FCC S.A. develops a virtual sensor model for predicting ammoniacal effluents treatment performance

As a main ally of the Digital Transformation movement, Fábrica Carioca de Catalisadores believes that digital technologies are essential support tools that add a unique value and thus it travels its journey to turn data into ever more intuitive solutions. 

One such example is the application of Machine Learning solutions to predict future events. In this context, FCC S.A. developed a virtual sensor model for predicting the company’s ammoniacal effluents treatment performance. The model is based on a neural network training with the history of the process data. The sensor results are calculated and made available in real time to the process follow-up software programs. 

This application reduces the waiting time for new analysis results, imparting higher accuracy to the process tracking and intelligence for the environmental impacts studies. It can also be a great tool for instrument failure events. Besides, the Project scalability is also included in the methodology to enable the utilization of the same modelling for other plant virtual sensors.