When manufacturing complex products, engineering teams typically need to establish a baseline monitoring process for equipment to ensure stability and productivity throughout the entire operation. In general, the simulation of a build takes longer than the actual build process itself, making process monitoring a priority for making design decisions. However, as product complexity increases and the tools needed require more intensive monitoring, efficient quality management can become an obstacle for manufacturers.
To address this challenge, Renishaw and Altair established an Al-driven quality assurance process using machine learning and advanced digital gauging to accelerate product development and production.
Challenge
During the additive manufacturing process, the laser fuses layers of powder together while the spectral emissions of the melt pool are monitored to ensure each subsequent run corresponds to the original.
Theoretically, teams could take the samples from one build (or print) and compare them to subsequent prints and determine whether or not process variation is occurring. The problem with this is that there is an incredible amount of data that gets produced during a print. For example, a typical print run samples at least 100,000 times per second, or about 20,000 samples in the time it takes to blink.
Generally, this data is filtered, and relatively basic statistical models are applied to try and derive meaning from it. This provides some useful information, but due to the sheer volume of data being created, it essentially becomes impossible for a human to efficiently glean in-depth insight from the data.
Renishaw leveraged Altair's AI technology that could flag anomalous builds and regions by analyzing spectral data in real time, allowing for quicker part development and more stable production. As machine learning algorithms have become more accurate with increased computation power, engineers can now depend on ML technology to make decisions from data without compromising floor efficiency.
Solution
The Altair signalAI tool was used to successfully detect anomalous builds during the metal additive manufacturing process. Without any prior knowledge of previous anomalous build patterns, the software was able to determine the anomalous volume and region, as well as reveal the specific 3D region of interest with further analysis. The technology's machine learning capabilities actively monitored the data from print runs, so manufacturing teams no longer had to physically compare datasets to find anomalies, leading to significant cost and time savings.
signalAI's anomaly detection software can be used in real-time locally or remotely on the cloud and includes pre-processing capabilities, allowing complete model build and analysis to be completed within a single platform.
Implementing data-driven decision making from the ground up allows design and manufacturing teams to holistically analyze processes to improve an organization's smart factory practices. Looking forward, an Al-driven manufacturing approach offers endless possibilities for machine improvement, such as identifying the best sensor type and location for an application, and shared learning to improve processes within an enterprise with a hybrid cloud approach.
For more information, visit: altair.com/ai-powered-design/