Il Yong Kim
Goldilocks and the 288 Batteries: Not Too Hot, Not Too Cold
Electric vehicles are regularly touted for their lack of emissions; the absence of anything like the noxious fumes generated by internal combustion engines. One thing they do emit, however, is heat; specifically, the heat put out by hundreds of on-board batteries that store energy for such vehicles. Managing this heat has become one of the key challenges to the success of these vehicles on the road.
And while heat is a normal consequence of the electrical current these batteries provide, performance is also affected by the surrounding temperature. The search for an optimal operating condition is something like Goldilocks’ taste for porridge that was “just right.” For batteries, that means around room temperature, 25-45 degrees Celsius. When it is much colder, say -20 degrees Celsius, resistance within the battery increases and its output diminishes. When it is much warmer, up to 60 degrees Celsius, the chemicals within the battery begin to degrade at an accelerated rate, decreasing its ability to return to a full charge after just a few dozen re-charging sessions. The battery’s life will be significantly shortened, compromising the long-term value of the vehicle it powers.
In other words, manufacturers have a vested interest in keeping batteries warm when it is cold outside, and cool when the battery becomes hot during operation. For just this reason, General Motors asked Dr. Il Yong Kim, professor of mechanical and materials engineering, to examine the best way of applying this Goldilocks principle.
“They have a good vision of long-term research,” he says, noting that the firm is the only auto-maker to have a technical research centre based in Canada.
The reference design of Kim’s work is the battery array for the company’s flagship electric car, the Chevrolet Volt. This structure is made up of 288 flat lithium-ion battery cells stacked together in the shape of a “T”; at 1.8 metres long and weighing some 190 kilograms, it is one of the vehicle’s largest single components. Its ability to store electricity determines how far the vehicle can travel without charging; its ability to hold a charge repeatedly determines how long this part of the vehicle will last without expensive maintenance.
The housing around the battery already contains a heating and cooling system that pumps liquid through channels to raise or lower the temperature in order to maintain the best performance. Although the system works well enough, GM approached Kim several years ago to take a closer look at the design of battery thermal management systems for next-generation electric vehicles. The topic, he explains, represents a frontier of automotive development.
“This research is still in its infancy,” Kim says. “At this point there are multiple design concepts – fundamental design concepts – for the thermal management of electrified vehicle battery systems.”
For example, while the Volt uses liquid to manage heat flow, a competing electric vehicle, the Nissan Leaf, uses only air. It remains unclear which approach is more effective.
Kim is addressing this question from the perspective of his speciality, a field called design optimization. Through intricate computer modelling, he and his colleagues are examining the physical characteristics of a battery array design that will provide the very best performance.
Among the details to be considered is the shape of the channels between batteries, where heat is conveyed to and from the array. Such studies represent a classic problem in computational fluid dynamics, a well-established engineering discipline that attempts to come up with quantifiable mathematical descriptions of complex flow patterns.
That task has always been difficult, although less so in recent decades as increasingly powerful computer technology processes these intricate mathematical models. Even so, when it comes to finding the optimal version of a temperature management system, it might be necessary to run through such models thousands or even tens of thousands of times to get an accurate picture of how a particular design would perform under real-world conditions.
“This could take as long as a year, for one optimization, unless you use an efficient method,” says Kim, adding that this kind of timeline can stretch the patience of a commercial operation like GM.
There is more at stake than improving the vehicle’s efficiency. If a battery gets too hot, it can generate additional heat, creating a feedback loop of continuously rising temperatures. This state, known as thermal runaway, can lead to fires and even explosions. Early in 2013, inadequate heat-venting in a much smaller array of lithium-ion batteries led to thermal runaway and an in-flight fire aboard one of Boeing’s latest passenger aircrafts, the 787 Dreamliner. The entire fleet was grounded until this design shortcoming could be resolved.
Kim is conscious of such disasters as he considers the batteries in electric vehicles. One preventative measure is a system of cooling plates inserted between batteries, with openings to drain excess heat before thermal runaway can begin. Yet these plates add weight, necessitating the need for a larger array – and more plates – to compensate.
“In a real-world design, it’s a really complex problem,” he says. “You want to minimize the weight, but at the same time your battery needs more power.”
Far from being daunted by this complexity, Kim welcomes the chance to move into uncharted design territory. He is preparing for the next phase of the project, which will take the already sophisticated computational models for the temperature dynamics of individual battery cells and combine them with the equally sophisticated models of the entire battery array. This unified model, he observes, will be unprecedented.
“That has never been done before,” Kim concludes. “I can say this with confidence.”
Counting the Uncountable
Tracking the flow of liquids and gases is a game with infinite possibilities. In order to cope with those possibilities in a practical way, engineers such as Il Yong Kim resort to a method known as finite element analysis, which combines large numbers of relatively simple mathematical functions into a more complex function that approximates the behaviour of an actual flow. In order to obtain this approximation, the curved, unmeasurable lines of this flow are rendered in the form of a large number of short, straight, and measurable lines. By carrying out these calculations over and over again – thousands of times, if necessary – the results get closer and closer to a correct picture of the desired flow.
Profile: Tim Lougheed
(e)Affect Issue 4, Fall 2013