Anomaly Detection and Alerting system
HVAC (Heating, Ventilation and Air Conditioning) is a complex system consisting of many electronic and mechanical equipment, sensors and controllers. The main task of this system is to ensure the comfort of the people in the building by regularly adjusting and improving the air quality and indoor temperature.
In industrial areas where many people live, anomaly detection (unexpected or abnormal behavior in the data) in the HVAC system has a very important place. Thanks to the machine learning based anomaly detection algorithm developed within the scope of early warning, anomaly situations are detected on the gas concentrations in the air obtained from the surrounding sensor data, and warnings are made to the user in cases where the air quality starts to decrease. Thus, it is planned to prevent health problems that may be encountered in case of prolonged exposure to this environment.
Developed ML-based algorithm allows computing operations to be carried out on the edge device, eliminating the need for a central server. By detecting excessive toxic gas accumulation by the edge device, it gives early warnings to inform users. A sensor monitoring system has been developed in order to monitor the air quality within the scope of the application. In this system, the sensors are visualized to monitor the air quality and their instant values are visualized. When an abnormal situation is encountered, information is sent to the user via e-mail and logged in service, enabling to take early action.
