Machine Learning Powered Soft Sensors Reduce Toxic Chemical Use in Paper Manufacturing

Implement machine learning (ML) soft sensors at multiple pulp and paper manufacturing facilities to demonstrate scalability within a company and across the industry.

Data-driven soft sensors are predictive models that use machine learning and historical process data to estimate process variables. Soft sensor technology can be used to control manufacturing processes that are difficult to monitor with conventional sensors. By leveraging predictive analytics, machine learning-based controls and soft sensors enable smarter decision-making and reduce energy consumption. Implementing these tools strengthens U.S. manufacturing competitiveness and contributes to a more resilient supply chain.

Soft sensors - which use historical data to estimate hard-to-measure variables - eliminate the need for expensive physical sensors. By leveraging vast amounts of data, machine learning-based controls can identify patterns and anomalies, enabling faster decision-making and process adjustments. This leads to increased productivity, reduced waste, and improved sustainability, enabling American industry to adapt to rapidly changing market demands.