Monday, 16 September 1996


Geoff Hill
Venco/Westwind
Kelmscott, WA
Fax: (09) 497 1335

Dear Geoff,
		re: Breamlea Wind generator

I was on site last  Thursday, 12th September, to meet a film crew for a
childrens' TV show.  The weather was too windy for them to go up the tower,
so they departed. I stayed to observe the W/G's performance in westerly gale
force winds.

At 2.28 pm in a particularly vicious blast of wind (longer duration than an
ordinary gust) the W/G went into overspeed, and within half a minute there
was a puff of black smoke out the rear vent as the brake pads vapourised.
Only two of the air brakes had deployed, and at one stage the rotor appeared
to reach the resonant freq. of the tower (1.33 Hz by my previous measurement,
or 80 rpm equivalent). What I saw was the tower swaying considerably in time
with each rev of the rotor, but fortunately this only occurred for a couple
of seconds during another severe gust. As I was not observing the LED display,
I unfortunately do not know how long the machine was in overspeed before the
gen. tripped, or how many more seconds until the 2 air brakes came out. My
recollection is that the black "smoke" from the pads appeared after the air
brakes deployed, indicating that the brake at no stage adequately controlled
the  rotor speed once the generator tripped.

I continued observing upwind at the meter hut, and decided to enter the tower
during a lull in the wind, as at these times it was apparent that the rotor
was returning to a relatively safe speed of rotation. I must say that the
sound of the air brakes renting the air at 200 km/hr is quite impressive, not
unlike a large steam safety valve venting high velocity steam!

Once in the tower, I finally learned how to slew the W/G from the relative
safety of the tower base, using the microprocessor safety override buttons.
The disc brake was then secured, while still way too hot to touch. Brake pad
compound had fused to the shiny segments of the disc. As the winds were still
pretty bad, I didn't hang about up there too long. This was only about
10 minutes after the initial overspeed event started.

Yesterday, Sunday 15 Sept. was our scheduled maintenance day, so we have
now replaced the brake pads, scraped the old melted pads off the disc, and
reset one of the blade tips, which remained flipped-out (it was very stiff
to re-set). Some large flakes of rust were found in the bung when it was
removed, possibly from the spring.  Unfortunately a test-run in high mode
revealed braking times as long as 20 seconds, and the air brake did not
remain properly seated.  In addition we have found significant cracks in
the leading edge of this blade, both in the air brake section and in the
main blade section. (The join between the two halves of the blade is failing
in both cases.)

There are three main issues arising from the current situation:

Air brakes: the machine is not safe to operate until we demonstrate that
all three are deploying at the design speed. I propose to remove them all
and overhaul the moving parts. Full details of how to do this would be
appreciated. As we cannot afford a cherry picker, I plan to secure the tips
by rope from the nacelle, and lower them to the ground, with the help of
another guide rope from the ground.

Cracks: We definitely cannot afford to remove a blade, so a method will
have to be found to effect a strong and durable repair in-situ. I think it
will be feasible to get access with a steel caving ladder suspended from the
hub, from which the lucky repair man (me?) will be safely suspended by the
usual harness/karabiner setup. Your advice about suitable compounds to
repair GRP would be appreciated. I envisage drilling small channels into
the cracks, injecting  a gel-consistency glue with a syringe, and possibly
clamping the whole length of the crack  while the glue sets. The surface
will then need repair with resin and tape as well.

Brake: The compound supplied is presumably an asbestos-free one, but appears
not to have a high-enough coeff. of friction at  low temperatures: stopping
times did improve considerably to about 8 seconds if a few stops were done
in rapid succession. In any case we cannot rely on it to stop the rotor in
a gale. I propose to immediately install a "crosswind slewing" windvane
which will slew the W/G out of the wind every time that the brake solenoid
is deactivated. The brake pads still have to be able to survive for the
1 minute 45 seconds it takes to slew 90 degrees, and from what I saw on
Thursday that is "a big ask" of the pads. Of course the slewing won't  work
if there happens to be a mains failure during a severe gale. In the longer
term, I believe it may be worthwhile to yaw the rotor partially out of the
wind, based on excessive slip being registered on the micro-controller. Or
do you think  automatic shut-down would be a safer response to excessive
slip?


			Yours Sincerely,           Michael Gunter