|
TheFargate posted:Thanks! but I already know about that stuff lol. Basically we have a department dedicated to developing all the settings for relays based on coordination, circuit load and impedance. So honestly what Im doing is a little outside my job description lol. What I guess I am trying to figure out how to use the impedance magnitude and angle to actually calculate the boundaries of each zone of protection. For instance: what should my zone 1 minimum pickup value be? My maximum? What value is the boundary between zone 1 and 2? Etc etc etc. But thanks! The rule of thumb that I have heard and the one that I have used for setting impedance relays (also called distance relays) is to set Zone 1 to 80% of the line impedance and set Zone 2 to 125% of the line impedance. Zone 1 is set instantaneous and will protect faults on the majority of the line (set less than 100% to avoid unintentional overreach due to error in the line model, error in the CTs and VTs, and error in the relay calculations). Zone 2 is delayed (may use a DCB scheme for faster tripping with a slower backup on receipt of block) and protects faults on the entire line with (set to more than 100% to avoid unintentional underreach due to the same errors as Zone 1). The Zone 2 setting may need to be reduced in case the adjacent transmission line is a short line, as the 125% setting is meant to reach past the protected line and into the adjacent line, but you wouldn't want it to reach past the adjacent and into a third line. Also, the rule of thumb is for a two terminal line; three terminal lines and lines with tapped loads become much more complicated. Protective Relaying by Blackburn is a great resource for relaying. You should be able to take the Zone 1 and Zone 2 set points that you have, convert them from secondary Ohms to primary Ohms (using the CT and VT ratios) and then compare that to the impedance of the protected line (in actual Ohms, you may need to convert from per unit if that is how the model spits out the values). I'm guessing you will find that Zone 1 is about 80% and Zone 2 is about 125% unless you are in a special case (e.g. three terminal, tapped loads, or short adjacent line). Protective relaying is one of my favorites parts of being a power systems engineer, so I would be happy to elaborate more, if I didn't answer your question completely. I've lurked this thread for a while, but never got around to posting before. I spent five years as a consultant doing substation design for wind farms, solar farms, and data centers; one year blowing stuff up in one of the most powerful short-circuit labs in the US  ; and am now doing substation relaying and commissioning work for an oil utility. If anyone has questions in any of those areas, I would be happy to share.
|
#
¿
Jan 23, 2015 04:01
|
|
|
|
| # ¿ Apr 28, 2024 17:20 |
|
|
GWBBQ posted:Are there any big differences between designing substations for different types of generation? The PV plants I worked on (5 - 20 MW) were smaller than the wind farms (80 - 120 MW) so there were differences just due to the size. Of course, these are both tiny compared to traditional plants (e.g. coal, or hydro, or nuke), but I didn't work on any of those. The collector systems (the cables from the turbines/inverters to the substation MV bus) are significantly different. The PV panels and inverters are all next to eachother in a one large fenced-in area. However, the wind farms are spread out of large distances all over farmers' land. My 90 MW project had 50 turbines and had 47 miles of trench to bring that back to the substation. The turbines were consolidated into four feeders, so that is 47 miles of trench with the turbines daisy-chained, not a single feed from each turbine directly to the substation. A big difference from the relaying side, is that wind turbines are spinning machines, so in the event of a fault they can source fault current since they are converting some of their rotational kinetic energy into electrical energy. However, PV solar is inverter-based so they have extremely limited ability to source fault currents (basically they provide full load current even for a bolted fault). This affects the protection elements used on both the MV and HV sides of the substation to properly detect faults. Also, the wind farm industry is much more mature than the PV industry. It is very apparent when looking at the owners/developers, but also the utilities and ISOs are still working out the interconnect requirements for PV. I've seen quit a few wind farm substations and they all follow the same basic design; high-side breaker, main power transformer, low-side bus with outdoor feeder breakers and grounding transformers. But I've seen a lot of different PV substation designs, in part due to the smaller size, but also everyone just trying to push their lowest cost option (outdoor breakers, metal-clad gear, metal-enclosed gear, padmount gear, reclosers, some weired combination of these, etc.). In all honesty, most of the time in the lab wasn't blowing stuff up (or at least they weren't supposed to blow up). A lot of the testing we did were things like fault interrupting, short-circuit withstand, and load interrupting testing. Not only was seeing/hearing/feeling the test cool, but we also had a high-speed camera (2000 fps). It is amazing to see how flexible metal is when subjected to high current/forces and viewed at high speed. We also did a lot of power fuse testing. Mostly 15 kV fuses (sounds like a gun when it goes off), but a couple of 35 kV fuses (sounds like a canon when it goes off), and even a 69 kV fuse (not as loud as the 35 kV, but it was tested at reduced currents due to the current/voltage limitation of the generators). We also did some internal arc fault testing, which is ridiculous to see. We did some for padmount interrupter gear, which was neat, but the coolest ones was for a set of 15 and 25 kV custom metal enclosed indoor switchgear units. They were tested to the indoor arc fault standard, so backed into a corner with a false ceiling hanging over it and then flammable cloths all around the gear (if the cloth is burned, the unit fails the test). The test was 25 kA at ~15 kV (enough to sustain the arc) for 1 second. I've been involved in arc flash hazard calculations, had training on arc flash PPE, and seen videos, but I still wasn't prepared for how much energy is released in those arc fault tests. It furthered my respect for lineman and for arc flash safety. Below is a GIF that I quickly put together from a fuse successfully clearing fault current, I think it is 10 kA at 15 kV. I think it is really cool to see the fireball come out in slow motion. 
|
|
#
¿
Jan 24, 2015 23:43
|
|
|
TheFargate posted:This is awesome and helps a lot. Pretty sure I know exactly where I messed up now. While I multplied by ct/vt ratios to for current/voltage values I completely forgot to use my primary line impedance lol. So then my next question is this: through the use of my line impedance along with a Mho circle, how do I calculate actual test values so when I test a relay I can prove the limits of say, zone 1? I have found the equation Z = (Ex - Ey) / (Ix- Iy) but what do I need to do to actually solve and come up with voltage and current values? Thanks man! To test any given impedance, you'd need a combination of current and voltage (both magnitude and relative phase angle) that would give you that impedance. While any values with that fixed ratio would work, you'd be limited on the upper-end by the test set and relay capabilities and on the lower end by the relay sensitivity (I believe the SEL-311 series has a minimum current setting as well for the impedance elements, so you'd need to be above that). If you are trying to simulate actual system conditions, the values would need to come from a system model, since the current would be a function of the sources behind each relay. So, for a fault half way into the line, both line relays will calculate the same impedance, but the relay with a stronger source will see a higher current (and correspondingly higher voltage). I'm not sure what values automated test set programs use to test out the mho circles, but for any given value that you want to test, you should only need the combination of current and voltage. If your Zone 1 is set to 1 Ohm @ 70 deg, then if you apply 14.25 V @ 0 deg and 15 A @ -70 deg (impedance of 0.95 Ohm @ -70 deg) it should trip on Zone 1. If you apply 15.75 V @ 0 deg and 15 A @ -70 deg (impedance of 1.05 Ohm @ -70 deg) , it should NOT trip on Zone 1. The relay may require some pre-fault voltage before the fault current and voltage are applied (in part to drop out and undervoltage or loss of potential elements, but also because some relays will use a memory function from the pre-fault voltage waveform for when the fault voltage is very low). It's been a while since I've used a test set on a directional relay, but I think that is most of the steps. H110Hawk posted:Got any more of this sort of thing? Pretty awesome watching it go off over and over. I have a few more videos that I can sort through and make some GIFs from.
|
|
#
¿
Jan 28, 2015 04:40
|
|
|
Not as exciting as the fuse operating, but I still think it's pretty interesting. Short-circuit withstand test on a 15 kV single-phase hookstick-operable disconnect switch. The fault current is 25 kA symmetric, the shot was taken with the appropriate X/R and closing angle to give a full asymmetric offset of ~52 kA peak. The current comes in through the top of the switch, out the bottom, across some bus (not shown), and then back up through the bus on the right. The current in the switch and in the bus are spatially in parallel, so they will create a force on another, because the current is in the opposite direction in the two conductors, the force will cause them to repel. The bus is solid 2" aluminum that is heavily braced, so it doesn't even appear to move. The best article I could find after a short Google search for asymmetric current. In the picture near the top of the article, the example current has a "major loop" for the first loop (initial start or current to subsequent zero crossing), meaning that the current has it's highest peak value in that loop. The next loop is a "minor loop" meaning the peak is less then the symmetric peak. The current waveform in the picture more or less matches that in my GIF. Because the force is proportional to the current squared, the peak force in the major loop is significantly higher than in the minor loop. So in the video you see the switch move a lot and take a set (major loop), and then you see it move a little bit again (minor loop). The whole test last for one second, but it stopped it after the first cycle, since the largest force is in the first loop.  The switch feels very solid when operating, so I think it is neat to see it look so flexible at high speed when hit with high forces. It also means that when there is a fault on the system, every thing that carries the fault current (every conductor between the source generators and the fault) all experience forces due to the current (both between adjacent conductors and internal to the equipment). This means that every time a recloser operates on a permanent fault, not only is a banging the fault over and over, it is subjecting all the equipment to high forces.
|
|
#
¿
Jan 31, 2015 00:09
|
|
|
Pander posted:There are other installation codes I'll check out like IEEE 484. So far the gist I'm getting is "there's nothing specific saying not to install these near potential high-energy arc flash hazards but boy engineering judgment suggests it's a bad idea". I have seen drawings with batteries in a separate room, but everyone building I've designed (data centers, wind farms, and solar plants) or visited (utilities, universities, and wind farms by other designers) had the batteries in the same room as the relay panels and switchgear, where applicable (only a few sites had indoor switchgear). Usually a fan is set to run every hour or so to vent potential build up of hydrogen, and there may be a hydrogen detector to send an alarm to SCADA. Fake edit: I worked in a high-power short circuit lab where the battery system was in it's own room. That's the only time that I have personally seen it.
|
|
#
¿
Oct 24, 2017 22:16
|
|