Connecting Solar to the Grid is Harder Than You Think

📖Signed copies of my book, Engineering in Plain Sight, are back on sale! store.practical.engineering/ 🌌Get Nebula using my link for 40% off an annual subscription: go.nebula.tv/Practical-Engineering

In early 2025, the Baltic states in North-Eastern Europe are set to disconnect their grid from the Russian power grid, and connect up with European power grid instead. This might sound like it's simply some sort of resynchronization task, however the reality is much more complicated, requiring the Baltic states to build out a fair bit of extra infrastructure and take on the task of maintaining grid frequency where before, Russia would have been doing so. This might make for an interesting video.

I work as a lineman in northern EU, and an interesting effect we've noticed with solar is that sometimes the inverters won't disconnect from the grid during a power outage. We've ran into this when doing service work on pad-mount transformers, when we open the switch to de-energize the transformer supplying a neighborhood (in our case an average of 150 households/transformer) with a lot of solar, on rare occasions it will remain energized, backfed from the households solar production. This happens because in the right conditions the solar panels will produce the same amount of power as the households consume, so there's no current flow through the transformer. So, when we open the switch - as far as the inverters are concerned nothing changes, so they remain on and backfeed the grid. It usually doesn't last very long since when the balance shifts they'll disconnect but you could still get a nasty surprise if you're not careful.

I worked at a factory that ran 24/7 and consumed megawatts of power. We were looking for backup power solutions. The problem was the few seconds backup generators take to start. There were "batteries" that consisted of carbon fiber flywheels in a vacuum spinning at 500,000 RPM. They could deliver 100% output (multiple megawatts) within 10 cycles, and for up to 10 seconds - enough time to start a generator kept in warm stand-by. Perhaps technology like this is what we need to stabilize the grid.

"Everything is a smoke machine if you operate it wrong enough."

I'm an engineer for a utility-scale solar developer and this is a fantastic summary of the challenges we face and design for.

the final project for one of my classes was to design a solar installation for a factory, we had to calculate power usage from power bills, design a capacitor bank for power factor correction, find out how many panels, in what arrangement, and at what angle and distance from each other they had to be installed, pick an MPPT and inverter, and design a battery bank. i couldn't finish in time, it was a nightmare, but at least it taught me that people way underestimate how much work goes into installing solar

Connecting it to the grid is easy. Making it not explode when it does is the problem.

I’m so happy the idea that “magic smoke” is what makes electronics work, is so ubiquitous.

The Odessa event was (edit: contributed to) by a particular phenomenon called Cessation, which is not an unexpected behavior of nebulous algorithms. Designers put it in intentionally to protect the hardware without violating the grid rules at the time, which say when a disturbance happens, you must keep your output breaker closed to help the system (low voltage ride-through.) But there were no requirements that your unit continue to output (real) power. Solar units therefore would drop their power output to nearly 0 to protect the hardware from harmonics generated by the panels trying to push power into a disturbed grid. There was also an automatic timeout on this cessation where the panel would not ramp up until the disturbance had passed for some time. This was often baked into the hardware and many farm owners didn't even know the behavior existed. 6:30 this explanation would have benefitted a lot from showing a raw PWM signal, then showing a capacitor low-pass filter wipe across it to smooth it out

I am a test engineer for high power frequency converters and be build our "VSD"'s not only for grid injectors like wind turbines (exactly like in the video), but also for applications where you would not suspect them. Is is for example sometimes easier to let a water turbine run at a non-constant speed and run all that power through a VSD, which constantly synchronizes to the grid, than try to maintain a constant speed with the turbine. The are also used as a soft starter for these very big motors and sometimes, they can even reverse the flow of energy to pump water back up into a lake for example. We are talking from 500kW to 100MW here. Great video!

This was interesting. I was a nuclear trained submarine officer and served on an improved LA class sub for 3 years. Going through the nuclear training pipeline, you learn a lot of this stuff. In fact, the crude demonstration in the video of a drill turning an AC generator is something we had two: the "ships service motor generator" (SSMG -- everything needs an abbreviation), which were approximately the size of a midsize vehicle. They converted power back and forth from the AC to DC bus, where the DC bus was powered by *VERY LARGE* lead-acid batteries in the bottom of the sub. Very interesting stuff. The control room ("Maneuvering") had the panel for managing the electrical buses, and the operators manually would synchronize various bus frequencies ("slow, in the fast direction" for the incoming bus, and throw the switch for the breaker at approximately the 10/11 o'clock position to shut the breaker at 12). Depended on the operator -- not automatic. I believe newer subs use a solid state inverter system and not the SSMGs now. Anyone really interested in hands on, great education in this sort of stuff, either as enlisted or officer, join the Navy and go to Nuclear Power School and get to a ship. I wasn't an engineer in college, but you more or less become one working the engine room of a nuclear submarine.

power electronics engineer here: Good video. I do have a few additional thoughts, though: 1. You said a home PV system cannot provide power during a blackout. At the end of the video you talk about grid forming inverters being able to support e.g. a backstart or island grids. The good news is: Those systems do exist for home PV applications. They are usually more expensive, but can supply power during a blackout. (There are different types being able to do different things exactly - I'm not going into details.) 2. One problem not addressed in the video (or maybe I missed it?) is the fact that mechanical generators can provide high short circuit currents that will trip breakers (not the ones in your house, the large ones being part of the power grid). Since fault currents have always been high in the past (something like 10x the maximum normal current), circuit breakers were designed to not only handle them, but also require them for fault detection. But: Inverters don't deliver such high surge currents, even if they stay active during the fault event. They generally limit their own output current. While a mechanical generator can survive a current surge into the grid, a - for example - PV inverter couldn't (hence the current limiting). This is different from fault ride through (FRT) and is an additional problem. IMHO the solution to this is more modern protective equipment that will detect faults differently and trip not only during high current surges. 3. What you're showing at 6:06 is NOT pulse width modulation (PWM)! This is (kind of) what a multi-level inverter's output voltage could look like. A PWM still only has two voltages: plus and minus dc voltage (in the case of a two level inverter, that is), but switches between them rapidly, so that the resulting voltage resembles a sine-wave after filtering.

I learned all about MPPT while shopping for solar charge controllers. No tutorial I found actually showed the same panel in different conditions, they just said that the MPP changes. Your test results were nice to see, and helped with understanding. Thanks!

Power grid automation and protection are the industry I work in (R&D at Schweitzer Engineering Laboratories). This is one of the really pressing and interesting challenges which modern grid operators are dealing with. Wind and solar both add quite a few challenges when it comes to grid stability, but the economics are such that they are going to continue to become a larger piece of the puzzle moving forward, so we have to continue learning how to protect and regulate our evolving power grid. This was a very nicely produced video which strikes a good balance of talking about the challenges and reasons why power production still is going in this direction.

This helps me to appreciate what is going on inside my off grid solar system controller. Cooling fans spooling up and down in proportion to the amount of sun juice coming from the panels, amp meters on battery management units counting out the juice coming and going to keep me connected, comfortable, and conserving in an independent way. Super technology and we are gleaning the results of a massive effort. Thanks for being the one who explains the mega-systems under construction today.

Fun fact about the frequency. Microwave ovens usually use grid frequency to keep clock in time. That's why microwave ovens' clock never stays on time, at least very long. Where I live the clock ticks ever so slightly too fast, because our grid provides slightly over 50Hz power, but MWO still ticks on one second every 50 oscillations.

It sure is difficult. But engineering is transformative—no matter what field or area you're in.

A dive into grid forming inverters would be an interesting future video.

I work in power generation ( a good portion of my work is in the renewable sector). Solar farms present a unique challenge in the fact that there is no inertia behind the power production. A regular power plant ( gas or steam turbine) has an effect almost of a synchronous condenser ( a small power reserve). Inverter based generation doesn’t provide that. Which can be weird once you start getting into crazy demand days. Most inverters used on the grid are what’s called voltage source inverters, meaning that the dc voltage MUST always be higher than the inverter AC output voltage. This is to ensure the inverter can “pump” power out to the grid. From a dispatch standpoint , most solar farms are not dispatched, they produce what they produce to the grid. The ac voltage that gets exported to the grid is a relatively stable voltage regardless of the dc production, although the power flow will adjust with dc production. There is very little tolerances when it comes to grid tied systems. Most solar inverters and battery storage inverters ( again utility scale) are the exact same equipment. They can be configured for each operation. Almost all inverters from the major manufacturers, are capable of grid forming, which means they can start with no grid reference. These units can be paralleled to create a backbone to which thermal generation can actually sync to. Basically the more solar farms that pop up, the more battery storage systems will be needed. All utility scale inverters have some sort or primary, secondary, and reserve frequency response algorithms built in. They’ll just convert this mismatch into heat ( vars). Let me know if you have any questions. I can provide clear explanations on anything else. 😊