Allowable input to solar converter

[markfrencken]

Hi, I'm new to this forum so I'll introduce myself and the issue I'm struggling with...

In the Netherlands, electricity is going to hit 1 Eu/kWh likely in the next couple of months for most people when renewing their contract.
I'm a chemical technology engineer with quite some interests and experience in mechatronics and would like to investigate the use of a diesel generator to shave off the peak loads (and use the waste energy to heat my house) to enable using a 'day ahead'electricity contract. Day ahead means you know the hour by hour prices of electricity and when 0.4 l/kWh is cheaper compared to the net, I want to cap my net energy usage by feeding in with the generator. I do have solar panels, but these typically do not produce during peak price hours and batteries are not yet economically feasible, hence the approach.

Now the question: the diesel generator can deliver 1 phase 230V @ 50 Hz, with a capacity of 6 kW.
I do not want to run it at max load, so I would like to connect 2 2kW SMA inverters (with galvanic isolation) to the rectified output of the generator. I'll take the conversion losses etc at face value for the enormeous simplicity of this approach at the net side (if the net drops, I will  not have power I understand but I also do not need complicated and convoluted switchover solutions etc. I might be looking into that direction when I invest in a battery solution).

Rectifying the gen output is one thing, and due to the galvanic isolated SMA's I can choose to feed in on different phases of the 3 pahse system without worrying about earthed neutrals etc.

The SMA converters I want to use also give me the option to select either MPPT tracking or constant voltage for input. This of course also helps a lot.

What I'm not sure about and cannot find in specifications is the sensitivity of the (or any) inverter against ripple voltage. My proposed 'solution' will introduce quite some ripple @ 100Hz that I'm unsure about whether or not a converter can cope with.
I can filter the rectified output with for example a PI filter, but that could cause spikes during switching of the converter and cause all kind of issues as well.

How sensitive are (in general) these solar converters for ripple voltage, ifI were to use a MPPT tracker setting I can imagine it would / could severely affect performance of the control loop, using constant voltage I have no idea but expect it to be better?

What kiind of filter would be the best choice for this application? A PI filter, or something active even? How about voltage spikes when teh converter is switching off ?

[Circlotron]

and when 0.4 l/kWh is cheaper compared to the net, I want to cap my net energy usage by feeding in with the generator.

An alternative and simpler method that doesn't need inverters to couple to the AC mains is to use a normal induction motor as a generator. As a motor they resist being slowed below synchronous speed, and as a generator they resist being rotated above synchronous speed. Have your diesel engine coupled to this motor, bring it up to 1500 rpm (4 pole 50Hz) or 3000 rpm (2 pole 50Hz), close the big switch and dial up the power on the diesel while watching the AC amps doesn't go above rated motor amps. No synchronising necessary.

https://en.wikipedia.org/wiki/Induction_generator

[Faringdon]

I think you cant just rectify the output of the generator and then feed that to the inverters. Well, you can , but its unwise.
A better way, would be to connect a 4Kw offline SMPS to the gen output…and then connect the two inverters to the dc output of that.
You can either buy a 4kw smps….or buy several parallelable lower power smps’s and do like that, to get the 4kW.

Or you could just use a Boost PFC stage and run the two SMA's from the output of that.....which woudl be say 390Vdc or so...Is the gernerator output pure sine or quasi-sine?

Just rectifiying the gen output will  mean a very low power factor, (between gen and rectifier) and you wont be able to get much power out of it.

when you say you want to rectifiy the gen output.....I understand you mean, rectify and smooth...ie with capacitors following it......so just imagine, 1...the poor power factor...and 2....the hideously high peak currents going on in that rectifier.

By the way, if you can get a gov't grant or something, then ill come over and we can build a 4kW SMPS up for you......i have a quick, cheap way of building high power smps's........a quick way to get the wanted power at wanted voltage........not beautifully packaged etc, but do the job. You can maybe get a gov grant for this....specially now....no gas and all that.

[markfrencken]

Hi, thanks for the input.
Unfortunately, any government grants in the Netherlands are only given out to ready made product installers, not DIY, so the grant is currently already absorbed in the prices
Having said that, your remarks have set me to the path of using PFC circuitry of existing large power supplies like for example HP/delta ever psu's. On the other hand, passive massive solutions also have a positive ring to them for the robustness. and, moreover, using a massive inductor before the rectifier would (at least to some amount) mimic the impedance of actual solar panels whilst at the same time largely correct the peak currents of the rectifier/capacitor bank. Now on to looking for a 10 Amp 50 mH inductor, which might be already present in the genset since it was oriinally a floodlight system with metal halide lamps.

[Faringdon]

...You could just build yourself a 4kW Boost PFC and  supply the SMA's from that.
As you can see from attached LTspice, PFC circuits are just boost converter ccts with some control...not difficult....you can parallel them with eg ucc28070A......or do a load of cheap easy BCM PFCs in parallel with a good controller....BCM PFCs are very very easy.....You can buy controllers which can paralell them.

You can also get a very good power factor by having eg a constant off time controller with  eg L6564 in COT mode....very easy way to give you a 4kW 390V Bus from which to run your SMA's.....COT just follows the mains sine wave....so you dont even have the current error amplifier to worry about....its almost impossible to get that kind  of PFC wrong......PFC is not perfect like that, but is very very good...0.95 or so.

These Cosel PFCs can be paralleled, but you coudl build  PFC yourself as discussed.
https://en.cosel.co.jp/tool/tag/pdf/SFE_DPF.pdf

A passive PFC circuit for 4kW is going to be absolutely enormous......so much easier to hack a high fequency Boost  PFC up and go like that....just do a simple L6564 COT mode one first to get you started...then develop a CCM Boost PFC one...or several in pllel.

[ejeffrey]

Keep in mind that connecting a diesel generator to grid tied circuits is almost certainly illegal, even if configured for zero back feed.  That is true basically everywhere that has a functioning electrical grid.  In most places you can use a diesel generator as backup but you must have a transfer switch that disconnects the utility feed.  You can't run from both at the same time.

The ability to have grid tied solar or in some cases wind is a renewable energy exception, not a license to run your own fossil fuel plant.

On the other hand at EUR1/ kWhr it's hard to see how battery storage wouldn't be worth it.  Maybe the installer costs are so high that it's not worth it but if you are planning to do an unauthorized DIY, do it with batteries not diesel.

[Faringdon]

If you dont want to do grid tied you can always go "island".

Attached is the way to parallel 1kW PFCs.....up to as many as you like.......so give your 4kW

I hacked the sim so it runs quicker (LTspice)......so i fixed the error amp output of the "master"...otherwise it takes ages to run.......just connect up the slave like the master....and feed in the MOUT signal with an opamp...that will overrride the slaves MOUT signal and make it give same current as master.....in theiry, prang the slave so its Multiplier outputs its max current all the time....in theory that is a more bona fide way to connect up an opamp output to the multiplier pin....but it may not be too important.

Parallelling BCM PFCs is even easier than this.

PFC is basically just a Boost converter with a few extras.....you can hack it to work properly its that simple....or use the calcs if you want.

[Faringdon]

Also, attached is a "Jack" PFC for 4kW....these are really easy...just a BCM PFC controller hacked to work in COT mode........you can see how they can be paralleled.......i again hacked it to run quicker by  having the error amp output fixed....(LTspice sim)........you just copy across the CTRL voltage to the slave....the CTRL would be set by an external error amp that you can put in.

As you can see, at max power, the Power factor of the "Jack" PFC is actually 0.99.

Another way to pllel them woudl be to use an auxiliary resistor in the output divider...and draw current through it via an error amplifier which was looking at the max stage current, and the individual stage current.....(ie UC3907 style)....ie, doing this to make the  paralleled PFC stages  each ship equal current.

[Faringdon]

Attached is the "Jack" PFC in LTspice....this time with the vout regulation added, but unfortunately it takes ages to run...you can see  that as many as you want of these can be paralelled...easily up to 4kW.

[markfrencken]

Hi,
The generator is not going to be grid tied (that's why I have the unaltered hence safe solution with hte galvanic isolated SMA's in my mind).

Wrt batteries, yes, there will be those in the future, have to balance costs for now and the diesel generator came cheap, just likethe two SMA's. Only a fraction of the costs of a diy battery solution.

[markfrencken]

Againthanks for your input, I'm sourcing the parts for a 1kW pfc as a test on the generator.
Although I can only agree with you on the tremendous weight and copper involved in passive pfc (and I did some sims myself to come up with some numbers) I currently have 4 autotransformers that were used for the lights on the genset that also might do the trick as these have a huge inductance of about 300 mH at 100 Hz, 6 amps. Problem though is that the voltage drop would still be massive and the SMA's likely would be running on their teeth with huge currents. To be continued.

[Faringdon]

Another  problem with passive  PFC is that you can end up with maybe too low  DC bus voltage, as the big inductor filters down the mains sine.....the attached is LTspice and shows a 50mH inductor which doesnt filter it down too much , but only has a power factor of approx 0.7. (1000W real power)

[jonpaul]

Modify EV , use engine as gen on 380V input inverter

Priups.com

J

[Simon]

You need more than 6kW to run your house? stick to the grid or switch over to the generator. Don't take anything you are told as professional advice. Any custom solution may be dangerous or illegal.

[hve]

Also living in the small minded provinces of the Netherlands.

In 2004 the energy companies here in the Netherlands have been privatized in order to "benefit" the consumer.
Of course the opposite is true, the maintenance costs have increased significantly and the market is dominated by a few big players that regulate the prices for the consumer amongst themselves.
Yes you are able to choose between a few billing providers that lure you with sign on bonuses, and will hit you with big markups on the price if you are not carefull...

So interesting idea: diesel generator will that work?

Lets do the calculations:
Grabbing a random diesel generator: https://www.torros.nl/nl/aanbiedingen/super-aanbiedingen/diesel-generator-75kw-1x230v-3x400v-mw-tools-dg75e

• Full load output: 6kW
• Fuel consumption: 2.82 liters of diesel per hour at full load.
• One liter of diesel is 2 euro at the gas station

So 2 x 2.82 / 6 = 0.94 Eurocents per kWh

Apart from the noise and pollution I would say not an option
edit: And the generator has to be phase locked in before connected to grid or nasty things will happen I would imagine...

(Currently paying 0.35 euro per kWh, but that will change 1st of jan next year )