Predictive Analytics for Manufacturing Processes
The application of quantum computing to predict faults in the manufacturing process of advanced materials.
Project Overview
Updated March 31, 2023.
The Problem
The root cause of failures in product testing is often difficult to determine, particularly when the failure signals are sparse relative to the available background data.
Compounding the problem, the process must meet a variety of specifications for multiple customers simultaneously.
How We Are Solving It
The Predictive Analytics for Manufacturing Processes project aims to create a digital twin of the metal finishing line, leveraging predictive analytics to analyze data captured from the process line, such as chemical compositions, temperature and voltage. This technology will provide new insights to help optimize the manufacturing process.
The project is led by D-Wave in partnership with Avcorp Industries, Solid State AI and Simon Fraser University. The team is leveraging its research capabilities in data analytics, predictive analytics tools and advanced machine learning techniques on a quantum computer to address situations when failure signals are sparse, relative to the available background data. These tools will move Avcorp’s manufacturing fault detection processes from reactive to predictive.
This project will demonstrate predictive capabilities that can also be deployed in other manufacturing processes, including industrial manufacturing and semiconductor fabrication plants.
The Result
This project created a digital twin prototype of a metal finishing line resulting in new insights from data captured from the metal finishing process such as chemical compositions, temperature and voltage for optimized large equipment manufacturing processes, including more effective mapping, cleansing, and processing of industry data. Through the project, Solid State AI developed and commercialized AIMS (Artificial intelligence for Manufacturing Systems), which allows users to import, visualize, and execute machine learning models on manufacturing data. The newly developed approach has the potential to be applied to aerospace manufacturing and other sectors including semiconductor, automotive, oil and gas, and mining.
