Current or Power Reference Tracking with Power Electronics Converters

[og_ogun_srvs]

TL;DR Summary

Current or Power Reference Tracking with Power Electronics Converters

Hi. This is my first post. Thanks to all who give it a look.

I am working on an application that has very stringent power output requirements. There is a rechargeable battery that I am trying to manage through a bi-directional boost/buck converter as is typical of many modern-day electric vehicle control topologies. However, I am finding that due to electrical coupling, trying to control voltage alone to achieve desired power output is giving undesirable results. The truth is that I am interested in tracking power but I have not found anything in the literature that talks about power tracking through DC/DC converters.

Would appreciate any thoughts about the practicality of tracking voltage+current i.e. power, in DC/DC converters.

 

[Windadct]

Not really sure what you mean by " electrical coupling"

On the outset - if Power is essentially Voltage * Current, how can you set a desired power, by only controlling voltage?

 

[anorlunda]

When you have a load attached to a supply, the load determines the relationship of voltage to current. For example, a simple resistor with resistance R, establishes the ratio of voltage to current R=V/I according to Ohm's Law. The power to a resistor is therefore $P=\frac{V^2}{R}$.

Batteries are much more complex than a resistor, but still it is the battery determines the relationship.

Battery charging protocols can be elaborate, and can include time and temperature as well as voltage and current.

 

[og_ogun_srvs]

I am currently in the research phase. Sorry if I ask anything that sounds remedial but I have 2 aerospace engineering degrees and no electric engineering degrees so I am taking some time to get my bearings on this subject matter. The information provided so far has given me something to think about. Thanks.

 

[og_ogun_srvs]

When you have a load attached to a supply, the load determines the relationship of voltage to current. For example, a simple resistor with resistance R, establishes the ratio of voltage to current R=V/I according to Ohm's Law. The power to a resistor is therefore $P=\frac{V^2}{R}$.

Batteries are much more complex than a resistor, but still it is the battery determines the relationship.

Battery charging protocols can be elaborate, and can include time and temperature as well as voltage and current.

I understand what you are saying here. For a battery, there is probably a way to compute equivalent resistance, or impedance I imagine. It sounds like researching more about such concepts might lead me down the road I am interest (or contractually obligated) to go down.

 

[berkeman]

Welcome to the PF.

Yes, recharging a battery can be pretty complex. What battery chemistries are you considering? What capacity? How big is the EV? Will it be a Hybrid, or strictly electrical?

 

[anorlunda]

Here is a figure showing a typical charging protocol for lithium batteries. The figure comes from https://batteryuniversity.com/learn/article/charging_lithium_ion_batteries Note that voltage, current and power all vary with time.

 

[og_ogun_srvs]

Welcome to the PF.

Yes, recharging a battery can be pretty complex. What battery chemistries are you considering? What capacity? How big is the EV? Will it be a Hybrid, or strictly electrical?

Good questions. This is likely a Li-Ion chemistry. Capacity is an open variable. We will likely collaborate with a battery company design custom batteries for the final product. The power scale is on the order of megawatts and will be coupled to an IC engine so it will be a true hybrid.

This sounds like it may break down to inner-loop and outer-loop control often seen in aircraft flight control systems where a flight path trajectory is maintained in hierarchal control through individual thrust and control surface controllers. Here, we design a controller with a voltage inner loop and a power outer loop to achieve the desired "power trajectory".

 

[og_ogun_srvs]

Here is a figure showing a typical charging protocol for lithium batteries. The figure comes from https://batteryuniversity.com/learn/article/charging_lithium_ion_batteries Note that voltage, current and power all vary with time.

View attachment 245279

Looking at this, it seems that the current is passively controlled by whatever the voltage is over time, given a certain state of initial charge? And of course the voltage of the charging circuit would be controlled by typical boost or buck converter switching logic?

 

[anorlunda]

Looking at this, it seems that the current is passively controlled by whatever the voltage is over time, given a certain state of initial charge? And of course the voltage of the charging circuit would be controlled by typical boost or buck converter switching logic?

No. The charger must cover the entire range of 0% to 100% battery charge state. As you see on that curve, there are three regions. It is not as simple as an equivalent resistance.

I think most EV chargers are microprocessor controlled. They may use temperature feedback. They may use testing protocols, or vehicle history, to determine the state-of-charge when first connected to the charger. The intelligence for control of charging might reside in the vehicle, or in the charging station.

Is this a school project?

 

[essenmein]

Looking at this, it seems that the current is passively controlled by whatever the voltage is over time, given a certain state of initial charge? And of course the voltage of the charging circuit would be controlled by typical boost or buck converter switching logic?

Not quite, most battery charge cycles fall into two basic stages, constant current bulk charge, and voltage controlled finishing charge. Note current doesn't have to be constant, but its more that while below some SOC, you can pile the current into a battery without worrying too much about the cell voltage.

During constant current charging, the cell voltage is monitored to determine the end point, and eg in Li ion, a knee in the voltage is detected to determine when to switch to voltage control mode. The voltage points at which end of charge etc are indicated are temperature dependent, so bat temp should be incl during the determination of those set points.

I don't see any reason you couldn't have "power" as the control quantity, it would off course be bounded by the minimum/maximum voltage the converter can deliver, and there will be a ultimate current limit, but if within those boundaries, controlling for power should be no problem.

 

[og_ogun_srvs]

No. The charger must cover the entire range of 0% to 100% battery charge state. As you see on that curve, there are three regions. It is not as simple as an equivalent resistance.

I think most EV chargers are microprocessor controlled. They may use temperature feedback. They may use testing protocols, or vehicle history, to determine the state-of-charge when first connected to the charger. The intelligence for control of charging might reside in the vehicle, or in the charging station.

Is this a school project?

More or less. This is a post graduate university research project.

 

[berkeman]

Sorry if I ask anything that sounds remedial but I have 2 aerospace engineering degrees and no electric engineering degrees so I am taking some time to get my bearings on this subject matter.

The power scale is on the order of megawatts and will be coupled to an IC engine so it will be a true hybrid.

This is a post graduate university research project.

That is a very ambitions project! How many team members are there on the project? How many have strong experience with battery chemistry and with power electronics? How about real time embedded system software design?

Hopefully you can leverage off the published work of other Hybrid vehicle design efforts, but probably a lot of those would be proprietary design efforts (like at Toyota and other Hybrid vehicle manufacturing companies). IMO you won't be able to get very far on such a project without several very experienced team members...

 

[og_ogun_srvs]

That is a very ambitions project! How many team members are there on the project? How many have strong experience with battery chemistry and with power electronics? How about real time embedded system software design?

Hopefully you can leverage off the published work of other Hybrid vehicle design efforts, but probably a lot of those would be proprietary design efforts (like at Toyota and other Hybrid vehicle manufacturing companies). IMO you won't be able to get very far on such a project without several very experienced team members...

At the end of the day, these are all controls problems, just like aerospace ,chemical, biomechanical, etc engineering controls problems. It just take a bit of leg work to get unfamiliar systems into the right system dynamics models so that one can implement generalized control theories (PID, sliding mode, model predictive, etc...)

 

[essenmein]

What benefit does a hybrid system give you in a plane?

I can see it for stop and go (land) vehicles, makes less sense on long haul high way driving, some kind of sense in a boat (same reason its done on trains), but a plane?

 

[Tom.G]

What benefit does a hybrid system give you in a plane?

If I recall correctly, take-off is the big energy hog (with some usage during landing); that's when the battery is drawn down to gain altitude. The battery is recharged during cruise. The approach avoids running large IC engines at low throttle during the cruise portion of flight.

Among others, Airbus is working on them.
https://www.airbus.com/innovation.h...gle__ao-sem|innovation-electrification|em__TF
As is a startup:
https://spectrum.ieee.org/aerospace...c-airplane-aims-to-rejuvenate-regional-travel

 

[essenmein]

I guess I'm highly skeptical, its a field I work in on the land vehicle side (power electronics for HEV drive trains) so have reasonable understanding on a system level of how well these work (eg the generate/motor cycles for a vehicle driving WLTC).

I can see why it might be attractive, in theory, but somehow doubt once system inefficiencies and weight is considered that it really buys you much. Hard to say though, all it would take is a battery break through, basically that is the hold up for EV in general, the motors and inverters are basically there, its the energy density of the batteries that is the problem.

Interesting link to another article in the comments of that Zunum one.

https://leehamnews.com/2019/06/14/30405/#more-30405

 

[anorlunda]

This old thread https://www.physicsforums.com/threads/extremely-high-current-generators.806573/#post-5064587 is a related hybrid idea. It asked about a special case. Capturing energy during landing to be used to help the plane taxi to the gate. It is interesting because that is a special case. Jet engines are horribly inefficient at very low powers, such as during taxi. So the ability to shut them down entirely during taxi has appeal. If the same system could run in reverse to use stored energy and motors to assist during the departure taxi and takeoff roll, that adds to the appeal.

In that thread, we were skeptical about increasing the mass of the airplane for systems used only on the ground.

I mention that in this thread only to point out that the value of hybrid ideas is not limited to use in-flight.

 

[essenmein]

It is interesting because that is a special case. Jet engines are horribly inefficient at very low powers, such as during taxi. So the ability to shut them down entirely during taxi has appeal. If the same system could run in reverse to use stored energy and motors to assist during the departure taxi and takeoff roll, that adds to the appeal.

In that thread, we were skeptical about increasing the mass of the airplane for systems used only on the ground.

I mention that in this thread only to point out that the value of hybrid ideas is not limited to use in-flight.

IMO the taxi thing makes sense, I don't know what portion of fuel burn is for taxi vs flight (probably only v small percentage), but its a larger portion for short haul flights. However you don't need any fancy hybrid stuff or batteries to do this. Hub motors in the landing gear and run them off the apu, shut down the main engines.

 

[anorlunda]

At the end of the day, these are all controls problems, just like aerospace ,chemical, biomechanical, etc engineering controls problems. It just take a bit of leg work to get unfamiliar systems into the right system dynamics models so that one can implement generalized control theories (PID, sliding mode, model predictive, etc...)

The man with the hammer thinks everything looks like a nail.

A control engineer not only needs a bag of control methods and tricks, he must thoroughly understand the physics of the process being controlled. Perhaps the purpose of this project is to force control theory students that lesson.