How engineers cut development time in half for an EV fire warning system
To meet emerging safety requirements for electric vehicle (EV) batteries, Amphenol applied fast engineering development principles to design gas sensor modules that reliably provide advanced warning of lithium cell thermal runaway.
A typical product development process involves finding a lead customer, creating a custom design for the application space and then (hopefully) propagating the final design through the rest of the industry.
But with safety looming large as a major headwind due to growing concern over the potential for EV battery fires, engineers at Amphenol knew time wasn’t on their side. The challenge was to design a system to detect when a battery cell is failing beyond a recoverable state and threatening to initiate a fire—fast.
But scant data existed on the phenomenon of lithium-ion battery fire propagation, and what little data existed was inconsistent.
“We knew that we needed to collect as much information as possible, as early as possible, and that meant getting prototypes into the hands of many, many companies quickly,” says Brian Engle, transportation business development manager for Amphenol Advanced Sensors.
EV fires are complex
The issue with EV battery fires is that they are complicated. Two years ago, when some of the more catastrophic incidents began grabbing the headlines, the physics around the thermal runaway conditions in batteries were not well understood.
“There were theories out there, but we couldn’t be sure what sensing technology would work best and what the industry ultimately would settle on,” says Engel.
Coming up with an effective sensor package would require collecting information on the environment the sensors would be exposed to, the excursions likely to be indicators of problems, and the effectiveness of the detection technology in responding in a meaningful and reliable way.
Hedging its bets, Amphenol decided to leverage two existing development platforms with a variety of different gas sensors, along with temperature and relative humidity sensors.
The goal was to engage with a wide base of OEMs, as well as battery and battery management system (BMS) makers, and Amphenol started by meeting with customers, sharing the specs of the platforms and talking through their functionality. When customers requested parts, Amphenol provided those free of charge when only a few were needed and a critical test was coming up.
In exchange for access to test data. Amphenol committed to providing engineering support during critical test and assistance with diagnosing any issues that came up. “Discussing what they observed and what we observed immediately after a test helped to advance everyone’s understanding,” says Engel.
To keep customer data confidential, Amphenol hosted the data on encrypted servers and only shared findings in the aggregate.
Click image for larger version.
Staying close to customers and working through their early development phases was key to building trust and confidence on both sides—but that came at a cost.
“The challenge was basically the amount of engineering resources we had to provide in order to babysit hundreds of prototypes that we shipped out into the field. But the benefit was that we learned at an incredibly accelerated rate, maybe 10X faster. What would normally have been three- or four-year development cycles was accomplished in two years,” Engel stressed.
Learning on the fly
Customers very quickly coalesced around certain things that were important to them from both a software and hardware perspective, which was useful in maturing the design.
“We started out with a base design for a mechanical package and very quickly learned that for certain customers the package needed to be modified quite substantially in order to fit inside some of the more confined packs,” Engel explains. “So, we ended up with a configuration of the electronics—basically a fir tree mechanical mount—that allowed us the flexibility to have the same content in a low-profile version.”
Anticipating different customer needs, Amphenol didn’t push a plastic box solution on everyone; some customers wanted just the core sensors to place directly into their BMS.
Amphenol was also learning about the rate at which companies needed to acquire data and how much power they were willing to consume. One area of concern was having the device run when the vehicle was off, risking the potential to drain the auxiliary battery. “So very early on we decided we had to have a dual-mode operation, allowing us to go into a limited power consumption mode that turned the CAN transceiver off when the vehicle was not running,” says Engel.
Amphenol learned other things, too. When a thermal runway condition exists, the humidity jumps by 30% and a lot of water vapor collects in the pack when the cell vents. And in off-road applications, the battery cells are potted and could get damp, requiring some sort of splash protection.
“Everybody is going to want tweaks because their applications vary,” says Engel. “And in a typical development cycle that causes delays because you’re only learning about them after the fact. Here, we were learning about these things on the fly.”
One major finding involved some of the dynamics that occur leading up to battery cell failure, exposing a broader physics problem than just fire propagation. The vehicle in question had a traction battery that was undergoing a thermal runaway event when Amphenol engineers saw a reset in the CAN transceiver.
“The customer said they were not too worried because their BMS controller reset at the same time. We found it awfully hard to ignore the data,” explains Engel. “And that’s basically when we learned about this arc discharge phenomenon that causes the battery cell to evolve into highly ionized plasma, effectively turning the inside of the battery pack into an arc welder.”
The big lesson was that thermal runaway can be the result of a physics problem. One of the physics problems that evolves from this release of gas is that as it comes into contact with a high voltage conductor and a lower potential area, the electrons will conduct directly through the gas in an arc flash event.
“It took some time to collect enough data, but we ultimately discovered that this was not an anomaly and in fact was a fairly common occurrence. That’s when we reached out to OEMs and an atmospheric expert at NASA and learned very quickly that it could be modeled and understood. From there we were able to come up with passive design features that could fix at least part of the issues with thermal runaway.”
This led to the development of sensor modules that reliably detect CO2 and hydrogen from cell decomposition venting for early warning of lithium cell thermal runaway. Gas sensing is also ideal because CO2 and hydrogen are produced early in the battery cell decomposition process, versus an indicator like smoke, which either appears late in the process or not at all. Further, the volume of hydrogen and CO2 released during venting yields concentrations much higher than background, making detection relatively straightforward.
When the industry started to experience battery recalls, Amphenol engineers realized their design had a distinct advantage in that it could detect an impending cell failure anywhere from 40 seconds to a minute faster than the pressure-based systems that other companies were working on.
Fast engineering development principles pay off
Engel acknowledges that this approach to prototyping adds a lot of extra work on the engineering side because you are tracking multipole programs in component-level release as well as assembly-level release. The value in doing that is you end up with a family of products that are more of a fit across the industry in certain common design approaches.
“This approach really allowed us to achieve that commercial success, which was the ready adoption of the product by the customers. We learned at a rate that was unheard of for our business,” concludes Engel. “I guess the best way I can describe it is that it was our voyage to revenue.” Behind the recent headlines of EV fires, engineers are working to get in front of the thermal runaway conditions that lead to battery failure by developing technologies that provide early warning of an impending catastrophe. Some detection platforms are currently in production.
About the author:
Brian Engle is the Transportation business development manager, Amphenol Advanced Sensors.
