Digitise casting processes, avoid rejects

"Among other things, our project partner manufactures plain bearings for machines in the mining industry. The components, which weigh up to 1.5 tonnes, are produced in very small batches. Accordingly, the machine parameters have to be set anew for almost every part," explains Danka Katrakova-Krüger, Professor at the Institute for General Mechanical Engineering at the TH Köln. Until now, both the settings and the documentation were done manually.

According to the project partners, centrifugal casting could benefit greatly from digitalised and (partially) automated production. In the manufacturing process for rings, discs and pipes, liquid metal - in this specific case an alloy of copper, tin and lead - is filled into a mould rotating around the central axis, the so-called ingot mould. The molten metal is pressed against the mould wall by rotation and hardens there. 

Melt, press - and melt again?

"Sometimes it takes three attempts until a product is perfect - the failed attempts have to be melted down again. The associated expenditure of energy and resources, the tying up of capacities and the long delivery times are a burden on the company. AI-supported production systems can help solve these problems," says Christian Wolf, scientific director of the university's :metabolon project.

First, the project partners investigated which process parameters have a particularly large influence on the distribution of lead in the finished workpiece. "Since lead has a much lower melting point than bronze and is also significantly heavier, inhomogeneous lead distribution can occur during the cooling process, rendering the product unfit for use," Katrakova-Krüger explains. Relevant parameters include casting temperature, cooling conditions, rotation speed of the mould or quantity and temperature of the cooling water used.

Artificial intelligence makes optimisation suggestions

The project team also digitised the existing machinery so that the selected process parameters and machine settings can be automatically recorded and correlated with the casting results. An artificial intelligence was then trained with this data. "We can enter the geometry of the desired component into the resulting system. The AI then suggests relatively reliable parameters that have led to success with the same or similar components in the past," says Wolf.