利用基於個體模型評估資訊回饋對家庭節能行為的影響 – 以住宅空調使用為例

Abstract

自上世紀以來全球工業與經濟飛速發展，造成人口激增之餘也使都市化成為不可逆的趨勢，更帶動了都市內建築物能耗與相關溫室氣體排放量的逐年增長，根據統計，建築物已然是現今能源強度最高的部門之一，約佔每年總消費的30-40%，當中又以家用住宅的能源需求為大宗。為達成社會的永續發展以及維護能源供應安全，不少國家已針對家戶層級的能源消費控管策略進行研議。許多研究也指出，居住者行為是影響建築物能源效率的關鍵因素，且相對於改善硬體設施所需之大量金錢與時間成本，優化住戶的日常消費習慣乃是更經濟又有效之作法，估計能帶來十分可觀的節能效益 (約4-30%不等)。 我國住宅能源使用有超過八成來自電力，其屬於一種無形資源，在被使用的當下難以察覺，往往導致人們過度用電卻渾然不知。為了解決此問題，專家們於是提出 "回饋" 的概念，期望透過日漸普及的資通訊技術，如常見的智慧型電表等自動化計量工具，結合家用顯示器、行動裝置、電腦等設備作為媒介，向用戶展示即時或連續性的消費資訊，提升其對自身用電情形的認知，進而改變行為從而達到節電的目標。然而，目前的研究主要聚焦在觀察小規模場域試驗 (於特定範圍內的住家中安裝回饋設備) 前後的用電量變動與針對受試者的訪談，或是以模擬方法評估相關設備於不同政策情境下的擴散及採用情形 (技術普及率)，鮮少有關於回饋如何影響消費者行為的直接描述，因此容易忽略個人意願與行為間之差距；且試驗的時長、規模、取樣方法、參與者之社經背景等干擾因素皆可能造成研究結果的差異，故至今仍缺少能夠量化回饋技術節能潛力之有效手段。 基於個體模型 (Agent-Based Model, ABM) 是一種由下而上的動態模擬方法，其核心概念在於從微觀角度觀察個體決策行為對整個大環境的影響，擅長捕捉及分析複雜系統內的角色互動情形。因此，本研究利用GAMA平台建立Agent-Based Model，針對台北市大安區住戶的室內空調使用行為進行建模。模型將環境溫度作為居住者開啟冷氣之關鍵因素，並搭配該地區的實際人口統計資料，涵蓋家庭教育程度、成員人數，年齡結構和住宅占用模式等參數，進行為期一年的動態模擬，再結合台電公布之電價資訊，計算出住宅空調用電量和相應的家庭電費支出。緊接著，模型導入了多項回饋策略，包括時間電價和社會影響力等，形成共10種不同的情境組合，旨在以個體行為受資訊回饋的干預為基礎，清晰展示出反饋內容、節能動機與用電習慣之間的相互作用，並試圖量化相關技術帶來的節電與經濟效益。研究發現，採用二段式時間電價方案下計算出的空調電費，相較於累進式電價可節省達43.1%，對住戶而言更具有吸引力；此外，與基線情境相比，回饋措施一年內最多可削減高達17.1%的總耗電量、22.5%的尖峰負載以及51.3%的住宅空調電費。這表明在財務誘因與社會壓力的驅使下，若用戶願意調整消費習慣，不僅能有效節約電力浪費，亦可顯著減輕尖峰時段內的用電壓力，同時避免家庭負擔高昂金錢損失，最終達到能源供需兩端雙贏之局面。 本研究利用模擬方法，驗證了資訊回饋在推動住宅節能方面有效性，並深入分析了各種策略組合下的節能成果。而有鑑於到智慧型電表等先進計量基礎設施 (Advanced Metering Infrastructure, AMI) 在我國仍處於初期發展階段，目前僅有零星住家已完成安裝，或是於特定地區設立智慧電網技術示範場域，因此本研究的成果有望成為將來大規模推廣此類回饋裝置時的重要參考，除了能協助評估其在節能和經濟效益方面之潛力，也為日後的實施策略提供了引導方針。

Since the last century, the rapid growth of global industry and economy has caused a surge in population and made urbanization an irreversible trend, as well as the steady increase in building energy usage and building-related greenhouse gas emissions. According to statistics, buildings have been one of the sectors with the highest energy intensity nowadays, contributing about 30-40% of total world energy consumption annually, among which the demand for residential buildings accounts for the most significant part. As stated above, many countries have been working on strategies for controlling household energy consumption to achieve sustainable development and ensure energy security. Numerous studies have indicated that occupant behavior is the critical factor affecting the energy performance of buildings. Furthermore, compared with the substantial capital and time cost required for renovating physical facilities, improving the daily consumption habits of occupants may be a more economical and effective method estimated to bring considerable energy-saving benefits (about 4-30%). In Taiwan, more than 80% of residential energy use is in the form of electricity, a kind of invisible resource difficult for people to perceive when using it and hence often lead to overconsumption. To address this problem, experts put forward the concept of "feedback", expecting to utilize increasingly prevalent information and communication technologies (ICTs), such as commonly seen automated measuring tools like smart meters, in combination with in-home displays (IHD), mobile devices, and computers as media to provide users with real-time or continuous consumption information. This approach is considered able to enhance users’ awareness of their own electricity usage and thereby change behavior routines for the purpose of conserving energy. However, the existing literature primarily focuses on observing the discrepancies in electricity consumption before and after conducting small-scale field experiments, such as the installation of feedback devices in dwellings within a specific range. Research in this area also includes carrying out qualitative interviews with the participants, as well as evaluating the diffusion and adoption levels of such equipment, for example, the technological penetration of smart metering, under various policy scenarios using simulation approaches. In contrast, there are still few direct descriptions of how feedback affects consumer behaviors, and the intention-behavior gap of individuals is often ignored in relevant studies. Additionally, a variety of confounding factors can also cause variation in results, including the duration and scale of trials, sampling methodologies, socio-demographic characteristics of target groups, etc. Therefore, still no effective means that have been proposed to quantify the energy-saving potential of feedback technologies. Agent-Based Model (ABM) represents a dynamic bottom-up simulation approach, where the core concept involves inspecting the impact of individual decision-making behaviors on the whole environment from a micro perspective. This method excels in capturing and examining the interactions between roles within complex systems. Consequently, this study employed the GAMA platform to develop an Agent-Based Model concentrating on occupant usage patterns of residential air conditioners in Da''an District, Taipei City. Ambient temperature was applied as the trigger for residents to operate the appliances, incorporating actual demographic data of this region, including household educational level, family size, age distribution, and occupancy patterns of the dwelling. Through one-year dynamic simulation coupled with electricity pricing information published by Taiwan Power Company, we obtained the residential air conditioning load profile as well as the corresponding electricity costs. Then, several feedback strategies, such as time-of-use rates (TOU) and social influence (normative comparison、peer comparison) were introduced, creating ten distinct scenario combinations. These were designed to base individual behavior changes on information feedback interventions, clearly illustrating the interplay between feedback content, energy-saving motives and consumption habits, attempting to qualify the potential energy efficiency benefits of relevant technologies. The research found that adopting 2-tiered TOU pricing scheme for air-conditioning costs could result in saving up to 43.1% compared to progressive pricing scheme, making it more attractive to households. Moreover, in comparison with the baseline scenario, feedback measures were able to reduce total electricity consumption by up to 17.1%, peak load by 22.5%, and energy costs for residential AC by 51.3% within a year. This indicates that under the influence of financial incentive and social pressure, if users are willing to adjust their consumption patterns, not only can they effectively conserve energy, but also significantly alleviate the demand during peak hours and generate substantial cost savings for households, ultimately achieving a win-win situation for both energy supply and demand sides. This study utilized simulation methods to validate the effectiveness of information feedback in promoting residential energy conservation and conducted an in-depth analysis of energy-saving outcomes under multiple strategy combination. Given that Advanced Metering Infrastructure (AMI), including smart meters, is still in the early stages of development in our country, with only a few households currently equipped and specific areas designated for smart grid technology demonstration, the results of this study are therefore anticipated to serve as a crucial reference for future large-scale deployment of such feedback devices. Besides aiding in assessing their potential for energy-efficient and economic benefits, the findings also provide guidance for planning the implementation strategies.