Thermal Grace
AI & Data
Semester programme:Open Learning/Innovation
Client company:Workplace Vitality HUB (WPVH)
Project group members:Artur Kraskov
Bartosz Kaszuba
Project description
The core design challenge was to develop an integrated, modular pipeline capable of real-time thermal comfort analysis. This involved answering the research question: How can diverse data sources—ranging from indoor occupant heat maps to external weather patterns—be synthesized using AI to provide actionable comfort advice? Following a literature study, the project utilized rapid prototyping to bridge the gap between theoretical thermal indices and cloud-based Large Language Model (LLM) analysis.
Context
This project operates within the domain of Smart Building Management and Human-Centric Design, specifically focusing on occupant well-being. The context is the critical need for maintaining perceived thermal comfort in office environments to ensure productivity and energy efficiency. By moving beyond simple temperature monitoring, the project addresses the complex relationship between environmental variables and human thermal sensation in modern workplaces.
Results
The primary outcome is a modular software and hardware system featuring remote wireless sensor nodes and a multi-modal data pipeline. Key technical achievements and insights include:
- Integrated Sensing: Successful data collection from environmental sensors, mmWave radar for occupancy, Grid-EYE heat maps, and the Buienradar API for outdoor weather.
- Validated Metrics: Implementation of the PyThermalComfort library to calculate standard PMV (Predicted Mean Vote) and PPD (Predicted Percentage of Dissatisfied) values.
- AI-Driven Advice: Validation of a cloud-based workflow using GitHub Models to process data for real-time HVAC control recommendations and suitable zone identification.
- TRL Positioning: The project currently sits at TRL 4 (Technology Validation in Laboratory Environment), having demonstrated a functional prototype that integrates basic components for joint operation.
About the project group
The project started after the end of GLOW in November. 2 Delta students cooperated. Semester 7 and semester 2. The stakeholder required a smart system to sense and predict perceived thermal comfort in offices. One student worked full-time and another 2 days a week.