A stochastic model to predict occupants' activities at home for community-/urban-scale energy demand modelling. Issue 5 (2nd November 2017)

• Record Type: Journal Article
• Authors: Yamaguchi, Yohei; Shimoda, Yoshiyuki
• Is Part Of: Journal of building performance simulation. Volume 10:Issue 5/6(2017)
• Page Start: 565
• Page End: 581
• Publication Date: 2017-11-02
• Subjects: stochastic modelling -- community-/urban-scale model -- occupant behaviour -- time use data -- residential building; 690.0113
• DOI: 10.1080/19401493.2017.1336255
• Languages: English
• ISSNs: 1940-1493

Abstract

Abstract : This paper proposes a stochastic discrete-event model to predict occupants' activities at home to be used in community-/urban-scale energy demand modelling. The model is designed to consider interactions among household members and generate consistent behaviour across simulated days and times of day with specific time-dependent characteristics. Routine behaviours undertaken routinely everyday such as working and eating meals are placed on the timeline while considering interactions among household members. Gaps between routine behaviours are then filled by non-routine behaviours. Individual specificity and consistency are enhanced by (1) providing input data representing intrapersonal variation instead of interpersonal variation, (2) using relative time to time of routine behaviours, instead of clock time and (3) adjusting the input dataset affecting the selection of non-routine behaviours and their duration. Case studies demonstrated that interactions among household members are reproduced as observed in empirical data and the individual specificity and consistency are enhanced.