Vire AI Scanner - case ANDRITZ
The first autonomous woodyard cranes in the world are used at the new Kemi bioproduct mill in Finland. An advanced AI scanner system helps the cranes to locate and lift the logs from the incoming trains.
Metsä Group’s new state-of-the-art bioproduct plant in Kemi will be first factory in the world to use autonomous logyard cranes. The intelligent Vire AI Scanner system provides the data that the crane system needs to locate and lift the logs from the trains securely.
Metsä Fibre’s Kemi bioproduct mill is the largest investment in the history of the Finnish forest industry. The new mill will use the latest technology, and a good example of this is the electrically powered and autonomous crane system used for handling timber at the mill. The cranes will be supplied by ANDRITZ, one of the world’s leading suppliers of machinery for the forest and metal industry.
Manual logyard cranes, controlled by an operator in the crane cab, are standard technology for wood handling in North America, for example. However, a completely new and highly advanced wood handling solution has been built by ANDRITZ and it is used for the first time at the Kemi bioproducts plant. The system consists of two electrically powered linear cranes that are autonomously operated.
– It is the first crane system in the world that can operate autonomously, i.e. without an operator. An operator is only needed to monitor the operation, says Ilkka Manninen, who leads the crane business outside North America and the autonomous crane team at ANDRITZ
"This is the first woodyard crane system in the world that can operate autonomously, without an operator."
Ilkka Manninen, ANDRITZ
Vire Labs converts the train loads into data
Each freight train arriving at the Kemi timber yard passes the Vire AI Scanner poles, allowing the AI scanner to generate a three-dimensional point cloud of the train’s cargo. The AI-based scanner solution used to create the point cloud was supplied to ANDRITZ by Vire Labs, a Finnish company specialising in AI-based data collection and detection solutions.
The different semantic parts of the point cloud are identified by a neural network as edge computation, and the data is processed into values used to control autonomous cranes, which are immediately available to the crane system, and the waiting time for the trains is minimised.
– The data delivered by the scanner system tells the crane system what is timber, what is bank and what is the train car body. This way it immediately knows where to lift the load,” explains Ilkka Manninen.
The autonomous crane system needs highly accurate data
The autonomous cranes require a high degree of precision in their control. The boom comes from a height, but its grab has to hit exactly the right spot. The distance between the body of the trolley and the wooden stack is 25 centimetres, and the grab must fit between the two when lifting the stack. The point cloud produced by the Vire AI Scanner is accurate to within one centimetre. ire AI Scanner can use cameras and a wide range of other devices to collect data. In Kemi, the detection is based on data collected through LiDARs, or laser scanners.
– To create a point cloud, trains and cargo are scanned from both sides. The identification system supplied to ANDRITZ consists of two Vire AI Scanner columns, each with two LiDARs and a Virebox.AI terminal, says Ossi Porri, Business Development Director at Vire Labs.
– The autonomous cranes require point cloud data to be always and instantly available to them. The focus is on reliability and dependability, and reliability is guaranteed even in winter, Porri adds.
The operation of autonomous cranes requires that the point cloud data generated by the scanners is always and immediately available to them. That is why we have invested in the reliability and dependability of the Vire AI Scanner system."
Ossi Porri, Vire Labs Oy
Data collection enables further development
ANDRITZ has been very pleased with the cooperation with Vire Labs.
– Cooperation has been effortless. For example, it has been easy to discuss the challenges and opportunities that arise, which is important when we are building a completely new solution, says Ilkka Manninen.
Manninen and Porri emphasise that the most important thing about the autonomous system and the use of the AI scanner is not the replacement of human work, but the collection of data that enables further development.
– The number of roundwood plants using smart solutions will continue to grow at an accelerating pace around the world. In practice, we will need such a scanning system wherever autonomous cranes will be used, says Manninen.
– We also want to find out what other kinds of data Vire Labs’ systems could produce for us in the future and what we can do with the data we collect, concludes Ilkka Manninen.