It’s easy to feel overwhelmed by the rapid influx of new data-driven technologies being adopted by electric-vehicle manufacturers. From machine learning and artificial intelligence to predictive maintenance and predictive quality, it’s clear that the industry is undergoing a rapid transformation. How will these new technologies impact the industry? And how will these data-centric technologies help separate winners from losers in the years ahead? For some clarity, Automotive News Canada spoke with Greta Cutulenco, the chief executive officer and co-founder of Acerta Analytics. The company develops software solutions, driven by machine learning and artificial intelligence, that translate complex manufacturing data into actionable insights.
Sponsored Content From Acerta
Data Will Drive Success in EV Manufacturing
Greta Cutulenco
Q: Many technology companies from other disciplines are trying to cross into the electric-vehicle (EV) space. Acerta has been solely focused on the auto sector...How does that differentiate you from newcomers that are trying to get into EVs?
Greta Cutulenco: EV manufacturing is an attractive market sector for technology companies that customarily focus on other verticals – and especially so for companies that typically serve electronics manufacturers, since the number of electronics required in an EV is growing significantly. However, we have a couple of big advantages over them.
One, we have a proven track record of successful and scalable implementations within the automotive manufacturing environment. Our deep experience with the unique needs and realities of these facilities positions us well for a transition into EV component manufacturing – far more so than newcomers that are trying to enter the EV market from other industries outside automotive. We’ve built a solid foundation upon which we can dive into fuel cell, battery and electronics production for EVs.
Two, our focus on predictive quality is unique because we operationalize the use of machine learning (ML) and artificial intelligence (AI) in high-volume production – and it’s scalable globally. This is key in the automotive industry and will play a huge part in the growing EV transition. It’s also something that other newcomers will struggle with.
Q: Disruption in the automotive world is happening right before our eyes, and Industry 4.0 is upon us. Do you think this level of change is creating opportunities for companies like Acerta, a relatively new Canadian startup, to carve out a bigger niche in the auto sector?
Cutulenco: Absolutely. Many manufacturing companies have already put Industry 4.0 infrastructure in place, which allows them to collect and generate more data. But many of them aren’t actually using their data to its full extent. Manufacturers are struggling to derive actionable insights and value from the data that is being collected, which has provided a tremendous opportunity for companies like Acerta. By focusing on use cases that directly impact production quality, we help automotive manufacturers and part suppliers operationalize their manufacturing data to get real value from it, such as rework and scrap reduction or general staff productivity. And because predictive quality still is a niche category in the industry, none of the big incumbents have become the de facto standard yet, so it’s the perfect place for us to win in the market.
Q: There is so much new auto sector investment coming to Ontario, and billions of dollars are being directed toward EV and battery production. What effect do you think that will have on Canada’s importance within the global automotive industry?
Cutulenco: There is a huge opportunity for Canada to be a leader in EV and battery production on the global stage, but we need a solid strategy to get there as full EV adoption ramps up during the next few years. It is encouraging to see several OEMs already commit to new EV sites in Ontario. Personally, I’d like to see more investment from Ontario to support the Tier 1 and 2 suppliers, which would further help position the entire industry more strongly in the EV space. Additionally, Canada is already at the forefront of ML and AI innovation. Through continued investment in ML/AI for the EV manufacturing space, I believe we can become a world leader in the ML/AI required for EV and battery production.
Q: The automotive industry, and the automobile itself, is centered more and more on data. Does the role of data in EV manufacturing differ from its role in non-EV manufacturing?
Cutulenco: Data is important whether you’re building parts for internal-combustion-engine (ICE) vehicles, off-highway vehicles or EVs. With EV part production specifically, there is a need to use manufacturing data to accelerate speed to market. Data can help EV component manufacturers reduce test times, innovate faster during new-product introduction and ramp-up phases, understand how new part designs will impact production and introduce new products to market as quickly as possible.
A high level of precision is required when building EVs. For example, noise and weight reduction are both big problems for EV manufacturing, which means that manufacturers need new ways to innovate during production. More data will help them deal with problems quicker, iterate on designs, launch new designs and drive productivity. Manufacturers have had a hundred years to get to the point of stable, predictable production of ICE vehicle parts. If we want to be at full EV production in, let’s say, 10 years, we need faster innovation — which means manufacturers need data, and they need AI.
Q: Have you observed a difference in the uptake of AI within automotive manufacturing due to the EV transition or hype around generative AI?
Cutulenco: Apart from vision systems – where AI seems to have been accepted more fully in automotive manufacturing – we still see limited adoption of AI and ML in manufacturing environments. Many still question the ability to operationalize AI solutions in production, and a lot of effort is still required to drive trust in these solutions with line engineers. Having said that, we are making headway, and the EV transition is helping accelerate adoption. We’re seeing manufacturers move away from one-off AI projects, and we are leading the way in terms of operationalized ML solutions. We’re also expanding to more plants within our growing customer base. Moreover, we are seeing more acceptance of the idea of “predictive quality” in manufacturing, which is a concept we have been touting within the industry for years. I guess you can say we like to be ahead of trends.
Q: Data in automotive can be used for both predictive maintenance and predictive quality, but they are quite different. What are the benefits of predictive quality versus predictive maintenance?
Cutulenco: More people are familiar with predictive maintenance because it has been popular in manufacturing for some time. Predictive maintenance comes into play when manufacturers monitor activity on their equipment and use the data to predict when machines will wear down or fail, so they can perform maintenance ahead of time. Focusing on machine uptime and overall equipment efficiency is important, but it doesn’t help manufacturers with the core lifeblood of their business: the parts they produce. There is a tremendous cost to poor quality, from wasted materials and labor to brand damage and customer churn. Investments in predictive quality can have a significant impact on the bottom line of any plant, more so than an investment in predictive maintenance alone.
Through advanced analytics made possible with ML and AI, predictive quality enables us to identify anomalies in manufacturing – and do so in real time, so we can help a manufacturer understand when defects will likely occur or determine what is most likely contributing to a quality issue. We help them identify the root causes of defects and address potential quality issues proactively, not after a problem has already occurred or when a faulty part shows up at an end-of-line test station. By focusing on predictive quality, engineering teams can make adjustments and implement solutions before potential problems ever impact production.
Q: Why is predictive quality extra-important with the manufacturing of EVs?
Cutulenco: Again, in today’s race to EV adoption, speed to market is everything. So if an EV part producer can optimize production and ensure high quality and first-time through rates without rework, it puts them in a better position to beat their competition. And because manufacturers are investing in new EV lines and plants, now is the perfect opportunity to introduce advanced digital technology. It’s much easier to do things right from the start, rather than try to revamp lines later. For example, choosing a manufacturing execution system (MES) and equipment that makes it easy to share and export data allows manufacturers to gain value from predictive quality much faster, which will enable speed, productivity and the ability to ramp up and scale production.
Q: In the highly competitive automotive world, the manufacturing process is closely guarded. In broad terms, can you provide an example of how Acerta’s AI platform predicted or detected a malfunction and share the end result?
Cutulenco: We can’t share specific details about most of our work with OEMs like Nissan and Tier 1 parts suppliers like Dana, BorgWarner and Linamar. But you can check out various case studies on our website (www.acerta.ai) to learn how we’ve optimized production in different automotive manufacturing plants that were facing different quality issues.
I can tell you a bit about our latest innovation, which is in alternative fuels. We’re working with Ballard Power Systems on a project that’s supported by NGen AI for Manufacturing funding. Our new technology will help Ballard shorten their factory acceptance testing times and identify the sources of test failures in their fuel-cell stack assemblies. This will be a multiyear project, and we’ve already completed the initial proof of concept (POC) with great results. Tests that previously took 21/2 hours are running in less than 30 minutes. Ballard is excited to move forward because our new approach will not only reduce their factory acceptance test times by up to 80%, like we saw in the POC, but it also will enable them to deliver industry-leading quality to their customers at much higher speed.
ABOUT THE PANELIST
Greta Cutulenco
Chief Executive Officer and Co-Founder
Acerta Analytics
Greta Cutulenco was named to a Forbes 30 Under 30 list and was named an Industry All-Star and Canadian to Watch by Automotive News Canada. Cutulenco is a member of the Automotive Parts Manufacturers’ Association board of directors and holds a bachelor’s degree in software engineering from the University of Waterloo.www.acerta.ai.
