VOC & PM Emissionsfrom PCM Buildings
A sensing and analysis system monitoring volatile-organic-compound and particulate-matter emissions from phase-change building materials — integrating multi-sensor arrays with ML-based inference and quantification.
What was built
- 01
Developed a multi-sensor embedded system (ESP32) measuring temperature, humidity, VOC and PM across heating and cooling phases of PCM materials.
- 02
Filtered multiple sensor streams using adaptive signal processing and noise-reduction techniques tailored to each sensor type.
- 03
Developed LightGBM regression models to predict VOC and PM emission levels with low MAE and strong cross-validation performance.
- 04
Introduced a Relative Distortion Index (RDI) to quantify phase-change behaviour and detect anomalous emission events.
- 05
Designed a full evaluation framework comparing RMSE, MAE and variance metrics against physical measurements; publication pending.