Have you considered using a battery/capacitor hybrid system? These are being used in fuel/electric hybrids now, but could they be an advantage for BEVs as well?
Electro-chemical double-layer capacitors AKA ultra capacitors or super capacitors have a very large power density for both charge and discharge. When used in conjunction with a high energy density battery, the combination would be able to provide high acceleration as well as capturing a very high percentage of regenerative braking energy (Something a battery alone does not do too well). As far as I can tell, by providing a better means to supply/store peak currents, the battery life and range should also be extended.
It seems there are a couple of types of these capacitors, Asymmetrical and Symmetrical. The Asymmetrical units do not require balancing networks when used in series, but cannot be discharged fully under normal use. The Symmetrical units require balancing networks, but can be discharged fully under normal use. Both types seem to have a cycle life of 10,000 to 100,000 and beyond.
A Canadian company called Tavrima makes these capacitors for EVs and hybrids. Seems that a few electric drag cars use them without any batteries at all. Good application for a high power density device.
http://www.tavrima.com/home.html
Yet another company called JEOL has developed an epoch-making capacitor (Nanogate Capacitor) with an energy density of 50-75Wh/kg, which is on par or exceeds some batteries and has a cycle life of around 10,000. They have now formed a new company called Advanced Capacitor Technologies, Inc. This technology is still in the laboratory and not in production. I have a feeling this will be a very costly unit.
Michio Okamura of Japan has been working with specialized electronics used to get the most out of "super" capacitors used for energy storage. They call them Energy Capacitor Storage Systems (EcaSS). He has his own web site that is very informative about his work. http://www.okamura-lab.com/ultracapacitor/edlc/ecass1Eng.htm
I think you would need to size the capacitor bank to match or slightly exceed the energy requirements for acceleration and braking. Then the battery bank is mostly used for the range requirements.
Seems to me that either the capacitor or battery industries are close to a good solution, or maybe a combination of both. The capacitors do have the great advantage of being able to be charged very quickly.
-Bill- [ Parent ]
They're a better idea for lithium systems than for lead-acid, as lead-acid is actually very good at delivering high power on demand. The units I've bought are claimed to deliver 3800A at peak. Since I only paid UK260 (about $470) for the whole battery pack (including five spares) I think it would be difficult to justify the expense.
The other area they're worth considering is regenerative storage: but the area where I live is not particularly hilly, and I'm expecting to be able to develop my own electronics to give good regeneration characteristics.
One problem is that the most simple circuit for controlling motor current is a "buck" circuit, where a switch and an inductor are used to step down the voltage. During the off phase the battery is not connected, and the current is "stored" in the inductor, but during the "on" phase the battery must deliver the same current as the motors require. This means that when I stamp my foot at the lights the motors may be getting 1200A, but that is only going to be at about 40v (largely resistive), which means that a 180v battery will only be drawing an average 270A or so. The problem is that each motor controller will draw 600A at times, so if they both come on at the same time they'll draw 1200A.
The solution is to use a pseudo-random generator to gate the current sense, (which also means the motor windings hiss instead of whistling) and to use several smaller current control modules. It is a property of the type of pseudo-random generator that different modules can be connected to delayed versions of the bit series, which means that if there are 6 100A modules per motor, only 2 or 3 of the 12 will be on at any given time, even at that peak load. That reduces the load on the battery considerably, but uses the inductors as the energy storage devices to smooth the load.
A similar argument applies to regeneration. Now we use a "boost" circuit rather than a "buck" circuit, but again the current comes in peaks. These peaks are averaged out by using the delayed pseudo random sequence, which enables the batteries to be charged much faster, since the maximum charging current limits the average power, not the peak motor current. This should enable me to recover much more energy on downhill stretches.
That should enable me to get by without ultracapacitors, since the motor inductors will be fulfilling the same role.[ Parent ]
That is interesting, most of the DC EV motor controllers I have seen are a simple low side switch employing either FETs or IGBTs. These require the motor to be able to handle the highest battery voltage. They all seem to use PWM to control the motor speed. I have seen some very nice designs that use multi-phase switching (6 phases in some cases), where each Fet is only on for a portion of the PWM output pulse. This way, they can run at a nominal 15 KHz or so, but since each Fet is on at a different time, it has the effect of increasing the ripple frequency and it allows them to use smaller input and output capacitors as the ripple is much easier to filter. This also pulls current from the battery at a more constant rate.
I have been thinking about the regenerative braking, from what I understand, you waste about 30% of your power in braking. Out of the 30%, you can reclaim 50% or better using regen. I have been mulling it over and an external regen circuit may be easier to encorporate, possibly using capacitors to store the impulse energy to feed back to the controller when it is ready. Building the boost circuit to handle the surge currents as well as being able to control the amount of re-gen (brake smoothness) would be the fun part. I have not heard of too many commercial motor controllers with a re-gen that was not problematic.
Bill[ Parent ]
Yes, my understanding is that the motor itself is an inductive load, so the inductor part of the buck circuit is incorporated in the motor, as it were.
The problem with "rippling" the output without an inductor is that there is always a FET on, so the current never varies. It is the varying current in the "buck" circuit that enables the step-down from one voltage to another. To get over that, I'm expecting to use an inductor per FET block, separate from the motor's inductance.
I have been thinking about the regenerative braking, from what I understand, you waste about 30% of your power in braking. Out of the 30%, you can reclaim 50% or better using regen.
A quick calculation shows that for a one-ton car, the kinetic energy at 75mph is about 0.15 of a kWh. Unless the journey consists of a series of rapid changes from 75mph to 0 and back to 75mph again (welcome to the M25) this sort of amount of energy is not going to make that much difference.
The place where regenerative braking does make a difference is in climbing hills. 1kWh takes a 1-ton car up 367 metres (about 1200 feet) if you drive slowly and the motors are 100% efficient at moving the car. If your journey consists of several hill-climbs then that will make a big difference.
A small problem comes if, like me, you live on the top of a hill. That means that if I fully recharge and then drive down the hill, halfway down the battery is full again and the regenerative braking gives out. That's bad. Any regenerative system needs a big dump load (and for 1C regeneration on my 180v 100Ah battery pack I'm thinking of a 0.56Ω 18kW resistor). That 18kW sounds like a lot, but it's equivalent to a vertical descent speed of 1.8m/s, which is about six feet a second or about 4mph vertically. Driving 12mph down a 1 in 3 slope would generate this much energy, and so would driving 55mph down a 1 in 9 slope - even including about 10kW loss for air resistance at this speed. That is normal driving conditions in hilly country, and the reason why brakes get hot.
My understanding is that most regenerative braking systems use big relays, which seems a bit strange to me. But my electronic development for this project is at the "scratch head and look at the magic smoke" stage: I know the principles, as a qualified and experienced electronic engineer, but I don't have final circuits because I haven't built them yet. It seems to me that a "buck" circuit from battery through inductor to motor should provide drive current, and a "boost" circuit in parallel from inductor through battery should provide regeneration. Most IGBT modules have protection diodes included, and if you figure the direction of the protection diodes, the diode for the "boost" IGBT is the freewheel diode for the "buck" circuit, and the diode for the "buck" IGBT is the freewheel for the "boost" circuit.
Here is a "back of envelope" circuit from my project website, to give you an idea:-
The load inductor is not shown: the Hall sensor is a current "clamp meter" type device, used to form a magnetic bridge to allow control of the motor current while maintaining isolation from the traction batteries. The motor current wire will pass through the ring once, but the control current will be between 25 and 100 turns, depending on the control current range I select. The optically isolated switches will probably be IGBTs. The upper-left one is the "buck" switch, and the lower right one is the "boost" switch. The two comparators detect when the magnetic bridge is out of step, then this is clocked onto the two switches by the pseudo-random sequence. If more than one of these modules is used (I'm planning to build them for +100A/-33A, and gang them together) then inductors will be needed on the output.
That's the theory - the practice will almost certainly need some ceremonial releasing of "magic smoke" before it's achieved.[ Parent ]
One idea I would like to share would be to use the vehicles brake system in the feedback loop for regen. One thought was to use the normal brake switch and a pressure sensor in the brake lines. The switch would trigger an A-D conversion of the pressure sensor and when the circuit noticed an increase in pressure, it would apply more regen, if the pressure drops, the regen would also drop. If the batteries are already full, then the regen circuit lays dormant until needed. The vehicles brakes still operate normally but during regen, they would seem to work even better. The goal of the regen circuit would be to keep the brake line pressure constant.
By the way, my work is in electronics engineering as well. Mostly sensor interfacing with microcontrollers and automotive networks. Not much in the high power realm, for the most part under 1000 Watts.
