A Thermal Comfort Index for Healthy Indoor Environments: An Interpretable, Simulation‐Based Mixture‐of‐Experts Model

Many buildings comply with thermal comfort standards yet still feel uncomfortable to occupants, leading to complaints and suboptimal indoor environmental control. This gap arises because common comfort representations do not capture how humidity, air movement, local nonuniformities, and short‐term dynamics jointly influence perceived comfort under realistic indoor conditions. This study proposes a thermal comfort index (TCI), an interpretable, physics‐informed mixture‐of‐experts model for evaluating indoor thermal environments. The model integrates whole‐body drivers (operative temperature, humidity, air speed, clothing, and metabolism), local effects (radiant asymmetry, vertical temperature gradient, and floor temperature), and transient exposure into a bounded comfort score with three actionable classes. Training and calibration are based on standard‐relevant synthetic scenes anchored to ASHRAE and ISO criteria. Scenario‐based validation, sensitivity analysis, counterfactual audits, and psychrometric comparisons show that TCI captures regime‐dependent air‐movement effects and cumulative discomfort mechanisms that are often overlooked by conventional approaches. A lightweight MATLAB application supports practical assessment and decision‐making. The framework provides a transparent tool for managing thermal conditions as part of a healthy indoor environment.