Bat Deterrent Device - Bats and Wind

Privileged and Confidential Information of DEI, Copyright 2012

Bat Deterrent Device

PREPARED FOR


Objective

•  To deter bats from approaching wind turbines using broadcast ultrasound in an effort to reduce bat fatalities at wind farms


Performance Targets

•  122 dB SPL at 1m distance (50 kHz)

•  Effective range of 25m at 50% RH

•  Effective range of 40m at 10% RH

•  Frequency range of 20 kHz – 100 kHz


Performance Targets


Initial Design

•  SensComp Series 600 Environmental Grade Electrostatic Transducer §  Intended for operation in air at ultrasonic frequencies

§  Resonant frequency of 50 kHz, with a defined typical transmit response over 20 kHz – 100 kHz range

§  Better suited for harsh/outdoor environments

•  16 transducers (4x4 matrix) mounted on each device to achieve desired SPL §  110 dB with 1 transducer (per data sheet)

§  122 dB with 16 transducers (per calculations)

•  Transducers driven with 300V pulses at pseudo-random frequencies across 10 kHz – 100 kHz range


Initial Design

•  Devices intended to be weather-resistant, but no environmental testing performed

•  LED on each device to show supply power is present

•  Timers used to power devices on for 14 hours each night

•  8 devices per wind turbine §  3 on each side of nacelle, pointing down

§  1 on each end of nacelle, with reflectors to direct ultrasound up

•  300 VDC power supply §  120 VAC input power

§  4 discrete channels per supply

§  1 supply powered 4 devices

§  Installed inside nacelle with cables routed to each device


Initial Design

•  $7740 (materials) per wind turbine §  $340 per power supply (2x)

§  $570 per device (8x)

§  $2500 per mounting, cabling, etc. (1x)

Initial Design


Initial Design


Initial Design



Field Test - 2009

•  80 devices deployed in summer of 2009

•  Installed on 10 turbines at Locust Ridge Wind Farm in PA


Field Test Results

•  Water caused component damage in nearly all devices within first 6 weeks of installation §  Indicated by LEDs that turned off

•  All devices and most power supplies required rework/replacement before or during the field test §  Transducers, transistors, and control boards replaced as needed

§  Additional sealant added to all seals

§  Vent added to allow operation with improved seal

•  Significant downtime for many of the devices, resulting in poor coverage during portions of the field test

Failure Analysis

•  Devices were not properly sealed §  Water was able to enter the devices around the heat

sink and cause damage to the internal electronics

•  Circuits did not have sufficient protection §  Water damage resulted in permanent damage to the

electronics

§  Water damage resulted in shut down of the entire device


Design Changes - Sealing

•  Designed & installed a custom gasket between the heat sink and the enclosure

•  Gasket provided a better seal than the silicone used in the previous design


•  Custom gasket inserted between heat sink and enclosure •  4 extra holes drilled in enclosure & heat sink to provide more

uniform compression around gasket •  Thread-sealing Loc-Tite or rubber washers used on all screws




•  Pressure leak (pressurized to 0.5 psi, bubble-tested for leaks) §  No leaks detected on 3 different gasket installations

•  Full submersion (submerged to 17” depth for 12-48 hours) §  No water entry detected on any enclosure

•  Simulated “rain” (subjected to “rain” per UL standard test method) §  No water entry detected on any enclosure (1 hour min.)


Design Changes – Circuit Protection

•  Added thermistors to protect each pair of transducers •  Added thermistors to protect each output channel of the

power supplies •  Components allow for recovery of circuitry once the fault

condition is removed (open when hot, close when cool) •  Components in device only shut down 2 transducers at a

time so the entire device is not rendered nonoperational


•  No impact in normal operating conditions •  Little to no impact when rear of transducers

sprayed with water •  Current limited (i.e., transducers off) when 1 or

both transducers submerged in water •  Current limited (i.e., transducers off) when rear of

transducer flooded with water



•  Performance returned to normal after fault was removed (i.e., transducers were dried off)

•  No PCB components were damaged •  No fuses were blown

§  Only exception was a cold start with all 16 transducers flooded with water

§  Still no damaged components



Design Changes – Other

•  Alternate enclosure venting method §  Water-proof vent from Gore

•  GFCI circuit breakers added on incoming AC •  Mounting methods redesigned for ease of

installation and anti-rotation

•  $13,500 (materials) per wind turbine

§  $355 per power supply (2x)

§  $620 per device (8x)

§  $7830 per mounting, cabling, etc. (1x)



Field Test - 2010

•  Prior to field test, devices were installed on 1 tower in spring of 2010 at Locust Ridge

§  Upon inspection in July, main GFCI CB was tripped, but when devices were powered back up, all were functional

§  Design was changed to utilize non-GFCI CB

§  Of the 7 devices inspected, all were dry on the inside

§  Fit & function of new mounting methods verified

•  80 devices deployed in summer of 2010

•  Installed on 10 turbines at Locust Ridge


Field Test Results

•  Anti-rotation clamps performed well §  1 tower retained the old mounting method; the mounting arms

rotated on this tower, but not on the others

•  5 welded joints on the mounting arms were broken §  Some were completely broken prior to removal, others were

damaged and then broke completely during the removal process

§  It was observed that the mounting arms experienced a great deal of bending and twisting, which likely led to the failure

§  These same mounting arms withstood the previous field test in 2009


Field Test Results

•  Devices still allowed water to enter, but they were much more resistant than previous efforts §  Water was detected in 27 of the 80 devices

§  Water was not detected in any of the devices that were mounted farthest from the turbine blades

§  Water was not detected in any of the devices with the reflector

•  Fuses were blown in 27 of the 80 devices, but it is unclear whether the fuses blew before, during, or after the field test


Field Test Results

•  During weekly checks of the 60 devices visible over 8 week period (8/5 - 9/26) §  2 devices were not working (no LED); these were replaced

§  12 devices had blinking LEDs, indicating that the power supply channel was current-limiting; most of these recovered on their own

•  Intense storm (heavy rains, 70 mph winds) passed through on the week of 10/3 §  26 devices had blinking LEDs; only a few recovered before last

check on 10/8 and 2 stopped working


Field Test Results

•  Most of the power supplies did not have any damage §  62 were still regulating at approx. 300 VDC

§  16 were supplying unregulated voltage (approx. 370 VDC)

§  2 were supplying low voltages (approx. 185 VDC)

•  Devices with a high voltage supply will continue to operate, but with more downtime as temperatures will be higher inside the device

•  Devices with a low voltage supply will continue to operate, but they will be less effective


Field Test Results

•  Data retrieved from the microcontroller was only partially revealing §  Number of full 14-hour power cycles was evident, but this does not

necessarily mean the transducers were operational

§  Number of overheat events was recorded, but not their duration or time-stamp

§  Finite (& small) available memory means that data was overwritten in many cases

•  Water-proof vents were effective, but need to be better protected


Field Test Results

•  Overall, much better performance & coverage than previous field test

•  Devices were able to recover from many failure modes

•  Design still needs additional ruggedization to reduce downtime & increase coverage


Proposed Design Changes

•  Improve the environmental rating of the devices •  Protect water-proof vents from physical damage •  Improve the strength of the mounting arm welded joints •  Redesign mounting arms to withstand the fatigue loading of the

wind


Proposed Design Changes

•  Prevent over-temperature events (i.e., blinking LEDs) •  Prevent over-current events (i.e., blown fuses) •  Add software control to automatically shut down device if poor

performance is detected •  Improve on-board diagnostics & data logging •  Add real-time remote monitoring & control •  Reduce audible noise •  Quantify and optimize ultrasonic performance •  Investigate driving the transducers at +/-150 VDC with 150V

bias, rather than 0 to 300 VDC pulses •  Protect against incoming power transients


Development Plans – Short Term

•  Acoustic lab testing (in progress) §  0–100 kHz spectrum at 1m on devices w/1 & 16 transducers

§  Spherical dispersion tests on devices with spectra measured at 0.5m, 1m, 2m etc.

§  Directivity tests, if possible

•  Research & development ($30k)

§  Proposed design changes

§  Reassessment of selected transducer

§  Analysis of quantity/location of devices per wind turbine

•  Definition of new design requirements ($10k)


Development Plans – Short Term Initial acoustic test results


Development Plans – Long Term

•  Deployment of 600 devices (20 devices on 30 wind turbines) ($1.229m)

•  Labor Estimate ($471.5k) §  Design (6 months, $220k)

§  Assembly & Test ($187.5k)

§  Installation/Removal & Field Test Support ($64k)

•  Materials Estimate ($757.5k) §  Devices ($600k)

§  Power Supplies ($97.5k)

§  Mounting Hardware ($60k)

Bat Deterrent Device - Bats and Wind

Documents