Soft Sensor for Exhaust Gas Emissions Monitoring: Application in Industrial SCADA

This paper presents the online implementation of an exhaust gas concentration soft sensor in an industrial SCADA platform. The soft sensor is a data-driven model based on machine learning techniques. It was tested and validated in a pilot plant of hybrid automation technologies and networks that include analog, discrete and Fieldbus protocol-based instruments. It may be used to monitor CO2 emission levels in order to achieve reliable and consistent operation regarding production efficiency and environmental safety. The automation solution is a friendly and generic framework wherein plant personnel can create, update and use local MLP ANN soft sensors. Even though the framework was implemented in a combustion process it can be generalized for different soft sensors and distinct processes.