Non-intrusive measurement method for the window opening behavior

Abstract The occupant windows opening behavior has a great impact on indoor air quality and building energy consumption. Therefore, measuring the window opening behavior and factors that affect it are important for the occupant behavior modeling and architectural design. In this study, we proposed non-intrusive measurement methods which can achieve large-scale sampling for the window state. An image recognition code based on MATLAB was used to conduct projective transform of the building elevation maps, identify the window positions and determine their opening proportions. The method can recognize most of window open states with the error of 8%. Based on this method, the window opening states for a hospital building from August to December in 2018 (about 6000 samples) was collected. Then the significance of the influencing factors and window opening distributions under different factors were analyzed. The results showed the outdoor temperature had the most significant effect, and the frequency of the window opening proportion in the range of 0.3–0.5 within 20–30 °C is significantly higher than other temperature ranges. This large-scale sampling method proposed in this paper provided a powerful tool for building modeling and energy analysis.