Grid Stability Basics

Guest post by Rud Istvan

Charles asked whether I would critique this new nonsense. Sure, why not.

The highlighted ‘sciency’ article purports to create an ‘open source’ grid data resource using 17 ‘synchronous grid’ European regions to design and test how “that will help develop new power systems capable of meeting changing demands, such as the move towards renewable energy sources.” NOPE, this won’t help at all.

The body of this silly paper found observationally lower grid frequency variations in larger ‘isolated’ European subgrids!  Of course larger grids have less frequency variation— something trivially true and known for many decades. The reasons are intuitive and simple. Larger grids by definition have more generation feeding the grid, so more grid inertia, and at the same time less ‘instantaneous load variation’ by simple virtue of the statistical law of large numbers.

Grid inertia is just the frequency stabilization that comes from the kinetic energy of large rotating generator masses. As grid load ‘instantly’ increases, the frequency sags and the generator wants to slow down. But by slowing, it also ‘instantly’ injects some of its rotational kinetic energy until the driving source (steam, or natural gas in CCGT) can be turned up a bit to compensate.

CCGT is particularly good at this. It runs about 61% efficient at full load, 60% efficient at 80% load, and still 59% efficient at just 40% load. The gas turbine power reaction time is literally about a second no different than a jet engine.

None of this ‘open source data resource’ helps ‘the move to renewables’. Basic grid electrical engineering (EE) is simple and inescapable concerning renewables— they are uneconomic (investment always falls when subsidies dry up), intermittent (requiring underutilized backup generation), and provide no grid inertia (which is automatic with big fossil fuel fired generators). ALL BAD.

Irsching Bavaria‘s about 10 year old CCGT units 4 (375MW) and 5 (860MW) illustrate the problems that intermittent renewables lacking grid inertia create for any grid. The owner wants to take them out of service because very unprofitable when run mainly as wind backup rather than grid load. The German government refuses to make up the difference, yet also refuses to let them shut down.

In the bigger German renewable picture, northern Germany exports surplus wind power to Norway as essentially a giveaway. That lets Norway throttle back its hydro. Then, when Germany needs power because wind is insufficient, Norway spools up its hydropower and sells it to Germany at exorbitant rates. To make matters worse, the German NIMBY crowd won’t let the utilities install more much needed North-South German transmission interconnect ‘eyesores’—despite wind turbines that are bigger and uglier.

A small amount of renewables in a large stable grid presents no problems–and no need for this new study. The backup capacity (spinning reserve) and grid inertia are already there. But as renewable penetration grows as a proportion of any stable ‘synchronous’ grid, these issues grow and compound. Depending on grid details (like how much flexible hydro), renewable penetration above about 8-10% ALWAYS creates extra costs and stability problems.  Alternating current grid engineering has been known now for well over 100 years, and nothing ‘new’ can change the established maths, physics, and electrical engineering (EE).

There is also some ‘basic EE stuff’ this new ‘open source data resource’ ignored.

First, their ‘new’ grid frequency data is almost always available from the utilities (grid operators), since required to manage the grid. It is captured both at the generating plants and at the transmission to distribution substations. Perhaps they should have just asked for the utility data rather than gathering their own.

Second, their super duper ‘proprietary’ grid frequency detectors for capturing their new open source data can be purchased almost anywhere. Such detectors are a common electrician’s tool. A very high-end multipurpose ‘pocket’ unit is the Extech PQ3450 power quality analyzer, pictured.

Third, their finding that grid frequency sags when demand exceeds supply and rises when supply exceeds demand has been known since Tesla first envisioned AC generation by using complex numbers (a+bi). The US is designed to be a stable 60 Hertz, EU 50 Hertz. Because electric clocks depend on grid frequency,  ‘stable’ is about plus/minus 0.2Hz in the US.

Just for fun (because the above is all well trod ground at WUWT and Climate Etc), we also provide a simple math/science fun challenge to those at WUWT unfamiliar with AC electrical stuff. (Thanks to a physics book full of hundreds of such challenges, sadly in Chicago rather than Fort Lauderdale, which I half reconstructed from memory and half from Google ‘facts’). It is about right.

Imagine a high voltage AC transmission line between Chicago and New York. (Hint: the distance does not matter much for this challenge, voltage does some, but is always ‘high’.)  The challenge: how far does a single electron charge travel during its journey along this transmission line? Choices:

1. Kilometers
2. Meters
3. Centimeters

The speed of light (aka electromagnetic radiation) is about 299792458 meters/second in a vacuum. But ‘electromagnetic force (emf) wave guide’ wires are definitely not vacuum; their metals have electrical resistance (technically grid AC impedance, a combination of circuit resistance and reactance). Although electrons pushed along by emf have essentially no mass, they still slow down lots in metals.

Skipping a bit of complex math explaining very simple physics, the reason you buy ‘fat’ DC audio cables with gold terminals is simple: DC current flows across the entire conductor cross section fairly uniformly. A fatter cable has uniformly less cross sectional resistance (DC speaker output is analog), and gold terminals lower connection resistance further since gold doesn’t corrode. The result is truer high frequency pitch amplitudes. As a side factoid, this same physics is why ‘fat’ high voltage DC transmission lines are preferred for long grid distances, like the Germany to Norway undersea interconnects mentioned above.

The ‘fat’ AC cable answer is different. AC current has a conductor skin effect. Unlike DC current, AC current travels mainly in an annular ‘skin’ ring whose thickness depends on frequency (higherèthinner) thanks to interesting AC consequences of Maxwell’s equations as explained by Feynman’s ‘Lectures on Physics’ V. 2 chapter 32, ’skin effect’. In a ‘pure’ copper conductor, that annular skin travel (the center of the conductor essentially carries no AC current) is only about 3.2 meters per second. The much more common (cheaper, lighter) aluminum HV transmission conductor is only about 61% of copper. So in aluminum high voltage transmission lines electrons travel at best about (3.2m/s*0.61) ~1.95 m/s ignoring skin oxidation.

 AC is a sine wave varying from 0 to plus emf to 0 to minus emf to 0 at x times a second, (US 60Hz, EU 50Hz). One cycle is from zero voltage to max emf to zero to opposite polarity max emf back to zero. The emf back and forth is a nominal US 120 peaks/second, 60 times in each direction. (Side note: unlike most DC circuits, AC emf is related to, but not the same as, voltage.) For simplicity imagine this AC challenge as simply digital rather than sine wave (all on or all off emf rather than a varying emf sine wave), thereby removing any need for the correct calculus challenge formulation. Then the most distance a virtually massless electron could travel in its aluminum high voltage skin is about (1.95/60) 3.2 cm, back and forth and back and forth. The actual (sine wave emf calculus) US answer is less than one centimeter back and forth, and EU less than 2. That is because there isn’t a lot of emf except near the peak and trough of the sine wave, slowing down intermediate time things further. Either answer (over simplified or correct) results in not very much electron travel distance (c).

Returning to the main topic of this guest post, despite this new European ‘ grid research’:

Anywhere, anytime, renewable subsidies get reduced, so does investment in them. They are provably uneconomic, stand alone.

Renewables are intermittent; yet do not cover the backup costs of intermittency.

Renewables provide no grid inertia; yet do not cover the costs of providing frequency stability using massive synchronous condensers.

More research and new ‘open source’ data resources concerning fundamentally bad propositions does not improve them. NOPE!