GM considering bringing power electronics production back in-house
24 June 2014
GM is almost two-thirds of the way through a $16.6-million project ($6 million in funding support from the DOE) to design and to build a next-generation power inverter capable of 55 kW peak/30 kW continuous power. The inverter is to improve the cost of the power electronics to $3.30/kW produced in quantities of 100,000 units, and the power density to 13.4kW/l, and a specific power of 14.1kW/kg, with an efficiency >94% (10%-100% speed at 20% rated torque) to meet DOE 2020 goals.
In his presentation on the project at the DOE 2014 Annual Merit Review in Washington, DC, Sean Gleason from GM noted that General Motors has not made prototype power electronics in an internal production facility since 1999. The current work on the Next Gen Inverter Project has put GM on the path to considering bringing power electronics production back in-house, he said. The project, which began in October 2011, will wrap in January 2016.
For the project, GM is working with Tier 1, 2, and 3 suppliers (Hitachi, Delphi, Infineon, HRL, Panasonic, AVX, Kemet, and VePoint) along with the National Renewable Energy Laboraotry and Oak Ridge National Laboratory to co-develop technology that reduces cost and increase efficiency, without increasing volume or mass.
The inverter is intended to be modular and scalable to meet all vehicle applications.
Reamining work includes a critical design review; early build and test; final technology and production cost assessment; and delivery of 3 next-gen inverter units for testing.
Resources
(DOE will publish presentations from the Annual Merit Review on its website.)
Sean Gleason (2014) “Next Generation Inverter” DOE 2014 Annual Merit Review Project APE40
Good for you, GM -- this is something that any major auto manufacturer must be able to do with the same aggressive targets for cost, quality, variability, production footprint, etc. as they do for any present core competency. Power electronics fabrication will soon become as essential (and in some cases already is)as body stamping, complex casting, metal joining, and other battleground manufacturing segments.
I remember (about 1990) when some colleagues at a client aerospace company had a meeting with Alan Cocconi and dudes from Hughes Electronics to discuss power electronics. They revealed a target of $10/kW (this in 1990 economics -- about $19/kW today). My buddies found this hilarious. They were hoping to get down to $200/kW from $1000/kW. Notwithstanding aviation requirements, the chasm looked to be infinite.
At that time the jewel project at the client was a 300kW inverter intended for an electric torpedo. It was about 45cm dia., 65cm long. As I recall the lab unit was not to be operated without the presence of the lead Project Engineer. Nobody else in the industry had matched their work.
I assume their metrics are based on steady-state ratings. If so...
$100
2.25L
2.12kg
This probably doesn't include various inductor/capacitor components for power conditioning, but still...
Wow. How times have wonderfully changed.
Posted by: Herman | 24 June 2014 at 10:37 AM
Great story Herman.
I've been convinced for several years that the reason the majors have been slow to the party with electric traction is that they don't own a significant (competitive) IP power electronics portfolio and manufacturing capability.
This news item supports that view, and hopefully is one of the markers for a sea change.
I realize that the biggest factor is battery energy density & price, but there seems to be a similar tidal change there too.
I guess that's what happens when there's real competition.
Posted by: electric-car-insider.com | 25 June 2014 at 12:21 AM
e-c-i:
I think you're definitely correct about the pwr elex design/mfg shortfall.
Back in 2003 or so, the same company I mentioned above had racked up a number of significant contract wins to develop aircraft electrical systems, with some real cost challenges facing them. One of their investigative actions was to purchase a Toyota Prius and tear it apart to see how TM could sell a whole car for the price of what an aero company would spend just to make the propulsion motor inverter/controller.
It was very revealing. Without getting into details, the investment in Toyota's Hirose facility (as well as partners like Denso) had clearly paid off with some well-packaged, robust, and affordable assemblies. While many of the manufacturing techniques and component selections would not pass muster for aviation, the guts of this modest car were humbling.
When Uchiyamada-san took on Project G21 in 1994, his team was looking at the complete automotive industrial framework needed to bring this car to the world. Toyota remains THE industry leader in electrification (Tesla's narrow market niche notwithstanding). It enrages me to see forum denizens insulting Toyota as a "shill for big oil" because of their caution in committing to BEV battery technology. From the expense of hundreds of $M for battery technology licensing fees/penalties to the had-to-hand combat required for power electronics affordability in TRUE mass manufacture (feeding hundreds of cars per shift with six sigma yield) they have led the world, and still do.
I am ecstatic to see GM aspire to the same.
Posted by: Herman | 25 June 2014 at 07:41 AM
I won't insult Toyota or anyone else, but I will say that it is really disappointing to see such a tiny battery in the Plug-in Prius.
Posted by: electric-car-insider.com | 25 June 2014 at 02:21 PM
Nothing wrong with picking on TM when they deserve it. I agree. The rationale for the PiP battery sizing is clearly CARB ZEV points first, consumer second.
Posted by: Herman | 25 June 2014 at 03:07 PM
Investing in semiconductor fab facilities etc. seems excessive and expensive. Why doesn't GM just buy an Intel competitor and move on?
Posted by: william g irwin | 29 June 2014 at 04:30 AM
William, a couple of things...
First of all you're right: the auto builder doesn't need to manufacture the semiconductor itself. Ideally, they will probably do as Toyota does (or at least did -- not sure today) and purchase the devices on a wafer from the manufacturer, cutting the individual components and mounting them. MHI can manufacture their own devices, though interestingly they are not a major EV seller yet. We'll see where their SiC work takes them.
Secondly again this is POWER electronics we are talking about, not processing. Before anyone points out an error on my part: yes, there are certainly processing elements in the inverter/controller product. But the supply chain for processing power from enterprises like Infineon already exists. What auto companies need to do is not to fabricate components, but to design, spec, select, and assemble fully integrated power electronic products with the same relentless focus that they do with transmissions, driver/auto interfaces, etc.
Posted by: Herman | 30 June 2014 at 11:55 AM