MTS AI Helps Segezha Pulp and Paper Mill Trim Production Flaws

MTS AI product team suggests incorporating a smart video surveillance system in the manufacturing cycle and trains a neural network to detect flaws in produced paper.

Paper web is produced by repeated rolling and drying of pulp stock in paper and cardboard machines. This process yields a roll of finished product whose dimensions depend on the unique technological properties of each machine.

At this stage of the cycle, disregard of machine maintenance routine and operator-contributed process violations result in product flaws, such as paper breaks, tail cutter clogging, as well as folds, creases and holes through the paper web. Every production flaw has to be detected at this stage and reported to operators to make sure that substandard paper web is not cut into rolls and sent to storage.

The pulp and paper mill called on MTS AI to help them automate the flaw detection process and improve the mill’s overall production performance.

Before presenting a solution, MTS AI experts carry out a business process function test and a manufacturing survey to find any bottlenecks that may impact the project outcome. Next, they engineer algorithms and design the business logic and interfaces, develop a solution, and finally integrate it into the customer’s existing systems.

After examining Segezha PPM’s manufacturing process, the MTS AI team suggested deploying a video surveillance system with industrial high-speed cameras and lights, and engineered custom software for business process automation routine. The developers also trained a neural network to detect flaws in the paper web.

The system installed at the enterprise is currently running in test mode and helps detect over 90% of paper web flaws. Having fewer flaws and being able to forecast paper web ruptures, the company will streamline its manufacturing processes, minimize equipment wear, and increase operating efficiency and production profitability.