4 ms·

Just as with nuclear power reactors, I hope that we can develop standard transformer modules that would reduce the need for "one off" production, and utilize th

Just as with nuclear power reactors, I hope that we can develop standard transformer modules that would reduce the need for "one off" production, and utilize the economies of scale that modern manufacturing is capable of delivering. Smaller standard designs would make it easier to handle disruptions, make a power grid more resilient in the face of natural or man-made disaster. It would also make it easier to scale. Having a power grid with interchangeable parts would reduce the logistical complexity of spares management.

Gud 4y ago
There are already standardized transformer sizes(e. g. 40MVA). Electrical transmission is pretty standardized already IMO.
- mschuster91 4y ago
  The problem is that large scale transformers are expensive to produce which means that it makes no sense to keep stockpiles for a disaster case. Usually that calculation works out, as most electricity grids, even on the low-voltage distribution level, are meshed enough to tolerate one or sometimes two large-scale transformers going offline. You get into problems though if more of these transformers go offline - e.g. because someone shoots them with rifles or artillery - because the lead time is weeks to months. And completely forget about a nuclear EMP scenario... it would probably take a decade to recover from that.
  - WJW 4y ago
    Transformer protection equipment is quick enough to disconnect them in case of an EMP, it's usually the control electronics that are identified in vulnerability studies as the weakest link. Those have a much shorter lead time than the multi-ton copper parts and are also much easier to have a stockpile of.
pjc50 4y ago
They are standardized, they're just not "mass produced" due to difficulties in automation at that size. I suspect transformer production could be ramped up if the manufacturers were convinced that it wasn't just a temporary blip.
- bluGill 4y ago
  Automation isn't worth it unless you have a high demand. A skilled machinist can create many different parts, one day he makes parts for a transformer, then next he is building parts for a ship, and a week later he is doing custom parts for the rail roads. A skilled automation engineer would still need a few months to automate each of those separate parts, while the machinist is is done in a couple days with each. Once the automation is done the automated process can produce a lot more parts a lot faster, and cheaper - but if you only need one/a few the automation isn't worth it. Of course automation is getting better/cheaper. Things that were not wroth automating in 1960 are now so easy that you would automate it even if you only need one. Modern CAD/CAM systems can often go from drawing to automated production with the press of a button. Over time more and more things will be automated. Skilled machinists need to be trained. Sometimes you automate production just because you can get an engineer to design the automation faster than you can get access to skilled machinists to do the job. (this assumes the automation can be set up with less skilled labor - often automation itself requires more machinist time than one of the part it is making)
Kalium 4y ago
Standardization isn't really the issue described here. One of the issues is the manufacturing capacity for producing standardized parts is limited. The manufacturers are not interested in investing in the costly expansion of capacity required to meet a temporary surge in demand. Other issues are regulatory in nature. The article goes into some detail on the matter.
pyrale 4y ago
The issue is not that much about the ingress or about a single piece of equipment, as it is that adding new power sources (or consumers) creates topological changes in the network. That new sources changes the power flow not only locally, but potentially all over the grid. It also changes the way TSOs react to equipment failures: typically, when an equipment like a power line opens, the power flows differently in the network. In order for other equipments not to become overloaded, operators study the network situation and have go-to solutions that involve changing the network topology, rerouting power and, in extreme situations, cutting some consumers. If power sources within the network change rapidly, that study work becomes obsolete fast, and that becomes a problem for operators and their ability to maintain a secure transport grid. TSOs are working on it with both R&D and industrial applications, but that's not a fast process, and sometimes it challenges assumptions that used to be safe, and uncovers big question marks.
mikewarot 4y ago
The point I failed to make is because I brought a machinists understanding to an electrical engineering conversation. I assumed a specific definition of the word "standardized" that isn't common to electrical engineering. Here's what I should have said.... What's needed is standardized, interchangeable transformer modules. We need to be able to parallel as many of them as we need to manage a load, and if one dies, swap it out with a new module, possibly 30 years later, with NO need for adjustment. Just as precision makes it possible to deliver parts that fit within 100 microns from countries on the opposite sides of the globe (and has for almost a century thanks to Johansson[1]), we should be able to produce a standard, easily paralleled transformer module that can simply be bolted to others to match situational requirements. Bolting in a new module shouldn't result in the need for balancing or any other adjustment, the parts should just work together. This would have helped immensely in Ukraine, for example... as missile damaged sections of a transformer stack could be individually replaced, instead of the whole thing. There are enough CNC tools and measurement techniques in the world to make this possible, we just need to make it happen. [1] https://en.wikipedia.org/wiki/Gauge_block https://en.wikipedia.org/wiki/Gauge_block

Gud 4y ago

There are already standardized transformer sizes(e. g. 40MVA). Electrical transmission is pretty standardized already IMO.

mschuster91 4y ago
The problem is that large scale transformers are expensive to produce which means that it makes no sense to keep stockpiles for a disaster case. Usually that calculation works out, as most electricity grids, even on the low-voltage distribution level, are meshed enough to tolerate one or sometimes two large-scale transformers going offline. You get into problems though if more of these transformers go offline - e.g. because someone shoots them with rifles or artillery - because the lead time is weeks to months. And completely forget about a nuclear EMP scenario... it would probably take a decade to recover from that.
- WJW 4y ago
  Transformer protection equipment is quick enough to disconnect them in case of an EMP, it's usually the control electronics that are identified in vulnerability studies as the weakest link. Those have a much shorter lead time than the multi-ton copper parts and are also much easier to have a stockpile of.

pjc50 4y ago

They are standardized, they're just not "mass produced" due to difficulties in automation at that size. I suspect transformer production could be ramped up if the manufacturers were convinced that it wasn't just a temporary blip.

bluGill 4y ago
Automation isn't worth it unless you have a high demand. A skilled machinist can create many different parts, one day he makes parts for a transformer, then next he is building parts for a ship, and a week later he is doing custom parts for the rail roads. A skilled automation engineer would still need a few months to automate each of those separate parts, while the machinist is is done in a couple days with each. Once the automation is done the automated process can produce a lot more parts a lot faster, and cheaper - but if you only need one/a few the automation isn't worth it. Of course automation is getting better/cheaper. Things that were not wroth automating in 1960 are now so easy that you would automate it even if you only need one. Modern CAD/CAM systems can often go from drawing to automated production with the press of a button. Over time more and more things will be automated. Skilled machinists need to be trained. Sometimes you automate production just because you can get an engineer to design the automation faster than you can get access to skilled machinists to do the job. (this assumes the automation can be set up with less skilled labor - often automation itself requires more machinist time than one of the part it is making)

Kalium 4y ago

Standardization isn't really the issue described here. One of the issues is the manufacturing capacity for producing standardized parts is limited. The manufacturers are not interested in investing in the costly expansion of capacity required to meet a temporary surge in demand. Other issues are regulatory in nature. The article goes into some detail on the matter.

pyrale 4y ago

The issue is not that much about the ingress or about a single piece of equipment, as it is that adding new power sources (or consumers) creates topological changes in the network. That new sources changes the power flow not only locally, but potentially all over the grid. It also changes the way TSOs react to equipment failures: typically, when an equipment like a power line opens, the power flows differently in the network. In order for other equipments not to become overloaded, operators study the network situation and have go-to solutions that involve changing the network topology, rerouting power and, in extreme situations, cutting some consumers. If power sources within the network change rapidly, that study work becomes obsolete fast, and that becomes a problem for operators and their ability to maintain a secure transport grid. TSOs are working on it with both R&D and industrial applications, but that's not a fast process, and sometimes it challenges assumptions that used to be safe, and uncovers big question marks.

mikewarot 4y ago

The point I failed to make is because I brought a machinists understanding to an electrical engineering conversation. I assumed a specific definition of the word "standardized" that isn't common to electrical engineering. Here's what I should have said.... What's needed is standardized, interchangeable transformer modules. We need to be able to parallel as many of them as we need to manage a load, and if one dies, swap it out with a new module, possibly 30 years later, with NO need for adjustment. Just as precision makes it possible to deliver parts that fit within 100 microns from countries on the opposite sides of the globe (and has for almost a century thanks to Johansson[1]), we should be able to produce a standard, easily paralleled transformer module that can simply be bolted to others to match situational requirements. Bolting in a new module shouldn't result in the need for balancing or any other adjustment, the parts should just work together. This would have helped immensely in Ukraine, for example... as missile damaged sections of a transformer stack could be individually replaced, instead of the whole thing. There are enough CNC tools and measurement techniques in the world to make this possible, we just need to make it happen. [1] https://en.wikipedia.org/wiki/Gauge_block https://en.wikipedia.org/wiki/Gauge_block