Sprecher Automation cutting pattern optimization
For Sprecher Automation GmbH, an automatic side goods optimization was developed for the ideal cutting pattern yield for an automatic saw line.
Sawmills also face the challenge of making ideal use of all their resources. In order to achieve maximum added value, existing roundwood must be processed as quickly as possible, with the lowest possible waste and into the most profitable products. In addition to production adapted to the market, local dynamic changes such as stock levels and machine conditions must also be taken into account at short notice.
The processing sequence along a chipping line can be summarized into the following exemplary steps: Log infeed, debarking, turning, chip removal, side cut, main cut, and removal by conveyor to the lumber sorting line. The valuable main product is obtained from the inside of the log. Lateral goods are sawn timber at the edge of the roundwood with a defined forest edge portion:
Where the forest edge is considered the part of the sawn timber with the original (and debarked) surface of the log. The main product is usually cut sharp-edged (without forest edge), because it is fitted into the interior of the diameter-sorted log. However, since each log is grown differently and has its own curvature, the ideal position and size of the desired sideboard per cantling varies, especially when mostly price-optimal sharp-edged sideboard is the goal.
In the cycle time of conveying on the chipper line, which is often operated at a feed rate of 2.5 meters per second (and beyond), not only the most profitable sideboard for the current log must be found, but also their ideal positions so that the cutters can be suitably positioned.
Innovative collaboration for a high-performance optimization solution
In cooperation with Sprecher Automation, RISC Software GmbH has created a high-performance optimization solution for the price-optimal cutting of side goods. In addition to the consideration of large product variety of possible side goods boards, given by their dimensions in length and width, their price valences and their permitted dimensions of the forest edge characteristics, mainly shortest computing time of clearly below one second per log was the ambitious goal of the development.
Several well coordinated and interlocked software components harmonize ideally and ensure the high performance of the optimization solution. A solid multi-core architecture coupled with efficient memory management provide the basis for the fastest calculations. The intelligent search algorithm based on this not only reliably delivers optimal solutions in terms of product selection and cutter positions, but also comes up with the information as to whether a global optimum has been found, which usually occurs in just a few milliseconds.
Since spring 2019 and 2021, the optimization software has been in use at two plants, where it supports optimal log yield. Further deployments are planned for the future.
Dr. Roman Stainko
Mathematical Optimization Specialist