Convert electrical frequency and change square to sin wave 1

[atheeanne]

I am a plasma physics innovator. I just completed the design of a power generator using plasma-MHD. The system outputs square waves at approximately 3000 Hz with powers output between 2.5 MW (500 Hz) to 20 MW (3 kHz). Does anybody know how I can convert this to useable utility power frequency (50/60 Hz)? I can do this by slowing down the system, but I am loosing too much power. Also, I am extracting power through multi pin electrodes at maximum 60 Amps per pin, so each pin can create its own circuit (at high frequency the pin current is under 1 amp). Is there a simple method to digitally change the shape from the square wave output to a sin wave to match the shape and frequency of the local utility power lines?

 

[jghrist]

My first thought is to rectify the current then use an inverter. With such a high frequency, however, you might explore modulation to a 50 or 60 Hz signal, maybe similar to AM radio modulation, then filter out the high frequency.

 

[itsmoked]

20 Mega Watts!?? 60A/pin? 100,000V??

Yes, rectify to DC then invert.

At 20MW you are at utility levels. Find the companies that convert DC to 3phase AC for power transmission. Your application should be easy.



Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

Google 'cycloconverter' or look it up in a power electronics text. They are horrible things which cause all manner of supply disturbances when connected to the utility network, but are great for converting a relatively high frequency source into a relatively low frequency output. Although most of the texts will show a sinusoidal input, the principle should be applicable to a square wave source. You will need some fairly heavy filtering after the converter.

Cycloconverters were used to drive very large variable speed low speed motors many years ago. They are more-or-less extinct since the advent of PWM drives (and all their subsets).


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[DaveScott]

Can cycloconverters operate with single phase input? I have only seen them with 3 phase input.
Another method for atheeanne would be to drive a motor and use this to turn a generator. I think maybe all other methods would require the use of harmonic filters.

 

[itsmoked]

Drive a 20MW motor with a square wave?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ScottyUK]

DaveScott,

A star for you. I don't normally tend to think of 20MW systems as single phase. Oops! I haven't seen a single phase cycloconverter - I'm not sure that such an animal even exists.


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[Snoball]

Is the frequency stable, clearly the voltage is not. Can you frequency regulate?

 

[jghrist]

The OP doesn't state what the voltage output is. With pin currents at 1-60A, there must be a pretty high voltage or one hell of a lot of pins to get 20MW. I don't understand the statement that maximum pin current is 60A, but at high frequency (where power is maximum), the pin current is under 1A. If at 3 kHz, power is 20MW and pin current is 1A, there would need to be 20,000 pins at 1 kV.

 

[jraef]

Cycloconverter in a 3kHz environment? That would require some extremely fast switching! And for that matter, running an MG set with a 3kHz motor?

I'm in the rectifier camp on this one. Too many other problems trying to convert something at that high of a frequency.



JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[atheeanne]

Thank you for your responses. I guess you deserve more information. The devise is a scalable MHD system. We create plasma at a rate of 6,000 pulses per second (or less) and deliver it through a discharge tube. At the other end of the tube we are able to recycle, reenergize and redirect the plasma stream back and forth along this linear plasma tube under a strong magnetic field where extraction through Hall Effects is at approximately 70% efficiency (demonstrated). Because the plasma is oscillating back and forth at 360,000 pulses per minute (6 kHz) for a long time, the length of the conductive thread (plasma) can be as long as 850,000 meters (~256,000 ft) per hour (~530 miles/hour) for a tube approximately 10 inch (254 mm) long. We can slow down the process at a rate of 120 Hz (7,200 pulses per minute) to obtain a square wave at 60 Hz, but we are loosing efficiency and power. The plasma tube is equipped with many pins along the way, approximately 10 inch, extended by the oscillation to hundreds of miles per hour. You do not have to have 20,000 pins to match the output current at 60 Amps, but only 100 or so. If you multiply 100 pins per 256,000 feet/hr oscillation, the virtual number of pins is extremly large. With a square wave at a single frequency of 3KH (it takes 2 turns to make a full wave) and a power output of approximately 14.5 MWe; 100 pins at 60 amps each will mean 6,000 amps total. The voltage can be calculated from there. This is just a rough example. 20 MWe or higher power are easily achievable by scaling the pumping system up or down.
Each pin constitutes an independent circuit, so even though we are speaking at overall utility power levels we need to change the shape of the pulse at much lower levels per each circuit of each pin. I guess, this will be easier to deal with. The generation and reenergizing of plasma in the plasma reactor is in the order of microseconds, so it is conceivable that higher frequencies are possible to nearly continuous wave (DC), but it would create more problems with cooling. For aerospace and ionic levitation, a new field in aerospace and space propulsion, the power generator is fine, since high voltages (40kV in average) and low current levels can be brnached from each pin. One pound of weight has been lifted (levitated) with 716 watts trough air ionization (demonstrated in 1964).
In order to have the plasma-MHD generator useable for power distribution (utility), however, we need a good sine wave at an acceptable frequency. Each pin circuit can be treated independently and than connected together to the grid. I need help from here; any comments?

 

[DaveScott]

atheeanne, I must admit that I am confused about the following:
There is a wave which bounces from one end of the tube to the other.
The wave makes an output of 60A flow on each pin, but not at the same time, since the wave is travelling across the pins.
So you cannot simply add all the currents on all the pins to get 6000A.
Maybe I misunderstood?

itsmoked, a square wave to a motor: yes, but a great deal of harmonic heating (and noise). It could be rectified and applied to a dc machine, but then isn't the voltage applied to the individual windings of a dc machine similar to a square wave? It seems to be a dc voltage switched by the commutator.

 

[itsmoked]

Hi Dave, Yeah, good point. Still I doubt that motor is very common and might result in an entire new development effort.

The rectifier method seems to make the most sense since it would allow complete flexibility of the MHD generator allowing throttling etc that would be very limited with any other methods so far mentioned.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[atheeanne]

Dave,
Good point, however, think of this: in my previous example 100 pins are set along a 10 inch plasma tube. The plasma is oscillating back and forth at 6 kHz. This means that a full square wave (positive and negative) is obtained at 3 kHz (a period of 333 microseconds); 167 microseconds for half of square wave. Since the pins are set up along the 10 inch plasma tube the current seen and extracted by each pin will be some 1.67 microseconds apart; 100 pins (the length of the tube and the plasma traveling in one direction) will extract 60 amps per each pin (not at the same time of course but very close together in the same direction of plasma travel) that yields a total of 6,000 amps along the tube 1.67 microseconds exposure per pin per half wave (60 amps X 100 pins). In reality, with the plasma moving back and forth exposed to a perpendicular magnetic field, one will see a square wave on the positive (above the axis) side 167 microseconds long (in one direction) and on the negative side (in the other direction). So one square wave above and below the line will have a total period of 333 microseconds, each formed by 100 segments above the axis and 100 segments bellow the axis 1.67 microseconds apart. With an output power of a generator capable of producing 14.5 MW, and the output extracted voltage of 7,200 Volts, the total current extracted should be about 2,000 amps (in one direction of plasma tube travel) or 20 amps per pin (1.67 microseconds apart) and 2,000 amps on the return of the plasma putting the wave below the axis, also at 2,000 amps per length of the plasma tube or 20 amps per pin 1.67 microseconds apart.. Are there any comments? I am not an expert, so I hope you can help me find a way to go to 50-60 Hz frequencies sin waves.

 

[atheeanne]

Itsmoked, thank you for your comments. If the current exposed to the pins along the tube where plasma oscillates back and forth why is it that the current seen by all pins along the tube (in one direction) cannot be additive? If rectifying each pin is the appropriate solution. Getting to DC current would require conversion to AC per each pin (?). Is that complicated to obtain or expensive?

 

[DaveScott]

See the following links for info: we are considering Siemens SM150 for our reversing roughing mill 2 x 8MW drives. TMEIC have some very good semiconductor systems, with high power and high voltage.
From memory, TMEIC is using a development of the transistor base (old = IGBT, new =IEGT), and Siemens is using a development of the thyristor base.

http://www.automation.siemens.com/ld/ac-umrichter/html_76/sinamics-sm150/sinamics-sm150-1.html

http://www.tmge.com/upload/press_room/050505_TMEIC-GE-Fact-Sheet.doc

http://www.micromech.co.uk/dir_prod..._overview.pdf#search="Siemens SM150 brochure"

(see page 17)

IEGT:

http://library.aist.org/ISSStore/PD...nElement#search="Toshiba mill drive inverter"

IGBT:

http://www.toshiba.com/taec/news/press_releases/2003/to-371.jsp

TMEIC TM-70 specification and application guide (up to 30MW at 3.3kV):

http://www.tmge.com/upload/library_docs/TMdrive-70-app-Guide.pdf#search="TM-70 IEGT"

That's all for now.

 

[jghrist]

Will each pin generate its full current for the full duration of the time that the plasma is going in one direction (167 microseconds)? Will the current be constant during that time? If both answers are yes, then it seems that there is no time gap between pins. They each generate 20 amps or 60 amps (which is it?) from the start to the end of the time of plasma travel.

Or will each pin generate only a short pulse while the plasma moves past, say a constant 20 amps for 1.67 microseconds, then no current until the plasma comes back the other direction? In this case, if you connect all the pins together, you will collect 100 pulses of current that are 1.67 microseconds apart, and you will get a 20 amp square wave at 3 kHz, not a 2000 amp square wave.

 

[itsmoked]

Are you running this as a Faraday generator or a Hall generator?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[atheeanne]

This is a Hall closed-loop generator, and yes, plasma is travelling from one end to the other, but each pin is collecting the current continuously over the entire travel from one end to the other until the plasma reverses travel. There are two plasma pumping chambers at each end of the plasma tube where the collecting pins are placed. Only 3/4 of the pumping chamber volume plasma travels to the tube, which gives full exposure to all the pins and therefore current extraction during one way travel (167 microsecond). Obviously, the pins set 1.67 microseconds apart from each other extract the current continuously in each direction. To answer jghrist's questions "each pin generates its full current for the full duration of the time that the plasma is going in one direction (167 microseconds). The current appears to be constant during that time. The plasma will gain its energy again from the opposite end pumping chamber and begin its return immediately. The pins are once again exposed in the same fashion upon the second cycle return and so on.".
Dave, thank you for your advise, I will look into the specified links for the TMEIC from Siemens.

 

[atheeanne]

Oh, one more thing, the 20 amps - 60 amps confusion is my fault. I am speaking of two different systems with different pumping cavities one that produces 20 MW at 3 kHz and the other 14.5 MW at 3 kHz but different voltages. Consider the system that produces 20 amps accross the tube travel in one direction. 2,000 amps cummulated over one full direction travel. Each pin is extracting 20 amps continuously during the 167 microseconds travel.