designing a charge controller for hybrid system

[diksha.k]

understanding about MPPT and how to build a charge controller.
Now question is about battery charging cycle:
Here as i have 12V battery say 110Ah then safe charging rate when battery discharged is 0.5 to 0.7 of charge rate/discharge rate.
I need to charge battery with mppt mode to put out maximum current into it provided the charging rate doesnot cross permissible level.
My question is:
what are the voltages i should refer as discharged,charged.
I am using voltage sensing at battery:
here i can assume that when voltage<=12V (not charged) MPPT mode
12V-14.4V constant voltage mode
14.4V and above termination.
Are these values correct?
What i have to refer to know the exact voltages i should sense to correctly charge the battery?

- - - Updated - - -

And it's specified to look for voltage per cell i.e:
2.2V/cell(discharged) to 4.2V/cell(charged)
And when starts charging in MPPT mode the voltage should reach around 60%(4.2)=2.22V or 70%(4.2)=2.94 then constant voltage mode till 90%of(4.2V)charged & later termination.
But how to find number of cells in battery to find out exact full charged or full discharged state?

 

[BradtheRad]

But how to find number of cells in battery to find out exact full charged or full discharged state?

For a 12V lead-acid battery, assume 6 cells.

The voltage at rest is 12.6-12.8V.

Discharged 10.5V

As a charging session ends, it is usually 13.8 to 14.4V (if you manage it properly).

However a fast charge may push it higher, maybe 15 or 16V. It's best to avoid this.

 

[Warpspeed]

Sophisticated mains powered battery chargers usually have several different modes of operation.
Terms such as bulk charge, absorb charge, and float operation are often used.

What we have is much more like a vehicle battery charging system.
The vehicle alternator is current limited by design, and is fitted with a voltage regulator set to usually about 14.2 volts. That is all there is, and it all works perfectly well.

We need something very similar to a vehicle battery charger system.

The solar panel can only output so much current, depending on size, so we do not need to current limit the battery charger.
Like a vehicle alternator, there will be a maximum current limited by the solar panel size, and there is no need to artificially limit the charging current, unlike a mains powered battery charger connected to an "unlimited" power source.

And just like a vehicle alternator, we need to set a voltage limit to the buck regulator output set to a similar voltage (14.2 to 14.4v recommended) to prevent overcharging.
Its all quite a simple requirement, and its all you really need.

 

[diksha.k]

thanks for your support tony,brathred,pradhan
Just a update:
I designed just a buck convereter first with L=200.2uH,C=470uf.With solar panel =19V-21V varying depending on sun intensity.
When i programmed the ucontroller i could see that the voltage across capacitor shooted up to solar o/p voltage=19V.
For whatever pwm it could always be around the same voltage.
Then i connected the load(battery) and set the PWM required to buck voltage to 12V and yes after connecting the battery it stepped down to the voltage required.
Now voltage across the battery is as per duty cycle and voltage is not shooting to max voltage.Hence its working as required.

 

[Warpspeed]

Great stuff diksha.k
Buck converters do not control very well with zero output load. As you have seen, the output capacitor charges right up to the input voltage and just sits there as there is nothing to discharge it.

Once the sun is up with a battery connected to it, the solar panel voltage should just sit steady at the MPPT voltage all day, until the battery reaches full charge.
Then the solar panel voltage should begin to gradually rise above the MPPT setting as the load current into the battery reduces.

 

[BradtheRad]

Absolutely, glad to hear you're making headway.

- - - Updated - - -

I'm still wondering about the developer's end of this. Are you the only student he has enlisted for this project? Can you be certain he won't choose someone else's design, and dismiss your efforts? Has he put anything in writing for you, in business terms?

Naturally there is every reason your design can be successful. Your electronic skills are a different thing from your business relationship.

 

[diksha.k]

Absolutely, glad to hear you're making headway.

thanks

- - - Updated - - -

I'm still wondering about the developer's end of this. Are you the only student he has enlisted for this project? Can you be certain he won't choose someone else's design, and dismiss your efforts? Has he put anything in writing for you, in business terms?

Naturally there is every reason your design can be successful. Your electronic skills are a different thing from your business relationship.

I dont care if he dismisses as my prime objective was to make this project a success /at least learn something out of it to get a good start in power electronics area as i am just a beginner.
Thank you.

 

[Warpspeed]

As a school project, the opinion of your teacher/tutor is far more important than the attitude of any outside developer.

This is going to be great fun, very valuable experience, and a very worthy project.

 

[diksha.k]

HI
Actually i just wanted to know that if i have designed buck converter on say 110W panel if i want to use it on 350W panel then i have to just change the duty cycle right?
As V=L*(di/dt)
with V=4V for 21V voc with 17V vmpp, L=250uH ,di=7% of(power rating/(Vmpp-V)).
Hence dt will depend on di and should only be varied if panel wattage change.

- - - Updated - - -

I also understand that: increasing the power would actually reduce the frequency and vice-versa.
But There may be a conflict if i go on reducing the power as the frequency demand for lower power rating may be much and uc has some max frequency limit.

 

[Warpspeed]

The input and output voltage ranges will stay exactly the same.
Hence the duty cycle and switching frequency will remain the same.

All that changes is the current will increase by very roughly three times.
That probably means larger mosfet and catch diode, and the choke will need to be rated for the higher current too.

You can reduce the inductance to about one third, but as before, the inductor just cannot have too much inductance, but too little inductance increases the ripple current.

I believe that you now use a 200uH choke probably rated about six amps.
Something like 60 to 100uH at twenty to twenty five amps might be about right.

Its not terribly critical, generally it is fine to work out minimum values for inductance and current, and choose whatever is commercially available in the next size up.

 

[diksha.k]

thanks,
Can you just tell me how much percentage variation do you see in the practical Vmpp and the one which is written or given by maunfacturer?
That is:
behind my panel its written Vmpp=17.89V But sir suggested that it's around 80%of vmpp not exactly vmpp.Is this true? Did your mppt tracking proved this point?Or was Vmpp exactly that what was specified?
I know you mentioned we can always check Vmpp using power meter in #2 post but just asking did you really found this difference?

 

[diksha.k]

As in this post you explained MPPT : https://www.edaboard.com/threads/348473/
So as i proposed to sir about tracking Vmpp rather than power but, he is opposed to this method as panel voltage may vary with 0.5V/cell as temp increase or decrease hence Vmpp may vary.

 

[BradtheRad]

panel voltage may vary with 0.5V/cell as temp increase or decrease hence Vmpp may vary.

I'm convinced the manufacturers test a new prototype on a mountaintop, clear day, sun overhead, clean connectors. Their data is in optimum conditions. Then they quote the results to us as the panel's specs.

We will never get equal performance. We're at ground level. We have haze in the atmosphere. Our panel gets dust on it. The sun is rarely overhead. Our wiring is barely adequate.

It's wise to drop our expectations 10 or 20 percent, and design accordingly.

 

[Warpspeed]

VMPP is a controversial thing to to discuss.
I know my own panels have written on them max power at 36v, and its nothing like that in practice. Its more like 29.5 to 30.0 volts year round.

Everyone tells me I am wrong about this, and you absolutely MUST use a complex software perturb and observe algorithm that corrects for temperature changes.

But practical testing with a power meter tells me that the peak power curve is not that peaky, and a volt either way only loses a very few percent of peak power.
What does change dramatically is the huge variation in ambient light from sun still below the horizon, to full total cloud cover, to clear blue summer sky.

My very simple system of always loading the panel down to a fixed panel voltage probably get an estimated 95%+ of what should theoretically be possible with something a lot more complicated.

Its just not worth chasing that extra few percent by adding cost and complexity.

Might as well just add one extra panel to the dozen you already have and get even more that way.
It may even be possible to do better with one simple MPPT buck regulator per panel, than hooking the whole lot up top a single more sophisticated MPPT controller, especially if the panels are of mixed sizes or origin as often happens.

It would be entirely possible to add a potentiometer to adjust the fixed MPPT voltage point up and down, and find the battery charging current peak manually.

I would like to bet that summer or winter the voltage will hardly change enough to worry about based on my own personal observations and measurements.

 

[diksha.k]

thanks,
At our college has 240W panel.
Hence the earlier(on 110W) buck converter setup that i tried with L=250uH gave me results as follows:
Earlier i had used a duty cycle of around 55.55% of time period due to which voltage from 20V stepped down to 12V.
But know along with 240W panel with same L value pwm=27.7%of time period. Due to this voltage from 20V stepped down to 14V and PWM is further going to reduce for 12V

 

[diksha.k]

And another thing is :
if i go on tracking Vmpp then, the algorithm has to change if we change the panel as Vmpp differs for various panels.So the algorithm what we write will not be universal it needs to be updated for different panels . But if max power is tracked then, there wont be such problem i suppose.

 

[BradtheRad]

There is the advice to 'fit the PV panel to the battery'. This applies to everything in your system. Commercial manufacturers specify a range of operation for a charge controller. Say, 15A for the inexpensive model. If you need 30A, they'll offer the next step up, at a higher price.

You can build a 300W unit and always operate it under 100W, but it is a waste of the larger, and more robust, and more expensive components which are needed to operate at 300W. Your customer will say 'No thanks, I'll buy a cheaper 100W model somewhere else.'

As it appears your 150AH battery is able to handle a 300W PV panel. 20A. You must minimize parasitic resistance to a few tenths of an ohm. Mosfets must turn on fully. Bulky large expensive inductor.

At some point you may consider multiple interleaved buck converters, each containing a smaller, cheaper inductor. You may want to do performance tests by doing bulk charging direct from the PV panel to the battery. It may be possible to enhance Ampere flow with a buck converter, but the how-to is not necessarily easy.

 

[Warpspeed]

And another thing is :
if i go on tracking Vmpp then, the algorithm has to change if we change the panel as Vmpp differs for various panels.So the algorithm what we write will not be universal it needs to be updated for different panels . But if max power is tracked then, there wont be such problem i suppose.

That is exactly right.

If you are manufacturing an MPPT controller, the user probably just expects to plug it in, and it will work perfectly without any tuning.

A controller that always loads the panel down to a fixed panel voltage requires the user to set the optimum voltage, and that is very easy to do, because under any solar conditions, the output power will always peak at the correct voltage setting.

It may only be a couple of watts, or hundreds of watts, but it always peaks at pretty close to the exact same voltage.

As the battery voltage stays almost constant at any particular time, a varying power input shows up as a varying charging current.

Once you have your buck regulator hooked up to a panel and battery, its pretty easy to see this effect for yourself.
You will need some way to change the voltage setting, perhaps a potentiometer connected to a spare a/d input so you can adjust the set point of the solar panel voltage up and down.

Don't be at all surprised if you can get more power out of your solar panel by loading it up more, where the panel voltage falls below the quoted 20v.
Try it and see !

 

[diksha.k]

thanks,
for your support brathred,tony..
As i was seeing the stepping down of the voltage by varying duty cycles then, when the duty cycle=2.77% then voltage change across battery was negligible it was as though panel was disconnected from load.
But when, i made duty cycle=99% then as per theory the voltage across load =solar panel voltage right?And hence solar panel voltage should be brought down to battery voltage but,this isn't happening..

 

[Warpspeed]

You should be able to load the solar panel right down to battery voltage at maximum full duty cycle.
At minimum duty cycle the solar panel voltage should be able to go right up to the full open circuit voltage.

When your mosfet is conducting almost all of the time, its almost a continuous dead short between solar panel and battery.