藉由人與建築的互動發展個人化熱舒適模型與空調控制策略

Abstract

現代人的生活有90%以上的時間與建築物息息相關，也因此我們愈來愈依賴空調系統來協助維持建築室內環境的舒適度。為了維持或提升室內環境使用者的舒適度，往往需耗費大量的能源來達成。然而，我們對於建築物營運維護過程中最主要服務對象，也就是每一位建築環境使用者，所知甚少。 熱舒適是表達對於熱環境滿意的心理狀態，至今研究人員尋求各種方法來評估使用者舒適度的量化模型，其中發展較久且廣為人知的就是預測平均投票指標，也就是PMV模型。由於每位使用者對於環境舒適度接受範圍不同；進一步來說，當使用者無法忍受熱舒適度時所造成的反應稱作使用者行為，這其中已包含各人容忍度的差異。因此我們藉由使用者的行為來反應其對於環境熱舒適的喜好。過去的研究亦指出環境使用者的行為對於能源使用效率有直接關係，因此如何用一套好的方式來分析環境使用者行為是很重要的課題。 本研究中，我們建立PMV指標與使用者行為兩者之間關係，並引入貝氏方法分析，希望藉此得到個人使用者熱舒適度檔案(模型)，再將成果用來擬訂讓使用者與建築構件互動次數最小化之單人環境空調控制策略，來達到節能的第一步。本研究使用先前自行研發建置的雲端管理平台進行實驗，該平台已經具備環境監測與控制建築構件(空調)的能力。藉由實驗來驗證個人化熱舒適檔案的建立與空調控制策略的成效。 我們從實驗的成果證實，透過使用者與建築構件互動可以用來建立個人化熱舒適檔案，形成一套在線學習系統並且依據熱舒適檔案對室內環境進行自動控制。

Nowadays, we spend more than 90% of our time inside of buildings. People rely on more heating, ventilation, and air conditioning (HVAC) systems to maintaining indoor environmental comfort. In order to achieve comfort in the indoor environment, it also costs a huge amount of energy. However, when we maintain the daily operation of the building, we are still don't know how to fit the thermal comfort needs of the main body of the service: the occupants. Thermal comfort is the condition of mind that expresses satisfaction with the thermal environment. Researchers use many methods want to find indicators that can effectively quantify occupant behavior. One of the indexes have long-term development and also widely known is the predicted mean vote (PMV) index. Each occupant accepts the environmental comfort range is different. Further, when the occupant cannot tolerate thermal comfort and trigger the activity, which is called occupant behavior; and it includes the difference tolerance of each person. Past research has pointed out that occupants directly affect the efficiency of energy use, and therefore develop a method to capture, simulate or establish occupant behavior model is an important topic. In this study, we establish a relationship between PMV index and occupant behavior, and then import the Bayesian approach analysis, to get a personalized occupant thermal comfort profile (model). Then we according to the thermal comfort profile to develop a strategy to control the environment to reduce the interaction between the occupant and the building components (like air conditioner), as the first step of energy saving. The study was conducted experiments through our own cloud management platform, which has environmental data monitoring and air conditioner (AC) control capabilities. The experiments were conducted to verify the effectiveness of the establishment of personalized thermal comfort profiles and AC control strategies. From the experimental results, we can through the occupant interactions with the building components to create a personalized thermal comfort profile, by the online learning system and apply automated control to the environment according to the thermal comfort profile.