Comparison of Wind Sensors Ultrasonic versus Wind Vane/Anemometer

October 2000

Kenneth G. Wastrack Doyle E. Pittman John E. Hatmaker L. Wayne Hamberger Tennessee Valley Authority

[NUMUG] Comparison of Wind Sensors 1 of 14

Background



Problem - Obsolete Wind Sensors

–

Difficulty obtaining spare parts (sensor, switch controller, etc)

–

Routine bearing replacement

–

Occasional history of mechanical failures



Requirements - Sensor Specifications

–

Comply with RG 1.23 requirements.

»

Wind direction +/-5 degrees.

»

Wind speed +/-0.5 mph.

–

Less maintenance than existing sensors.

October 2000 [NUMUG] Comparison of Wind Sensors 2 of 14

October 2000

Vaisala 425AH Ultrasonic Wind Sensor



Sonic Wind Sensors

– –

All-electronic Minimal maintenance



Vaisala* (formerly Handar) 425AH ultrasonic wind sensor most suitable

– – –

Manufacturer specifications Capabilities for integration System meets RG 1.23

[NUMUG] Comparison of Wind Sensors 3 of 14

System Performance

RG 1.23

Climet Vaisala Wind Direction Wind Speed +/- 5 deg +/- 4.6 deg +/- 4.7 deg +/- 0.5 mph +/- 0.48 mph +/- 0.48 mph

October 2000 [NUMUG] Comparison of Wind Sensors 4 of 14

October 2000

Principle of Operation



Transmit signal along axis



Send in reverse direction



Calculate time differences



Convert to WS along axis



Repeat for other axes



Compute WD and WS

[NUMUG] Comparison of Wind Sensors 5 of 14

1.0

0.5

0.0

-0.5

-1.0

Wind Tunnel Tests

0.6

0.4

0.2

0 -0.2

-0.4

-0.6

34 28 22

Upper Limit

16 10 6 3 3 6 10

Nominal Test Speed (mph)

16

Lower Limit

22 28 34 41 Knoxville Speed Error (@10 mph) Oak Ridge Speed Error (@30 mph)

Wind Direction (degrees)

[NUMUG] Comparison of Wind Sensors October 2000 6 of 14

Field Tests



Establish that sensor works as expected.



Document that data are comparable to wind vane/anemometer data.



Verify software and hardware changes.



Provide maintenance experience.

October 2000 [NUMUG] Comparison of Wind Sensors 7 of 14

October 2000

Field Sampling Configuration

wind vane/ anemometer ultrasonic wind sensor ultrasonic wind sensor

~3.4 meters Tower width ~ 1.2 meters (48 inches)

Meteorological Tower Anemometer mounted due North of wind vane.

[NUMUG] Comparison of Wind Sensors 8 of 14

Comparison of Wind Roses

Climet Vaisala

October 2000 [NUMUG] Comparison of Wind Sensors 9 of 14

Statistical Results

count (N) - nondimentional minimum difference maximum difference systematic difference (d) est. standard deviation of difference (s) operational comparability (C) skewness (M) - nondimentional kurtosis (K) - nondimentional

correlation coefficient - nondimentional

Note:

VWD (°) 7973 -10 10 0 2.4

2.4

0.2

-0.2

1.000

Differences in 15-minute values (Climet minus Vaisala) VWS (mph) AWS (mph) 7833 -3.8

1.0

0.0

0.37

0.37

-1.64

10.90

0.998

7827 -3.9

1.0

0.0

0.37

0.37

-1.62

10.80

0.998

N, d, s, C, M, K are specified in ASTM D 4430-96, “Standard Practice for Determining Operational Comparability of Meteorological Measurements”.

ST (°) 7976 -9 9 0 0.9

0.9

-1.4

24.3

0.975

October 2000 [NUMUG] Comparison of Wind Sensors 10 of 14

Problems Encountered



Wind Tunnel Tests

–

Tunnel Configuration

–

Echo Effect



Software Changes



Alignment Error



Error Handling



Tower Shadow Effect

October 2000 [NUMUG] Comparison of Wind Sensors 11 of 14

Results

Vaisala Ultrasonic Wind Sensor is equivalent to the existing Climet wind vane/anemometer system.

Additional Benefits



Lower starting threshold



Continued functioning during icing conditions

October 2000 [NUMUG] Comparison of Wind Sensors 12 of 14

Pre-Operational Issues



EMI/RFI potential



Uniformity of sensors



Changes to operational practices and revision of procedures

System Changes



Removal of wind translators



Connection of ultrasonic sensors directly to data logger computer



Software changes



Discontinue strip-charts

October 2000 [NUMUG] Comparison of Wind Sensors 13 of 14

Conclusion

The Vaisala ultrasonic wind sensor meets the accuracy requirements of RG 1.23 and can be used as a replacement for the Climet wind speed/anemometer system.

TVA plans to install the ultrasonic wind sensors at its nuclear plants as soon as EMI/RFI testing is completed and applicable documentation (procedures, FSARs, etc.) are updated.

October 2000 [NUMUG] Comparison of Wind Sensors 14 of 14