Analisis Presisi Klasifikasi Tahap Pengeringan Kayu pada Sistem Pemantauan Suhu Berbasis IoT

Abstract

Good wood drying depends on keeping the oven temperature within a predetermined range. Manual temperature monitoring makes process quality highly dependent on operator presence and accuracy. This study develops a web-based wood oven temperature monitoring system using NodeMCU ESP8266, DHT11 sensor, LDR light sensor, and OLED display. Sensor data are sent to the server in JSON format through HTTP POST every two seconds and displayed on a real-time dashboard. The system was developed using the Rapid Application Development method. Evaluation was carried out through automated black box testing, firmware compilation verification, data transmission testing, and Monte Carlo precision simulation with 10,000 samples per stage. All functional scenarios passed and the firmware compiled successfully. The precision simulation shows that the DHT11 tolerance of ±2 °C reduces classification accuracy for drying stages with a narrow range. This finding confirms that classification accuracy is determined not only by program logic but also by the relationship between sensor tolerance and decision class width.