We’ve been speaking with an owner of a few quite large land-based wind turbines. There are 5 wind turbines over various heights, smallest being 40m above ground level, the highest 60m above ground level. The ground actually being a reasonable sized hill, most of the wind turbines bases being 420m above sea level.
The owner has agreed to allow us to place one LoraWAN gateway with 4G backhaul inside the nacelle (the housing behind the hub of the turbine blades) of one of the wind turbines. Apparently it is not a problem to get a UK 3-pin 240v supply up there, the only requirement being it needs to be inside the nacelle. The nacelle is made from fibre-glass. The thickness of which, I don’t know. Does anyone know if fibre glass stops 868Mhz frequencies? I know that glass in windows isn’t great for LoraWAN signals and fibre glass is just blown glass covered with resin but it’s interlaced cross-crossing and there is more resin which is plasticky and plastic is kind of okay with LoraWAN signals…. :-/
If you can mount outside on the ‘A-Frame’ (or what ever shape the specific manufacturer has gone with ) that often sits atop the turbine nacelle - usually at the far end away from the actual blades. I looked at this a few years back (for a major set of sites some miles to the north and to the N.East of you ) and that seemed most viable option - They usually have weather stations/aneometers back there (to monitor if wind/weather conditions are turbine safe or it they should go into a ‘parked’ mode, safe for worker ‘up top’ etc.) often along with other antenna (Ground comms, cellular monitoring, mesh with other turbines etc.)
As for the fibre-glass itself - well it depends! Similar structure/construction often used for the Radomes on aircraft (nose cones), and need to be designed for specific RF windows (Radar, comms, IFF etc…depending on type of craft). I played with some fibregass structures a while back out of curiocity (sheds/temp office blocks, even kayaks/canoes) with variable results… one factor was the purity of any resins used - and deliberate contaminents - e.g. colourings added…metalic or black carbon a pain! Even more important is the type of weather proofing coating that might be applied - again metalic/high metal(oxide) or partially conductive paints generously applied can be worse than the actual structure. TBH just go for it and test (I’m jealous!) to see what you get…likely little worse than using an indoor gw/antenna - especially if it has to work through metalised tripple glazing, foil back plasterboard or foil covered cellotex type insulation boards etc. But what the hell, you have major height advantage to compensate! The other things to consider (Simulations showed little significant impact over time) is the rotational masking caused by the actual blade motions in the forward direction, also the area immediatly around the turbine at ground level below can be in a minor RF shadow due to the propogation profile of the antenna (not great in direct vertical), though ironically and helpfully the scattering effects (resiliance) of LoRa signals of the close proximity metal work can actually help! Getting outside also helps move you further away from the big internal metal blobs (generator, etc!) that will inevitably mask RF coverage to a certain extent in the direction of the metalworks alignment with the ant. You may also find that if tall and on a hill that many deployed nodes may not be effective at lower SF’s down in the valleys, especially if ‘below’ the GW, as they will have exhausted the free space link budget just reaching ‘up’, so you may find they bump up to SF8 or SF9 c/w normal operation in the horizontal at SF7. Dont make the common mistake of thinking that a higher gain ant will help/compensate as that will only concentrate coverage out to the horizon vs close proximity below and may even introduce undesirable ‘notching’ due to polar diagramme properties. A lower gain ant is typically closer to isotropic radiation/coverage pattern and gives more consistent coverage…
Certainly plenty of RF gets in & out of previous Reliant Robin of the air: ASW15 - generally the main radio antenna would be run down the inside of the fuselage - anything that sticks out is prone to being snapped off.
) that often sits atop the turbine nacelle - usually at the far end away from the actual blades. I looked at this a few years back (for a major set of sites some miles to the north and to the N.East of you 
I think we’ll just have to give it a “whirl”…