment and lighting ratios usually reach a maximum of 90% of specified maximum loads; an indication of diversity in use. Using sub-hourly occupancy input loads (i.e. metabolic heat discharged by occupants), rather than hourly loads, would typically have little outcome on annual energy simulation results, given the thermal lag of building mass and mechanical systems. If daily patterns of light and office equipment use remain close to predetermined schedules (e.g. if lights are kept on during normal occupancy hours, regardless of occupancy patterns), then the impact of using sub-hourly input data would remain equally trivial on annual energy requirements. The use of hourly utilization profiles instead becomes controversial when considering certain control strategies, as they are quite sensible to short-term variations in occupancy. Denis Bourgeois is a PhD student, École d'architecture, Université Laval, Québec, Canada ([email protected]); Dr. Jon Hand ([email protected]) and Dr. Iain Macdonald ([email protected]) are research fellows at the Energy Systems Research Unit (ESRU), University of Strathclyde, Glasgow, Scotland. Dr. Christoph Reinhart is a research officer, Institute for Research in Construction, National Research Council, Ottawa, Canada ([email protected]).