4 ms·

> The last thing which is important to know is how we can send the electricity to the grid. It is simple from a technical point of view [..] It's not simple fr

by elric 5y ago

> The last thing which is important to know is how we can send the electricity to the grid. It is simple from a technical point of view [..] It's not simple from a technical point of view. It's bewilderingly complex, in fact. If not done right, it can threaten the stability of the grid. It requires parts of the grid to have increased capacity. In/out balance needs to be maintained at all times. Decentralized production is great, but it could really benefit from more decentralized storage. Might be useful to start installing sizeable batteries near all medium->low voltage substations. Seems like the right place for it.

aerojoe23 5y ago
Also he didn't mention an auto disconnect if the power goes out, and as I understand it this is required in most places. You could be responsible for killing someone with out it. The power is out and the lines are being worked on. The line is supposed to be dead. Your solar system is making power and sends it out of the house, back on the grid.
- avianlyric 5y ago
  He did. He’s using a grid tie inverter which requires the grid to provide a synchronisation signal for the output AC waveform. No power, no grid signal, no output AC waveform.
  - jandrese 5y ago
    This is a standard feature of every grid tie inverter. Unfortunately it also means you can't power your house when the grid is out unless you also have a battery attached. Not that you would want to run appliances with the power dipping out randomly anyway. Maybe to charge a laptop/phone or run basic incandescent bulb/electric space heater/toaster type loads.
    - llampx 5y ago
      As the author mentioned, that can be done with a DC-DC converter.
  - marvin 5y ago
    Is this system guaranteed to not run into some sort of deadlock situation if your neighbor also has one of these installed?
    - avianlyric 5y ago
      Pretty much. Unless your neighbour has enough generating capacity to power the entire section of the grid your connected too (which would blow every fuse in the supply unit, and melt every cable between them and nearest transformer). Any power either you or your neighbour tired to put into the grid would be immediately consumed by all of your other neighbours (and of course your own appliances) that it would drag your AC waveform completely out of spec (I.e. it should be 230V AC, but your equipment can only supply enough juice to support 0.01V AC). So you’re inverter would shutdown due to lack of recognisable grid compliant waveform. Either that or catch fire from trying to drive thousands of amps (from 100 acre solar farm on your balcony) into your power socket, which would also catch fire, along with all the cable in the wall, well, you get the idea. In short physics and simple supply and demand ensure that no home scale inverter would ever be able to maintain an even vaguely compliant AC waveform during a blackout.
avianlyric 5y ago
Even balancing the grid is pretty simple from a technical perspective. You just watch the grid frequency, if it falls then add more power in (or disconnect load), if it rises add less power in (or connect more loads). No need to decentralise storage, decentralised production is very unlikely to ever be enough that distribution lines and unable to transport the power out. Additionally medium->low voltage substations are located near customers, which means the land is expensive. So putting in batteries with a material amount of storage is also expensive, assuming there’s even enough space.
- elric 5y ago
  I suspect you're not trying to be facetious, so I'll do my best to reply in kind. But seriously, just adding more power or just shed load is not easy. Where do you magically find more power in a decentralized system where something as fickle as a cloud can cause huge spikes in power generation? Load is easier to predict, but far harder to shed on short notice. Decentralized production absolutely can be enough to max out power lines. In most parts of the world, power grids aren't exactly crown jewels. Many grids are basically held together with proverbial duct tape. Here in Belgium the grid is such a pile of rubbish that instead of trying to adapt to decentralized production, we're now actively going to punish people for using power during peak hours. I agree that land and batteries are expensive. But they're a lot less expensive than rebuilding the grid.
  - avianlyric 5y ago
    > But seriously, just adding more power or just shed load is not easy. It’s all a question of scale, you need coordination to handle the macro level balancing of power. Making sure there’s enough power stations running to keep the lights on, and that the grid is roughly balanced. But that level of balancing is planned hours to days in advanced, because power stations take hours to days to spool up depending on the tech. So it’s a system incapable of responding to moment-by-moment changes. For handling instantaneous fluctuation, the first defence the the rotational inertia of all the generators in every power station. These things have tons of spinning mass, and are the reason grid frequency fluctuates with power draw, pull more power than provided, and the grid makes up for it by stealing stored spinning energy in the generators. Then speed governors on the generators notice the slow down and provide more steam/gas/coal whatever to speed the generator back up. Both covering the stolen inertial energy and bumping total grid input. For larger fluctuations that can’t be covered by rotational mass and modulating power station output, you’ll have a variety of fast response loads and supplies. Energy consumers that have agreements to modulate their usage or production in response to frequency fluctuations, and who get paid for providing the service. Some examples include steel and aluminium smelters. They can shutdown temporarily to shed load, and just rely on thermal mass to prevent their equips freezing. Water tower pumps, you’ll have plenty of stored water, pumps for canals, fridges and freezers can turn their compressors off, so can AC units, and even data centres that will swap to their backup generators and take their load off the grid. Now all of the above loads heading only works for a limited amount of time, all of those consumers can only reduce their load for so much time. So you also have power producers, people who have diesel generators sitting around that can be turned on and connected to the grid, rapid response power stations that can cold start in 15-20mins etc. All of these will be started by a drop in frequency, and expected to pickup the load without any central coordination. Between the fast response load shedding and fast response producers a grid can handle a substantial amount of production variation without any central coordination. In normal operations you’ll only be relying on fast response load shedding of non-critical loads like water pumps, and cascading to more drastic load shedding and power production if the grid can’t recover quickly by upping production at major producers (who won’t be running a 100% under normal conditions). > Decentralized production absolutely can be enough to max out power lines. > <snip> > Here in Belgium the grid is such a pile of rubbish that instead of trying to adapt to decentralized production, we're now actively going to punish people for using power during peak hours. Yes and no. Normally the behaviour you describe isn’t caused because the infrastructure can’t handle the production (after all peak distributed production is extremely unlikely to ever be higher than peak load, you simply can’t fit that many solar panels on a roof). But rather that distribution companies run on the assumption that power only passes in one direction in their network, from grid to consumer. All of their modelling, and more importantly, their financial agreements with the grid and each other depend on this predicate. As a result their objective is to ensure there is zero supply back to the grid, and that all distributed generation is always consumed within their grid, preferably within the same substation so there are zero backward flows anywhere. This doesn’t mean the equipment couldn’t support reverse flows, it’s just that figuring that out would cost money, and they have no incentive to figure it out. They get paid the same regardless of if they support reverse flows or not, so it’s better to obstruct and get government to pay the costs of checking, than it is to pay for it out of a profit margin. In short our grids are a lot more capable than the industry generally lets on. But using that capability would require profit making companies to invest in work that will never return a positive ROI (at least under the current model that distribution network operators are paid under). So it doesn’t happen, and network operators prefer to spread FUD, than actually innovate (which would risk their margins).

by elric 5y ago

aerojoe23 5y ago

Also he didn't mention an auto disconnect if the power goes out, and as I understand it this is required in most places. You could be responsible for killing someone with out it. The power is out and the lines are being worked on. The line is supposed to be dead. Your solar system is making power and sends it out of the house, back on the grid.

avianlyric 5y ago
He did. He’s using a grid tie inverter which requires the grid to provide a synchronisation signal for the output AC waveform. No power, no grid signal, no output AC waveform.
- jandrese 5y ago
  This is a standard feature of every grid tie inverter. Unfortunately it also means you can't power your house when the grid is out unless you also have a battery attached. Not that you would want to run appliances with the power dipping out randomly anyway. Maybe to charge a laptop/phone or run basic incandescent bulb/electric space heater/toaster type loads.
  - llampx 5y ago
    As the author mentioned, that can be done with a DC-DC converter.
- marvin 5y ago
  Is this system guaranteed to not run into some sort of deadlock situation if your neighbor also has one of these installed?
  - avianlyric 5y ago
    Pretty much. Unless your neighbour has enough generating capacity to power the entire section of the grid your connected too (which would blow every fuse in the supply unit, and melt every cable between them and nearest transformer). Any power either you or your neighbour tired to put into the grid would be immediately consumed by all of your other neighbours (and of course your own appliances) that it would drag your AC waveform completely out of spec (I.e. it should be 230V AC, but your equipment can only supply enough juice to support 0.01V AC). So you’re inverter would shutdown due to lack of recognisable grid compliant waveform. Either that or catch fire from trying to drive thousands of amps (from 100 acre solar farm on your balcony) into your power socket, which would also catch fire, along with all the cable in the wall, well, you get the idea. In short physics and simple supply and demand ensure that no home scale inverter would ever be able to maintain an even vaguely compliant AC waveform during a blackout.

avianlyric 5y ago

Even balancing the grid is pretty simple from a technical perspective. You just watch the grid frequency, if it falls then add more power in (or disconnect load), if it rises add less power in (or connect more loads). No need to decentralise storage, decentralised production is very unlikely to ever be enough that distribution lines and unable to transport the power out. Additionally medium->low voltage substations are located near customers, which means the land is expensive. So putting in batteries with a material amount of storage is also expensive, assuming there’s even enough space.

elric 5y ago
I suspect you're not trying to be facetious, so I'll do my best to reply in kind. But seriously, just adding more power or just shed load is not easy. Where do you magically find more power in a decentralized system where something as fickle as a cloud can cause huge spikes in power generation? Load is easier to predict, but far harder to shed on short notice. Decentralized production absolutely can be enough to max out power lines. In most parts of the world, power grids aren't exactly crown jewels. Many grids are basically held together with proverbial duct tape. Here in Belgium the grid is such a pile of rubbish that instead of trying to adapt to decentralized production, we're now actively going to punish people for using power during peak hours. I agree that land and batteries are expensive. But they're a lot less expensive than rebuilding the grid.
- avianlyric 5y ago
  > But seriously, just adding more power or just shed load is not easy. It’s all a question of scale, you need coordination to handle the macro level balancing of power. Making sure there’s enough power stations running to keep the lights on, and that the grid is roughly balanced. But that level of balancing is planned hours to days in advanced, because power stations take hours to days to spool up depending on the tech. So it’s a system incapable of responding to moment-by-moment changes. For handling instantaneous fluctuation, the first defence the the rotational inertia of all the generators in every power station. These things have tons of spinning mass, and are the reason grid frequency fluctuates with power draw, pull more power than provided, and the grid makes up for it by stealing stored spinning energy in the generators. Then speed governors on the generators notice the slow down and provide more steam/gas/coal whatever to speed the generator back up. Both covering the stolen inertial energy and bumping total grid input. For larger fluctuations that can’t be covered by rotational mass and modulating power station output, you’ll have a variety of fast response loads and supplies. Energy consumers that have agreements to modulate their usage or production in response to frequency fluctuations, and who get paid for providing the service. Some examples include steel and aluminium smelters. They can shutdown temporarily to shed load, and just rely on thermal mass to prevent their equips freezing. Water tower pumps, you’ll have plenty of stored water, pumps for canals, fridges and freezers can turn their compressors off, so can AC units, and even data centres that will swap to their backup generators and take their load off the grid. Now all of the above loads heading only works for a limited amount of time, all of those consumers can only reduce their load for so much time. So you also have power producers, people who have diesel generators sitting around that can be turned on and connected to the grid, rapid response power stations that can cold start in 15-20mins etc. All of these will be started by a drop in frequency, and expected to pickup the load without any central coordination. Between the fast response load shedding and fast response producers a grid can handle a substantial amount of production variation without any central coordination. In normal operations you’ll only be relying on fast response load shedding of non-critical loads like water pumps, and cascading to more drastic load shedding and power production if the grid can’t recover quickly by upping production at major producers (who won’t be running a 100% under normal conditions). > Decentralized production absolutely can be enough to max out power lines. > <snip> > Here in Belgium the grid is such a pile of rubbish that instead of trying to adapt to decentralized production, we're now actively going to punish people for using power during peak hours. Yes and no. Normally the behaviour you describe isn’t caused because the infrastructure can’t handle the production (after all peak distributed production is extremely unlikely to ever be higher than peak load, you simply can’t fit that many solar panels on a roof). But rather that distribution companies run on the assumption that power only passes in one direction in their network, from grid to consumer. All of their modelling, and more importantly, their financial agreements with the grid and each other depend on this predicate. As a result their objective is to ensure there is zero supply back to the grid, and that all distributed generation is always consumed within their grid, preferably within the same substation so there are zero backward flows anywhere. This doesn’t mean the equipment couldn’t support reverse flows, it’s just that figuring that out would cost money, and they have no incentive to figure it out. They get paid the same regardless of if they support reverse flows or not, so it’s better to obstruct and get government to pay the costs of checking, than it is to pay for it out of a profit margin. In short our grids are a lot more capable than the industry generally lets on. But using that capability would require profit making companies to invest in work that will never return a positive ROI (at least under the current model that distribution network operators are paid under). So it doesn’t happen, and network operators prefer to spread FUD, than actually innovate (which would risk their margins).