AFRY Kappa Prediction Tool

AFRY’s Smart Kappa Analyser is quick and easy to implement and provides vital help for the operators

System operations are usually complex and require adjustments for multiple variables simultaneously. Explainable machine learning together with multi-variable optimization enables to creation of autonomous automation systems.

Once the analyser has learned from the data, it optimises an estimate for variables affecting to kappa and gives a ”recipe” about how much other variables need to be adjusted to reach the desired state of the system. It helps to keep the process stable, so that high quality of the production can be maintained.

Kappa Prediction Tool Optimises Pulp and Paper Industry Efficiency

• Quality is a high-importance parameter in the pulp and paper industry. The kappa value is one key parameter affecting the product quality
• AFRY’s Kappa prediction tool enables improvements for controlling the Kappa value at different parts of the project
• Improved Kappa control enables for example:
  • Maximised material efficiency
  • Minimised chemical consumption
  • Improved product quality and cost-effectiveness
  • Better quality control over the process in general

Kappa Prediction Tool functionalities:

• Collects process data as a time series
• Creates several mathematical models from the process (linear regression, neural networks, decision trees) and choose the best fitting model
• Creates an interface that enables giving setpoints to chosen process parameters (such as Kappa) and calculates the required values for the other process parameters affecting the chosen process parameters
• Can be used to analyse other process parameters as well, if needed

How to implement the tool?

• The user interface is tailored based on the customer needs and available data
• The tool can be built into any analytics platform (PowerBI, Exel, SAS Analytics, browser-based solutions, etc.)
• Implementing the tool requires:
  • Some analytics platform (clients own or delivered by AFRY)
  • access to the online process data (creating new or using existing data interfaces), the first version can be done with history data only
  • Studying the customer process and available data
  • Programming the tool on the chosen analytics platform (2-4 weeks + fine tuning)

Step-by-step Implementation:

• Checking automation infrastructure: 1 visit to the site

If the client has all variables inside PLC/DCS exposed by OPC-UA:

• Install the platform and connection to the PLC/DCS: 3 days
• Connecting all variables with the Kappa Analyzer: 2 days
• Operation and configuration training: E-learning
• Assisted operation: Up to 20 hours in the first month
• Tech support: Up to 20 hours per month after assisted operation

* if the client doesn't have the minimum automation infrastructure, AFRY will provide a proposal to build it

Other benefits:

• Continuous improvement of the product and the Platform, including updates, fixes, performance upgrades, etc.)
• Freedom to customize the solution and also implement other solutions or algorithms inside the opensource AFRY Smart Platform