Evidence Hub

A Physics-Informed Digital Twin Framework for Health Vigilance and Indoor Air Quality Decision Support in Healthcare-Oriented Environments

Albashir A

2026

Abstract

Abstract Indoor air quality (IAQ) monitoring is important for health-vigilance-oriented indoor environments,yet conventional monitoring-only systems may not fully support proactive decision-making. This studyproposes a physics-informed IAQ digital twin framework designed for short-horizon PM2.5 forecastingand operational decision support under sensor-only constraints. The integrated methodology leveragesroutine IAQ sensor streams, temporal feature engineering, and an effective mass-balance interpretationlayer. The framework evaluates multiple forecasting models alongside baseline approaches usingwalk-forward validation, uncertainty quantification, and spike-aware reliability assessment. Aninteractive Streamlit-based dashboard translates these components into a scenario simulation workflow.Empirical evaluations selected LightGBM at a 15-minute horizon as the deployed predictive engine,achieving an RMSE of 11.20 𝜇g/m3, MAE of 1.73 𝜇g/m3, MedianAE of 0.244 𝜇g/m3, and an 𝑅2 of0.170. While the 95% prediction intervals provided useful uncertainty information, empirical coverageremained below the nominal level, with a PICP95 of 0.823. Spike-aware evaluation demonstrateduseful moderate-threshold warning behavior, though severe PM2.5 spike detection remained limited.The effective physics layer indicated source-dominated dynamics—with approximately 66.6% sourcecontribution and 33.4% removal contribution—exhibiting strong TVOC-driven source behavior. Thisframework demonstrates how…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free