However, the focus has been on physically measurable waste, be it materials or time lost on machines when a component has to be scrapped or reworked. What is often disregarded is the large and growing amount of data that has been going to waste because it has been so difficult or expensive to analyze.
Automotive Industries (AI) asked Canvass Analytics founder and Chief Executive Officer Humera Malik how artificial intelligence (AI) allows manufacturers to monetize data.
Malik: First, AI provides the ability to predict into the future. Secondly, AI brings the ability to discover and study correlations between various parameters in complex processes. In the auto industry today, there are many advanced process automation and condition-based monitoring and diagnostic systems. However, when a measurement is made or an alert is produced, it is based on something that’s already happened. This is where AI and its ability to forecast future values is fundamentally changing the game for manufacturers. By enabling proactive actions to be taken, AI ensures a desired outcome is achieved rather than reacting to an undesired event. As AI models continue to learn and self-optimize in an environment, they can make suggestions and even take action to make sure the desired state is achieved.
AI: Where is unused data to be found in the typical manufacturing plant?
Malik: There’s unused data in almost every area. From machine tools, to robotic welders and dosers, to stamping presses, stackers and 3D printers – and it’s not limited to production lines. Even areas such as energy production or consumption and environmental control systems are sources of data that present great opportunities for optimization and cost savings.
AI: What is the true value of this data?
Malik: In most cases, challenges such as quality and yield variations, scrap rate, machine failures, out of tolerance parts, etc. can be explained via the data. However, today, process control and automation solutions work on limits that are often too relaxed to catch small variations in process parameters. This is where AI is proving its potential to truly provide transformation within the automotive industry.
Imagine if 100% of parts or assemblies produced were “inspected” via the production data or if bad parts were stopped in mid production using real time quality prediction. Issues such as recalls, warranty claims, etc. that currently cost the industry billions of dollars and drive up the product cost could look very different. In fact, we have a number of examples where AI has helped address production issues that otherwise seemed impossible to solve. This is not a future vision or dream, applied AI is in use today with proven examples.
AI: Do you have automotive industry-specific solutions?
Malik: We have worked with a number of auto manufacturers and parts suppliers globally. Some of the examples include optimizing welding and predictive weld quality, assembly line predictive quality for functional tests (in-line and end of line testing), additive manufacturing quality inspections, and machine tool predictive maintenance, amongst many others.
AI: You are a relatively new company – what makes you different?
Malik: One of our biggest differentiators is the fact that we can operationalize and apply AI in industrial settings in less than 30 days and do so at scale, which creates a very low risk to entry and accelerated ROI. We have seen ROI ranging from less than a year to 24:1 in year 1.
AI: What is the significance of your partnership with Microsoft?
Malik: Canvass Analytics recently announced its partnership with Microsoft, where we’ll be collaborating to help industrial companies achieve the benefits of Industry 4.0. By using Azure with Canvass Analytics’ AI-powered predictive analytics, our customers will transform their Industry 4.0 data strategies, enabling them to derive commercial value out of the massive amounts of data in their factory floors.