A 3D print jobs can take several hours to complete. When a print fails, a user may be outside house or sleeping, and the printer will just keep running. That wastes can cause damage and even a fire hazard. This Ai-based failure detection solution utilizes the web camera of a printer or home computer to detect when a print job has gone wrong and started extruding ‘spaghetti’. It will then pause the process and alert the user through an email or text message.
The technology uses a deep learning model called YOLO to automatically detect print failures, and will pause the job for user to decide whether cancelling the job or not. After sufficient training, the algorithm is capable of generating coordinates of what it thinks are spaghetti errors with a confidence level. Currently it can run on cloud or on edge using Jetson Nano from NVIDIA. Historically it only had an inaccuracy rate of 6.9%, meaning it missed failures or flagged up false positives an average of 6.9 times every 100 prints. The team is working to maintain an inaccuracy rate of only 2% in the coming months.
It also comes with a host of other features so that you can print remotely with a peace of mind even when you are not home.
We are keen to explore R&D collaboration or proof-of-concept trials with SG local schools and institutes. We have previously worked with printing farms and provide solutions for trial in some US high schools.
It is a simple and user-friendly solution that can solve the big pains for 3D printer users. We have currently more 40.000 users in the builder community that support us.