Acerta to help Ballard shorten factory acceptance test times and identify sources of test failures in PEM fuel cell stack assemblies; POC showed 80% improvement
— Andreas Putz, Principal Data Scientist, Ballard
KITCHENER, ON, CANADA, July 6, 2023/EINPresswire.com/ — Acerta Analytics – provider of predictive quality analytics software solutions for automotive manufacturing – is pleased to announce a multi-year project to help Ballard Power Systems shorten factory acceptance test times and identify the sources of test failures in their PEM fuel cell stack assemblies.
The project is supported in part by AI for Manufacturing (AI4M) funding through NGen, the non-profit organization leading Canada’s Global Innovation Cluster for Advanced Manufacturing which promotes commercialization of artificial intelligence and machine learning innovations across Canada. This funding will enable Acerta to productize more AI for manufacturing.
“As a proof of concept, we developed custom machine learning models that ran on a subset of Ballard’s test data and demonstrated that we can reduce test cycle times by 80%. With the help of machine learning, acceptance tests that previously took 2.5 hours were running in under 30 minutes,” said Aaron Alberts, Senior Account Executive at Acerta.
“We proved that our models could identify a critical defect mode, predict good and bad parts, and shorten test times significantly, all while adhering to Ballard’s strict testing requirements,” Alberts continued.
“Our early results look very promising so we’re excited to start moving from Proof of Concept to the next phase of the project – and we’re doubly excited because NGen accepted our joint proposal for funding. Over the next two years, Acerta will be our backbone to create vastly improved models, manage their deployment onto our factory floor, and continuously monitor and correct model quality. This new approach will drastically reduce factory acceptance test times for our different products and will enable us to deliver industry-leading quality to our customers at significantly higher speed,” said Andreas Putz, Principal Data Scientist at Ballard.
To meet their high throughput demands, Ballard runs factory acceptance tests across multiple parallel test stations which compounds the number of machine learning models required. Acerta will host models on edge devices in test stations and facilitate communication between IIOT devices and the cloud where Acerta’s LinePulse will monitor and allow maintenance of machine learning models. Acerta will use Telit Cinterion’s deviceWISE to manage edge-to-cloud data transfers.
About Acerta
Forged from industrial experience and driven by data science, Acerta assists precision manufacturers to take their digital transformation beyond manually crunching sensor data. Our machine learning and AI-powered software solutions translate complex product data into actionable insights and enable companies to make the right decisions fast, optimize production, and improve product quality. Founded in 2017, Acerta Analytics Solutions Inc. is based in Kitchener, Ontario, Canada. We’re on a mission to understand the digital thread of data for every vehicle. Visit www.acerta.ai to learn more.
About Ballard Power Systems
Ballard Power Systems’ (NASDAQ: BLDP; TSX: BLDP) vision is to deliver fuel cell power for a sustainable planet. Ballard zero-emission PEM fuel cells are enabling electrification of mobility, including buses, commercial trucks, trains, marine vessels, and stationary power. To learn more about Ballard, please visit www.ballard.com.
About NGen
Next Generation Manufacturing Canada (NGen) is the industry-led, not-for-profit organization leading Canada’s Global Innovation Cluster for Advanced Manufacturing. NGen is dedicated to building world-leading advanced manufacturing capabilities in Canada, for the benefit of Canadians. For details about NGen’s $50M AI for Manufacturing funding, visit https://www.ngen.ca/funding/challenge/ai4m.
Heidi Marr
Acerta Analytics
+1 519-341-6080
[email protected]
Visit us on social media:
Twitter
LinkedIn
The content is by EIN Presswire. Headlines of Today Media is not responsible for the content provided or any links related to this content. Headlines of Today Media is not responsible for the correctness, topicality or the quality of the content.
