Banca de QUALIFICAÇÃO: EDUARDO GARCIA ZACCHARIAS

 This study aims to experimentally evaluate the metrological performance of the DHT22 and DS18B20 sensors, integrated into an ESP32-based acquisition architecture and the ScadaBR open-source supervisory system, by comparing their readings against calibrated reference instruments under representative laboratory operating conditions. The experimental infrastructure—comprising the ESP32, sensors, Modbus TCP/IP and MQTT communication protocols, ScadaBR, and Node-RED—forms a test bench that enables systematic data acquisition for the proposed assessment.

 The trials will be conducted across two distinct environments within the Microbiological Testing Laboratory at the Center for Food and Drug Research (NUPLAM/UFRN): an incubation room (temperature at 34 ± 1.0°C, humidity ≤ 65%) and an incubation chamber (43 ± 1.0°C). In the incubation room, two DHT22 sensors will be positioned at different locations on the shelves used for storing Petri dishes under analysis, with each DHT22 sensor paired with a reference sensor for direct comparison. Similarly, in the incubation chamber, two DS18B20 sensors will be deployed alongside the equipment’s factory-installed sensor and a calibrated reference sensor.

 A LogBox-RHT-LCD electronic temperature and humidity data logger will serve as the reference instrument, enabling the evaluation of accuracy (mean error, MAE, RMSE), repeatability (deviation between sensors of the same model), temporal stability (drift analysis via linear regression), and metrological capability (ratios $Q_{MS} = 2U_{MS}/T$ and $Q_{MP} = 2U_{MP}/T$ between process tolerance and the sensor’s expanded uncertainty). Furthermore, transient tests—simulating door openings and power outages—will assess sensor responsiveness and alarm system efficacy.

 The research hypotheses define quantitative acceptance criteria derived from manufacturer specifications and regulatory requirements. The findings are expected to delineate the operational conditions under which these low-cost sensors can be reliably deployed for continuous monitoring, identifying both their capabilities and constraints, and ultimately contributing practical guidelines for their implementation in regulated laboratory settings.