Model Predictive Control Signal For Natural Ventilation Optimization Using Data-Driven Techniques
Benedetta Zuccarelli, Chenjie Xiong, Patrick Margain, Elence Xinzhu Chen
Harvard University, United States of America
Indoor environmental quality, particularly in older buildings and low-income housing, is a global issue requiring cost-effective solutions. Window-based natural ventilation strategies can improve indoor air quality and thermal comfort but depend heavily on occupant behavior. However, occupants often lack the awareness or guidance needed to optimize window operations for ideal indoor temperatures and CO2 levels. The complex nature of building dynamics further complicates prediction and control. While adaptive model predictive control (MPC) systems have shown success in optimizing ventilation using sensor data, they typically rely on costly automated hardware, limiting their accessibility.
This research proposes an accessible alternative to automated systems: a sensor-driven, occupant-guided device that delivers real-time window operation recommendations. It predicts indoor CO2 and temperature using a Multi-layer Perceptron (MLP) model with multiple time steps trained on five-minute interval sensor data, outperforming linear regression, KNN, and random forest models. The system uses a light signal to prompt user action, with a display providing optimal window opening angle and duration based on predictive and control models. A genetic algorithm determines the best window settings to maintain CO2 levels below 800 ppm while preserving thermal comfort.
The device connects to a cloud-based MPC system via microcontroller, offering responsive, data-driven ventilation guidance. Results show that the system effectively reduced indoor CO2 concentrations and sustained them within the target range. Thermal comfort was maintained, with temperatures showing only slight variation. A 20° window angle was recommended in 48% of early morning periods, balancing fresh air intake and heat retention.
By implementing an MPC system that anticipates occupant and environmental conditions and provides actionable insights, this approach empowers communities to manage indoor air quality and comfort without costly retrofits, reducing reliance on mechanical systems. Future work will broaden testing across seasons and building types, enhancing the device’s adaptability and promoting sustainable human-environment-technology interaction.
Behaviour Nudging for User Acceptance in Smart Blinds Control: A Pilot Study on Behaviour Change and Indoor Environmental Enhancement
Wen-Ting Li1, Alessandra Luna-Navarro2, Pedro de la Barra Luegmayer2, Michele W.T. Mak1, Han Fang1
1School of Civil Engineering, Faculty of Engineering and Physical Sciences, University of Leeds; 2Faculty of Architecture and the Built Environment, TU Delft, Netherlands
Automated dynamic façades offer significant potential for enhancing building energy efficiency and occupant comfort (Luna-Navarro et al., 2022). Aligning building automation systems with user preferences maximizes these benefits by preventing negative interactions between building services and occupants that might impact indoor environmental conditions or energy efficiency (H. Li et al., 2021). Engaging users in the building operation process is proposed to achieve this alignment (Morton et al., 2020). Gamification—the application of game-design elements and user experience principles in non-game contexts—has emerged as an innovative technique to engage and motivate users in sectors such as education, healthcare, and finance (W. T. Li et al., 2024). However, there is a lack of information on how to effectively integrate gamification to improve user awareness, acceptance, and engagement with automated building systems.
This pilot study experimentally investigates integrating gamification into an automated blind control system to assess its effects on task completion, user acceptance, and satisfaction. Participants interacted with the automated blind system under sub-optimal conditions—including lighting, view, privacy, and thermal comfort—while different gamified mechanisms served as behavioural nudges to reduce occupant override disruptions. Behavioural metrics—override frequency, response time, and engagement experience—were observed and analysed using a repeated measures design. A repeated measures Analysis of Variance (ANOVA) evaluated the statistical significance of differences in occupant behaviour across gamification mechanisms.
The experimental results show that occupant override behaviour differs across gamification mechanisms. Specifically, reward-based gamification, integrated with thoughtful user experience design and effective behavioural nudges, significantly enhances occupant compliance and energy-saving behaviour, outperforming feedback and goal-oriented mechanisms. This indicates that user interface design and information provision are critical for effective user engagement. These insights highlight the potential of integrating well-designed gamification and user interfaces to transform occupant behaviour, promoting acceptance of automated façades and improving energy efficiency.
Let’s Play! What Serious Gaming Reveals about Occupant Behavior and IEQ
Debby Veillette, Jean Rouleau, Louis Gosselin
Departement of Mechanical Engineering, Université Laval, Canada
Despite technological advances and the development of standards (e.g. ASHRAE 55) and certifications (e.g. WELL) that aim to improve occupants' well-being at home, comfort remains a critical challenge for many households, especially vulnerable populations (e.g. low-income, elderly, remote areas, extremes climates, etc.). Various strategies are used to understand occupants' intentions better and to study the socio-energetic aspects of buildings - the most popular being the use of questionnaires and surveys. Some examples of these evaluate the occupants' interest in energy efficiency or their preferences and adaptive strategies to improve their indoor environment. This study presents a novel approach to assessing the impact of occupant actions on their comfort through a serious game. Serious gaming is a widely used tool for education and awareness, but its operationalization in the building industry has yet to be established. Serious gaming is a promising avenue to address the lack of empirical data on occupant behavior (OB), intentions behind actions, and decision-making. Designed as a proof of concept, the game simulates various indoor environmental quality (IEQ) challenges, allowing participants to make real-time decisions to control air quality and mitigate thermal and visual discomfort. Players of the game must navigate different daily scenarios that either reflect common residential building conditions or situations of severe discomfort. Available actions for the players include the choice of the heating and cooling temperature setpoints, clothing of inhabitants, opening of windows, and control of the blinds, mechanical ventilation system, and lights. Preliminary results discussed in the paper provide valuable insights into OB and decision-making processes. This approach highlights the potential of using serious games as an effective tool for research and education in building performance, indoor environmental quality and occupant-centric design. Finally, players provide insights and appreciation into their gaming experience.
|