AC line frequency stabilization

Scenario:

Power demand on a site varies between a minimum of about 200W, is typically around 500W and short term peaks (5 minutes at most) at about 10KW, the voltage is 240V and nominal frequency is 50Hz.

The site has three power sources, one from a power company, 4KW of PV from a grid tied inverter and a 6KW Diesel generator for back up.

Under normal operation, the PV assists the power arriving at the site by supplying some if not all of the demand, this is the good situation. Unfortunately, the incoming power is subject to frequent black outs and when that happens, the Diesel generator springs into action automatically to 'fill in the gap' until power is restored. However, the Diesel generator speed is not stable enough for the grid-tied inverters to maintain frequency sync so they automatically shut down, leaving a deficiency of 4KW in the supply. As a result, some machinery has to be shut down to keep the load within the generator's capacity.

A bigger generator is out of the question so a solution to stabilize it's output frequency is needed. The obvious answer is to convert the generator output to DC with a big rectifier, filter it and build a cystal controlled PWM sine wave output stage to convert it back to 50Hz again. Before looking into that possibility (reservoir caps on the DC being the biggest problem I foresee), can anyone see any other novel ways to achieve this?

Lateral thinking is needed - if it goes ahead the project will not commence for about 4 months so there is time to do a lot of planning.

Brian.

Hi,

To your question:
I'm sorry i know no such device.

But.....

I'm currently working on exactely a solution to this problem. I use a battery powered inverter (not the classic ones).
It acts like a big rotating mass, but completely electronic without moving parts.
It stabilizes frequency, voltage and automatically corrects powerfactor. It even improves sine waveform, if distorted.
It is able to source power from battery to load and in the next millisecond back to battery.
The power forward and back, as well as power factor is controlled, not regulated, therefore i expect it to be very stable.
All the design is very rugged and with low power loss.
The first solution is one phase, then three phases, 6kVA steady, 20kVA for some seconds..


The bidirectional powerstage is simulated and i have the necessary parts here.
But there is a lot of software needed, one part to use the system with the common power grid, one to use it offline, others to regulate the diesel, switch it off and on. And another big part will be all the safety mechanisms to get the permission to run it connected with the power grid. Automatically sychronize and connect if mains power is available and disconnect when power grid is shut down.

A lot of work.. and short of time to work on it...
I hope at end of the year a prototype runs with limited functionality...

I'm in germany, that means a huge ammount of bureaucracy before me...

Klaus

I understand the bureaucracy - if it is any consolation the same problem exists here as well. Too many overpaid EU 'consultants' in Brussels. :lol:

Your project sounds very interesting. I had considered solving mine in two stages, the first being a huge 24V DC 250 Amp PSU running from the generator, the second being a 24V to 240V pure sine inverter. Both are available on the market but the size and cost would be prohibitive and there would be considerable efficiency issues too. The simplest option is to connect a bridge rectifier across the generator output and use the 340V DC from it to directly power a PWM output stage but any ripple on the 340V would beat with the PWM frequency. It would be very difficult to keep high voltage reservoir capacitors topped up while drawing 25A from them.

Brian.

Hi,

yes, this is a solution.
What if the solar power is more than the load needs. Can you stop solar power or do you have to waste it?
It would be great to store it...

Is the solar power connected with the generator or is it connected at the output of the inverter?

Klaus

The bidirectional powerstage is simulated and i have the necessary parts here.
But there is a lot of software needed, one part to use the system with the common power grid, one to use it offline, others to regulate the diesel, switch it off and on. And another big part will be all the safety mechanisms to get the permission to run it connected with the power grid. Automatically sychronize and connect if mains power is available and disconnect when power grid is shut down.

A lot of work.. and short of time to work on it...
I hope at end of the year a prototype runs with limited functionality...

I'm in germany, that means a huge ammount of bureaucracy before me...

Click to expand...

I don't see a particular "bureaucracy" problem involved with the project. The inverter has to comply with basic safety regulations, power quality and EMI standards, as any other power electronic equipment.

Hi,

I don't see a particular "bureaucracy" problem involved with the project. The inverter has to comply with basic safety regulations, power quality and EMI standards, as any other power electronic equipment.

Click to expand...

Yes, a lot of work and time...

Klaus

What if the solar power is more than the load needs. Can you stop solar power or do you have to waste it?
It would be great to store it...

Is the solar power connected with the generator or is it connected at the output of the inverter?

Click to expand...

Sadly, the excess solar power is lost but that isn't a major concern, maintaining the 10KW capacity is more important. There is no DC to AC inverter except for the grid tied ones, there are two in parallel already in the system at the moment. All the PV output is connected to them, each having it's own solar panel array. That part works fine while stable AC from the power company is present, the problem is merging it with the unstable output of the generator which dips frequency down as low as 48Hz when the sudden load is placed on it. The shift in frequency is far more than the grid-tied inverters can adapt to so they automatically shut down and leave the generator coughing and spluttering black smoke until either someone switches the equipment off or the overload trip pops.

It's looking like the only viable solution is to use the generator to feed a hefty DC PSU and then use that DC to feed a 6KW pure sine inverter. Hopefully, the output of that inverter would be clean enough for the grid-tied ones to synchronize to. There is a hidden advantage in that method in that the DC (probably 24V) could be stored so at least temporarily it held the power up because at the moment, the generator has a 2 minute 'warm up' time before it's AC output is enabled. The batteries, probably 2 x 12V 100AH would act as a good DC filters as well.

Brian.

A economic implementation of the intended frequency converter would use a HV DC-bus. To utilize the generator's rated power, the AC/DC part must realize near 1 power factor. A bidirectional "active front end" as mentioned in post #2 would be the state-of-the-art solution.

I see another problem that hasn't been clearly addressed in your posts. To operate your local network in islanding mode during power failure, it has to be disconnected from the public power supply, both for safety reasons and to avoid shorting of your limited sources by external loads. How is the switchover accomplished?

May be starting current of some loads make the generator frequency to fluctuate. If it is so, then a method of soft starting for those loads can be the solution.

FvM: at the moment the excess PV power production leaks back to the grid but the internal anti-islanding circuits in the grid-tied inverters shut them down if the grid fails. It all meets G.83 and EU regulations and is regularly tested to comply. There is a monster sized three way switch which is manually operated to switch supplies when the grid fails, it is rated at 600V/100A and the center position is off so it is electrically impossible to connect the generator output or PV power to the grid when in 'local' mode. The PV power from the grid-tied inverters is fed in on the user side of the switch so it is also isolated from the grid in local mode. In theory it is completely safe. An opto-isolated line monitor is used to warn when the grid is back on again and the same signal is used to auto start the generator when it goes off so it warms up before anyone has time to reach the switch.

The idea of HV DC had occurred to me. The DC from the PV panels is typically 260V so it wouldn't be too difficult to mix it with DC from the generator (rectified 240V) then convert the combined power to a crystal controlled inverter to make 240V AC again. The drawback to that method is the existing grid-tied inverters become redundant and a new 10KW pure sine inverter is needed. That would work out expensive and I'm not sure I would feel happy to design and build such a big inverter myself. It's a one-off job so economy of scale would make it impractical to hand over to an expert company.

Ahsan_i_h: thanks for the suggestion. The load in this case is resistive (heating elements) but in any case the frequency stability has to be +/- 0.1Hz over a 3 minute period and the single cylinder Diesel generator has no chance of maintaining that. Even idling with no load it varies considerably more. It uses a mechanical speed governor so it's reaction time to engine variations is quite slow.

Brian.

I had no doubts about safe operation of your site, you didn't yet mention that the switchover involves shutdown and restart of the internal network.

I had expected that 6 kW generator and 10 kW total power are realized as a 400V three phase system. A three phase converter would use a 560 - 700 V DC bus.

I'll draw a diagram of the system when I get home tonight, it makes it easier to explain. There is a three phase + Neutral supply to the distribution board but only one phase is used for this part of the system and it is not possible to use the other two so it is essentially a single phase L-N-E supply. The 3 phases continue on to power a pair of 3HP motors but it doesn't matter short term if they lose power so they are not included in the backup supply requirement.

Brian.

It seems, frequency stability is only required by grid tie inverters.
If you can bypass inverters and supply with DC from solar panels and rectified DC from generator. Generator can also feed minor AC loads required.

Sketch of system attached. Unfortunately, the site is in operation and rewiring the supplies is not possible. The PV and generator are located in a different building some distance away from the load.


Looks like the only practical solution is to cut the generator output wires and add an AC-DC DC-AC conversion in line.
I should point out that everything is Earthed properly before someone points it out! I left it off to simplify the sketch.

Brian.

Back in the 1980s I saw a similar system realized with a DC motor coupled to an AC generator.
It fed a huge Burroughs mainframe computer.

The DC motor would have its speed controlled to maintain exactly 60 hz at the generator. A flywheel was also attached to the shaft.

Nowadays the rotating machinery would be substituted with solid state conversion, as you propose.