Development and application of advanced AI decision making for manufacturing

Born from the work of William Edwards Deming in postwar Japan, statistical manufacturing has remained largely unchanged since then. The advent and propagation of artificial intelligence and deep learning allows for nonlinear feedback and feed-forward systems to be integrated into statistical process control for real-time monitoring and evolution of each part assembly, based on individual build conditions. Industrial 3-D printers are typically governed by classical control systems that aim to normalize the production fluctuation of the entire part, not improve error correction and tolerance from layer to layer.

Vadim Pinskiy explores an AI platform that can be applied to existing 3-D printers that’s capable of detecting and classifying spatial errors at any printed layer within that part and uses that data to automatically adjust in the inner structure of subsequent layers to maximize the performance of the entire part. This system is able to increase the overall yield of a production process and reduce the overall variance of the produced parts, can be used in any 3-D printing application, and isn’t limited to adaptive manufacturing.

This is the power of integrating reinforcement learning and other advanced AI methods with 3-D printing to create parts that are of controlled quality and whose production is governed by the performance of the part and not just its appearance. This advance is critical for the continued development and proliferation of 3-D printing into medical and other regulated industries.