The project ran between September 2018 and March 2021.
The project aims to improve raw material management for customer-oriented products. This is done by using the latest technology for high-resolution information about the log and data of its sawn products in the sawmill process. The concept is based on multivariate prediction models that has been trained for industrial customers, and where the wood material descriptive parameters from x-rayed logs (CT) and planks and boards (vision system) are used.
The goal has been to demonstrate the concept’s potential compared to today’s rule-based method for customer management, to describe robustness and sensitivity to unwanted variation in input raw material and show how the quality dialogue between the customer and the sawmill in a B2B relationship, can be changed within the framework of the concept.
Instead of having a customer formulating and decide limits for all different types of wood properties that they wish to buy, the customer simply indicate which boards they wish to buy. This customer response is used as input for training prediction models, where all data (each log/board) measured in the sawmill process is used as descriptive parameters. The models can be used online for control of the sawing process.
Luleå University of Technology, Norra Skog, FinScan and Microtec.
The project’s budget was 12 MSEK.