Design and Development of a Real-Time IoT-Based System for Monitoring Essential Indoor Air Quality

Abstract

Controlled industrial environments require strict control of air quality to prevent contamination and to ensure the reliability of sensitive manufacturing processes, particularly in the pharmaceutical, electronics and semiconductor industries. Verification in such facilities has traditionally relied on scheduled manual particle counts or on costly proprietary facility-monitoring instruments, which leave temporal gaps in coverage and offer little spatial resolution. Each edge node integrates a Sensirion SEN55 multi-parameter sensor — reporting four particulate mass-concentration fractions (PM1.0, PM2.5, PM4 and PM10), a volatile organic compound (VOC) index, a nitrogen oxide (NOx) index, temperature and relative humidity — with an ESP32-S3 microcontroller. The microcontroller samples the sensor over I²C every five seconds, renders the readings on an integrated 7-inch LVGL touchscreen human-machine interface at the point of use, and publishes them over Wi-Fi to an EMQX message broker under a per-device MQTT topic. A Node.js server aggregates all nodes through a single wildcard subscription, evaluates each reading against configurable control and specification limits, persists an hourly snapshot to a TimescaleDB time-series database, and pushes live updates to a responsive browser dashboard over Socket.IO. The resulting four-layer architecture separates sensing, edge processing, communication and visualization, allowing additional nodes to be introduced without reconfiguring the back end. Four nodes were deployed concurrently across a single building, with independent per-device limit evaluation, status reporting, offline detection and spatial presentation on a floor plan. A continuous seventy-two-hour record demonstrated stable operation of the complete chain over three consecutive diurnal cycles, with particulate concentrations ranging from a baseline near 11 µg/m³ to a peak of 82 µg/m³ (PM10) and the four particulate fractions preserving their expected ordering throughout. At an indicative cost of approximately Rp2,460,000 per node, the work demonstrates that commodity sensing and edge hardware can deliver responsive, spatially distributed air-quality monitoring at a small fraction of the cost of traditional facility-monitoring instrumentation. The deployment reported here was conducted in an ordinary indoor environment rather than a classified cleanroom, and no comparison against a calibrated reference instrument was performed, so the reported concentrations are indicative rather than metrologically traceable; calibration against reference instrumentation and formal validation aligned with ISO 14644 monitoring procedures are identified as the necessary next steps.