Controls Eng'r needs help w/ Power Question

[powerhigh]

I am working out a control solution for a load switching application, and I'm much more qualified as a 20 yrs+ controls eng'r than power, where I have minimal experience - hence, I need some assistance, please.

I'm working with a heating system with 3 pairs of independantly controlled resistive loads. These are on the secondary of a D-Y that is 480/80. There is a need to selectively current limit the primary to either 60 or 100A.
The six loads, two on each point of the wye, are controlled by a six zone on/off temperature/process controller.

I have used three single-phase phase angle controllers on the primary - using only the current limit function. These have fixed compensation for phase shift included. Two problems surfaced: first, with different combinations of the 6 loads turned on/off there were secondary V variations, both high & low. These on their own are acceptable to the process, albeit undesirable.

Second, and more significantly (I think) is what happens as the loads are nearly all, or all switched off. Then the load resistance change makes for a change in the primary load inductance and my SCR controller's phase shift compensation is off. This leaves a nice measurable amount of reactive power in the primary. Also, switching loads on and off when there are only one or two in use has blown 200A slo-blo supply fuses, although none of the controllers or SCR's have been damaged.

Perhaps I am going about this the wrong way.
Is there an accepted method to limit current with this D-Y, six load configuration?
I am considering controlling the 6 loads with contactors and SCR's for soft starts and stops while measuring the current in series w/ the primary lines. With some logic, I intend to prioritize the loads so as to control max current. Even if this works, how can I calculate the primary currents to expect for all 36 (6 x 6) load on/off combinations?

Thanks in advance.
Chris

 

[killabyte]

hi, i think the scr are switching the load along with the windings of the transformer. why to do that if you need just to control resistive loads, you should not switch the windings. just over the resistive loads (scr´s on the secondary), so you´ll avoid inductive reactance.

you say that youre having different voltage supplies, this would sound stupid, but be sure that the 3 phase system has a neutral wire.

about current control, im electronic so the answer i could offer is this way. if that could be...

hope to be any helpful.

killa

 

[rbulsara]

I am not sure what exactly you are trying to do, by having SCRs on the primary of a D/Y transformer and whether the current limting is intended to vary the heat output or it is a protection featuer something like overload protection?

Killabyte's response is in the right direction.

If you want individual element's heat output variation (by contolling the voltage) simple application would be to have six individual SCRs for each heating element, arranged in series with the heating element on the secondary side, before or after the on/off contact. (If the two elements on each phase act as a single unit, then 3 SCRs will suffice).

If you are under impression that by controlling each line current (or voltage) of the delta primary side, you are controling the secondary line currents, you are gravely mistaken.

Also if you post rating of your heating elements and the transformers it will help to understand if 200A fuse is a correct size.

If you are trying to keep the current rating of the SCRs by keeping it on the 480V side, you may be better of using (3) or (6) single phase transformers for each set or element.

 

[jbartos]

Suggestion: There are different principles of operation of controllers. The phase angle controllers are creating harmonics and spikes. The zero voltage switching principle produces less harmonics and smaller spikes, if any.
Visit

http://www.pelsla.com/septopic.htm

http://www.heatersplus.com/a1z.htm

etc. for more info

 

[powerhigh]

Firstly, thanks.
The SCR's WERE on the primary simply as a means of limiting current. The No.1 objective of this project is to prevent circuit breakers from tripping. To clarify, this is a portable unit used in places where it can take hours to get access to a breaker somewhere across the facility in order to reset it. With the SCR's on the primary, the controllers limit the current well via phase-angle, but load switching on the secondary (6 individual loads) blew fuses in my test setup. The 200A fuses just happened to be in the disconnect that feeds a transformer that supplies the test rig. The transformer in the UUT is only a 75 kva transformer.

Yes, the idea of limiting at the primary reduces the cost of the SCR's significantly, but the concept was to limit current quickly and accurately, as my controller allows. This, as I mentiopned, failed and I am trying to flush out any unknowns before moving on to the secondary side.
I am in fact wondering if I will even need the SCR's there since I am switching the loads by PID controllers w/ 30 sec. cycle time (also one per load).
My thought was that there would be benefit in using the SCR's to soft-start and ramp down the load cycles. This is one thing I am not certain of.

Also, is there a way I can figure out what will happen to the current on the primary as regards imbalance with different numbers of loads on & off???

 

[powerhigh]

As for proverbial forests and trees - I designed the SCR controllers and they can be configured to do phase angle or zero-cross switching. I never gave the ZC method a thought for this application - and will do so now.

Thanks, jbartos!

Perhaps this will give me some new ideas.....Hmmmm...

 

[busbar]

Won’t the distorted, thyristor/phase-controlled waveforms have a very undesirable effect on the transformer? The first effect would seem to be a very noisy transformer, right?

 

[rbulsara]

I am at a loss, what is there to limit the current to a resistive load? It will be a fixed current at a given voltage, and no inrush!!

You are trying to devise a solution to problem that does not exist! In case of a short circuit nothing will help. A CB rated 125% of rated current will just do fine.

 

[rbulsara]

On the contrary, you are creating unwanted inrush curret by switching and single phasing the transformer!!! Which wont be the case if switch R on the secondary side.

 

[60hzhmmm]

Powerhigh, Me thinks you may be doing this the hard way... I would put a zero-crossing SCR on each of the elements. Then I would switch the heck out of them. I regularly switch 50KW elements under zero crossing up to 50 times a second. This means your average power can be controlled with great detail as compared to your 30sec cycle time. If the elements can't take full power at start up then the power limiting can be done thru the same quiet cycle by cycle pulse width modulation limiting the AVG on time to some appropriate level. I usually resort to phase angle torture only when the required cycle time drops to a few seconds or less.

I've used this for the switch element:

http://www.ciicontrols.com/products/int100.htm

 

[powerhigh]

1. I haven't yet placed SCR's on the secondary, but yes, on the primary they sure make for a noisy transformer. I originally proposed controlling the load current on the secondary side, but was shot down by an EE who somehow got the configuration changed to the screwy SCR's on Primary scheme that basically has failed.

2. There is inrush on resistive loads, but that was not the problem. Since this is on a mobile platform and is used for field service work in places such as power gen plants & oil refineries one never knows how much power is available on the line the tech is told to plug in to. At times he can be given a 50 or 60 Amp line and he may have to use all 6 80V heaters simultaneously at something like 130A each. The ideal situation would be to add a layer of control that would limit the current from the source while letting the already inplace process control system continue to switch loads according to it's 6 PID channels. In use, you can't simply say "I only have 50 Amps available - so I'll only run 3 loads at a time...."

3. Now the ideal concept is to switch the 6 loads in a priority arrangement. The priority of a particular load would be based on it's deviation from set point temp. Any loads left intentionally lagging because of sensed over current at the primary would get their priority elevated for the next cycle and the highest temp load (least deviation) would then be turned off and allowed to lag. This would be done in a round-robin fashion. I admit this would be an odd solution, but I'm told this is the only acceptable method for the application. Either the people I'm working with don't understand proportional control or they just don't think it will work in this case. It is believed that the process requires loads to be full on in order for this process to function. Apparently splitting 50 amps across 6 loads at 1/2 power would be something like running them at 5kva ea, which they tell me would be useless because the entire process would never reach set point.

4. I like the idea of ZC switching the secondaries more & more. I'll have to think of a way to implement a control scheme for this method. I see many benefits - great suggestion, thanks.

5. Finally, if I draw a 6 x 6 matrix depicting all the possible combinations of on and off loads, how can I figure what the primary currents will (or should) be? I will need this in order to model the process and develop the control software. I am at a complete loss because of the voltage irregularities seen when the loads are imbalanced in various ways. Is observation and measurement the only way to find this out???

BTW, many thanks!!!

4.

 

[60hzhmmm]

Hi again Powerhigh, hmmm, Thanks for the whole picture, sounds like fun..

Point 1. A noisy transformer is an unhappy transformer.. Won't last. May go with more fan fair than expected.

Point 2. Why the expensive heavy transformer? Is there a reason the heaters can't be your line voltage? This would alleviate the switching element cost and increase the selection. 80V is way over the "safe" voltages anyway. By the way, is it assumed the supply will always be 480? The 'places' you mention will most likely have that.

Point 3.
I'm afraid, the "ideal" system is gonna be the expensive system... The only good way I can see of limiting the peak instantaneous power is to modify the voltage. A variable transformer...?? Wow. Or a tap switcher, problematic. If you could use line voltage heaters you could essentially check the available power then switch their wiring. Perhaps use twice as many elements and don't turn some on in low supply situations. This could let you select how many elements are used compared with how much power is available in much smaller jumps. It would also allow you to use very inexpensive control elements as they would be lower current since they are running smaller wattages. They could be down in the $30 bucks a unit range. With smaller current ones you essentially get the voltage rating for free. They're all like 600VAC these days. Now your controller could select any kind of patterns running the SSRs in parallel. You could also go with the cycling soccer ball kinda thing (who gets the power now?)

Point 4. Given.

Point 5. I believe,(somebody correct me), that if you switch on the secondary side you just do the math.. Winding ratio, current ratio, voltage ratio. Straight forward as opposed to the hideous primary switching you are effectively changing the transformer structure when rapidly rewiring it by disconnecting different legs while leaving the loads on it.
This device sounds like it will always be suffering from voltage loss in its power cord... 200 ft of SO cord today 20 feet tomorrow... With the "field of heaters" method you could put in a volage monitor that allows automatic heater selection. Remember SSR's are all isolated so by adding more you could also rewire the elements. Hooking them in series or parallel or both ect.

 

[powerhigh]

There are too many non-variables here. This is so because there are several units in the field to be retrofitted. Each one has a set of loads (actually flexible 'blanket type heaters) that the technician connects according to a set of instructions - get it now? The resulting solution must be useable by primates - who have already been conditioned to use it a certain way. Replacing the heater loads is way too costly, as would be swapping the transformer. My solution has to be a field retrofit also.

I just found that the center of the wye is connected to an earth ground (chassis) which I think is supposed to be a N.
I was assuming that it was wired as documented - with the wye floating. Once again, it's a different animal.

If the Wye was afloat, it's a simple phase to line / turns ratio thing, is it not? Now it's become a bit more complex....Ugh!

 

[killabyte]

well, let me tell you that switching on the primary is not a good practice, why to deal with inductive reactance having just resistive loads.

anyway use snubbers on scr`s this will limit peaks.

have a good look at the neutral winding if you switch on the primary side. neutral calble should be thicker than phase cables because the harmonics.

 

[jbartos]

Suggestion: If possible, consider the elimination of the transformer since there are 480V rated heaters on the market in many applications. Visit

http://www.chromaloxheaters.com/

http://www.mychromalox.com/literature/pg433.pdf

etc. for more info.
Please notice that it is not unusual to have heaters rated at 680V, for example. Click on the second link above.