Towards Broad Implementations of Date-Driven Soft Sensors in Industrial Manufacturing

Project Lead: Auburn University
Partners: Rayonier Advanced Materials & Georgia Pacific

Member % Cost Share: 50.1%
CESMII % Cost Share: 49.9%
Duration: 6 Months

Problem Statement

Data-driven soft sensors are very attractive modelling approaches for energy efficient operations and reducing maintenance costs and downtime in different industries. However, commercial implementations of the sensors are still very limited in some industrial manufacturing processes such as  the pulp and paper industry and textile industry.

Project Goal

T specific objectives of the proposed project are to implement the soft sensor technology: 1) at other stages and plants/mills at Rayonier Advanced Materials; 2) at two more mills of other leading U.S. pulp and paper companies, Georgia-Pacific, or International Paper or Packaging Corporation of America; 3) we will assist CESMII to develop SM profiles for brown-stock washing to enable/facilitate additional implementations.

Technical Approach

We will leverage the unique experience we have learned and follow the proven successful approach: 1) build excellent teamwork with domain experts in different mills/plants to develop the soft sensor by analyzing historical operational data; 2) verify and optimize the soft sensor, first in passive mode, then in active mode; 3) incorporate the soft senor for enabling multi-objective process controls; 4) trial implementations of this technology in selected short operation periods; 5) full implementations; and 6) carry out the techno-economic analysis.

Deliverables/Outcomes/SM Marketplace

  • To be listed at project completion.

Potential Impact

  • Improve washing efficiency by 15% or reduce energy consumption in  recovery evaporation in U.S. kraft mills by around 24.6 trillion Btu while reducing defoamer  usage by 20%, providing the industry with a unique opportunity to both reduce cost and  improve environmental sustainability


  • Develop and deploy SM solutions to ALL major processes in pulp and paper manufacturing that are reusable for process modeling and control across a wide range of high-energy consumption manufacturing processes, building on the SM Platform core technologies.
  • Train two Ph.D. students on SM in the areas of advanced ML sensing, data analytics, and model-based control and optimization of pulp and paper processes in close collaboration with our industry partners throughout this project

Project Selection & Announcements


Leverage project outcomes for your manufacturing operations

Learn More