Reminds me of an extremely similar case with a long distance microwave link at a mobile telecom provider in Australia that I worked for. They relied quite heavily on microwave link chains and this particular one was in northern Queensland where fixed lines were hard to find and no local engineers were locally present/aware of the changing environment.
Every week day + Saturday, from 7-3 the link would keep cutting out intermittently. Then work fine and the rest of the day and on Sunday… a crane, building a new residential building would operate during those hours right in the middle of the microwave path. Many weeks of theories and time wasted until someone had a chance to visit. :)
Amazing. Reminds me of the fact that militaries really don't want wind turbines in areas where good radar coverage is important (case in point: the Finnish Defense Forces anywhere near the Russian border); even though the blades aren't metal, they're still a source of noise and radar shadow.
I'm not at all knowledgable about this, but: is it feasible (and if so, hard?) or impossible to have some sort of live reporting from the turbines about the speed/position of their blades that connects to the radar system allowing it to ignore what it knows to be turbine noise/shadow and therefore be able to have turbines there and still get good radar?
Well, you can't just ignore a radar shadow or noise. Just like GP's point-to-point microwave link couldn't just "ignore" the crane. You can't make a bad Wi-fi connection faster by just telling your computer to ignore the wall between you and the hot spot. A wind turbine is a solid obstacle that conceals stuff behind it, and even if a single turbine might not be a big deal, most commonly someone interested in harnessing wind power would want to build an entire farm, which would be much worse.
My (probably ignorant) thinking was that if there weren't propellers, just stalks, they could be arranged in a pattern that radar setups at multiple locations could see through gaps and between them have no dark spaces caused by the towers. Leaving the problem that the blades essentially block out an entire circle, but if the radar software knows the position of every turbine's blades (by a combination of turbines reporting location/speed/acceleration of blades in real time and maybe modelling so the radar system can know in time at least fairly if not very accurately where all blades are in a field of them) then when a radar bounces of one it can say "that doesn't count as a hit" leaving only non-turbine objects showing up in the UI that the radar setup outputs?
Is it that radar can't ping / receive at a high enough frequency to distinguish the difference between "this fraction of a ms the blade was at that location so we don't care about the radar hitting something, but the next fraction of a ms the blade had moved and we still got a ping from just behind it so there's something there"? Or some other problem with the idea?
Radars work in the 3cm or 10cm ranges[0], so you'll always get a return from the towers and the blades, and that significantly reduces the energy that gets beyond the turbine to illuminate potential targets.
Reflections from targets also need to pass the turbine(s), reducing the return even more.
Beamwidths are significant, depending on the antenna configuration. You can have beamwidths of 0.5 degrees and more. That helps detection, but reduces target discrimination. It helps get the energy past the turbines, but you tend to have "wound down" the power output to allow for that. So it doesn't really help.[1]
You know where the turbines are ... knowing the exact positions of the blades won't help at the frequencies being used.
In short, it's complicated. Radar pulses are like HUGE lumps of energy flying through space, and delicate things like turbine blades chop them up and defect them around, preventing them from getting to, and then from, the targets.
[0] Marine radars, X-Band and S-Band. There are other ranges and bands.
[1] Yes, this is all extremely simplified and inaccurate, but I'm trying to give a sense of what's going on.
Thanks! I wouldn't have guessed the sizes were so large for radar, and just assumed that in current era radar would have progressed to the point of even being able to not send pings at the rotating blades in sync with their rotations (ie always firing radar somewhere within the circle but only at the spots where the blades are not) the same way machine guns fired past propellers on WW2 (or WW1?) fighter planes figured out. But of course a bullet and a radar setup isn't at all the same thing.
The details I give are for marine radar, where off-shore wind-farms are a significant issue. But radar in general works on similar wavelengths, and a lot of clever processing can be done to "improve" the images you get back, especially when you have a lot of images that can be combined, rather than sweeping the radar around and only get 20 to 120 sweeps per minute.
There are trade-offs between how often you can ping versus how far you want to see versus how much energy you can send out versus how big the "lumps" are, etc. The sums/calculations are pretty simple, but not obvious until you've seen them, and then they are.
But if you want to see something at 32NM ~ 60km, you have to send out a lot of energy, so the "lumps" need to be pretty big, especially as the returns drop off as a fourth power with distance.
Any idea to what extent the problems can or can't be solved by money vs. a limit on maximum quality of radar before needing to break the laws of physics to improve further?
E.g. are the US military saying "yep those are the issues we face too" or "haha our tech has more than enough energy to get around all these problems"?
Most of what I know in these areas is either something I can't talk about, or is surmise based on conversations between others when I was around. There's a lot to do with the laws of physics, but equally, many times people have found clever things to do with the data they have.
I wouldn't care to speculate further, but you might want to read about phased array antennas, and FMCW radar. There are some very deep rabbit holes here, and I certainly would not be your best guide.
Even if you ignored the turbines themselves, there would still be a "shadow" behind the turbines though I think? Which means you'd have a blind spot every now and then (when the wind is blowing and the turbines are active) which could be exploited by an enemy
Even if you are looking at the turbines "edge-on," there's probably going to be a noticeable Doppler return as well.
Plastic drones with plastic propellers are still visible on radar because the tiny propellers spin super fast, so they light up like a Christmas tree on Doppler radar because the approaching vs receding velocities of the blades are so different.
I've done a ton of low-budget analog hardware debugging, and the major problem with hardware debugging is each attempt to fix the problem takes a long time. If I had wanted to test every idea I had I could easily waste a week. Not to mention that I can't just run some automated test suite after the fact. For hardware, approaching debugging methodically is a necessity, not just best practice.
We don't typically have log files for hardware, but I'm always surprised when otherwise extremely intelligent people first try to debug by applying "fixes" that shouldn't have any causal effect on any weird observations we've gotten. I have no problem with people coming up with theories because each modification takes time, but each theory should ideally explain the data...
Reading log files with really obscure error messages might as well be reading a magical grimoire. Especially when the solution turned out to have nothing to do with the error message.
Every week day + Saturday, from 7-3 the link would keep cutting out intermittently. Then work fine and the rest of the day and on Sunday… a crane, building a new residential building would operate during those hours right in the middle of the microwave path. Many weeks of theories and time wasted until someone had a chance to visit. :)