The researchers of the Laboratory of Lightweight Materials and Structures of Peter the Great St. Petersburg Polytechnic University faced the problem while they were 3D printing. The process of the mathematical modeling takes lots of time and sometimes simple calculations could take a week.
In order to get rid of this problem they developed neural network which obtains 3D printing process parameters and ensure the stability of the process. Their software uses large data sets but their next step is to create an online neural system with automatical input of data sets and output of parameters to make it learn continuously.
Their neural network is already used in the assessment of quality parameters of manufactured parts such as quality of welding process or melting and transferring of metal.
Where is the money?
The algorithm developed by scientists from Saint-Petersburg could decrease the time of calculation process for all sphere of industrial 3D printing. Using their neural network there would be no need in lots of well paid specialists and, of course, the time for production of one detail would be decreased from weeks to days or maybe hours, so the rate of details per time would be way bigger. And the market of only industrial 3D printing is quite big. It is estimated to grow from $1.73 billion in 2018 to $5.56 billions in 2023. So, it is quite good idea to invest in this sphere.
Techers: Raisa Bestugina