Comparison of sensors for the measurement of temperature with (small) UAV/RPA

Transcription

1 Comparison of sensors for the measurement of temperature with (small) UAV/RPA Norman Wildmann, Jens Bange Center for Applied Geosciences, Eberhard Karls University Tübingen 19 February 2013, Palma de Mallorca 8.98 Flight path 1000 Static pressure Longitude [ ] pressure [hpa] Latitude [ ] time [s] 0 5 hole probe differential pressures roll [ ] pressure [Pa] time [s] pitch [ ] yaw [ ] time [s]

2 Outline 1 1. Overview of temperature measurement principles 2. Description of sensor under investigation 3. Comparison to 99 m tower 10 minute averages 4. Comparison to radiosondes and remote sensing instruments 5. Spectral analysis 6. Error analysis 7. Conclusion

3 Temperature sensing principles 2 contact type temperature sensors mechanical thermal expansion liquid thermometer, bimetal thermometer vibration quartz thermometer pressure gas pressure thermometer electrical resistance PT100, NTC (thermistor), PTC thermoelectric (Seebeck) Thermocouple, thermopile capacitance capacitance thermometer inductance inductance thermometer bandgap IC thermometer (diode) thermal phase change liquid crystal, temperature sensitive paint optical fluorescence fiber optical thermometer non-contact type temperature sensors optical thermal radiation IR thermometer, thermograph, microwave temperature profiler back scattering (Rayleigh, Mie) temperature LIDAR, OADS (optical air data system) absorption TDL acoustical sound velocity ultrasonic thermometer

4 Temperature sensing principles 3 contact type temperature sensors mechanical thermal expansion liquid thermometer, bimetal thermometer vibration quartz thermometer pressure gas pressure thermometer electrical resistance PT100, NTC (thermistor), PTC thermoelectric (Seebeck) Thermocouple, thermopile capacitance capacitance thermometer inductance inductance thermometer bandgap IC thermometer (diode) thermal phase change liquid crystal, temperature sensitive paint optical fluorescence fiber optical thermometer non-contact type temperature sensors optical thermal radiation IR thermometer, thermograph, microwave temperature profiler back scattering (Rayleigh, Mie) temperature LIDAR, OADS (optical air data system) absorption TDL acoustical sound velocity ultrasonic thermometer

5 Description of sensors under investigation - Thermocouple 4 type E - (CHROMEGA R chromium nickel alloy and Constantan R ) measurement range: C resolution: 0.01 K Wire diameter: 25 µm/13 µm

6 5 Description of sensors under investigation - FWPRT Center for Applied Geoscience Resistance: 100 Ω/10 Ω measurement range: C resolution: K Wire diameter: 25 µm/13 µm

7 Comparison to 99 m tower 10 minute averages 6 Altitude agl / m :39:57 Tower PCAP/TC FWPRT temperature / C Altitude agl / m :07:45 Tower PCAP/TC FWPRT temperature / C

8 Comparison to 99 m tower 10 minute averages temperature / C FWPRT mean error: 0.45 K σ = 0.13 K temperature / C Thermocouple mean error: K σ = 0.13 K Center for Applied Geoscience Temperature measurement error / K Temperature measurement error / K

9 Comparison to radiosondes and remote sensing instruments 8 FWPRT, 17:00:00 UTC Thermocouple, 17:00:00 UTC Altitude above sea level / m Radiosonde 16:45:00 UAV 16:30:51 asc UAV 16:42:31 des UAV 16:45:11 asc UAV 16:47:31 des Tower 16:40:00 Profiler, 16:29:00 Sodar, 16:30:00 Ground level 112m asl Altitude above sea level / m Radiosonde 16:45:00 UAV 16:30:51 asc UAV 16:42:31 des UAV 16:45:11 asc UAV 16:47:31 des Tower 16:40:00 Profiler, 16:29:00 Sodar, 16:30:00 Ground level 112m asl Virtual temperature / C FWPRT Virtual temperature / C Thermocouple

10 Spectral Analysis - Variance Spectrum 9 Temperature variance / K 2 Hz 1 1e 09 1e 07 1e 05 1e 03 Thermocouple 25µm FWPRT 25µm frequency / Hz k 5 3

11 Spectral Analysis - Structure Function 10 Structure function Thermocouple 25µm FWPRT 25µm x lag / s

12 Error Analysis Temperatur [Â C] FWPRT time series PCAP time series linear trend without radiation Zeit [min] Windtunnel test at 20 m s 1. Artificial radiation of 800 W m 2 is switched on three times for short periods.

13 Error Analysis 12 no shield shield Average temperature / C leg North South leg South North leg East West leg West East square number Average temperature / C leg North South leg South North leg East West leg West East square number

14 Error Analysis 13 Average temperature / C no shield leg North South leg South North leg East West leg West East square number Average temperature / C Other errors: shield leg North South leg South North leg East West leg West East square number Adiabatic heating Icing other contamination not observed / smaller than measurement accuracy fatal for both sensors fatal for both sensors

15 Conclusions and outlook 14 Absolute accuracies of both sensors within ±0.2 K when calibration offsets are removed. Thermocouple is fast enough to resolve turbulence up to 10 Hz. Thermocouple cold junction not stable enough for the compensating temperature sensor. FWPRT with same diameter as thermocouple has a slower response time. Total hardware cost of less than 100 EUR per sensor.

16 Conclusions and outlook 15 Absolute accuracies of both sensors within ±0.2 K when calibration offsets are removed. Thermocouple is fast enough to resolve turbulence up to 10 Hz. Thermocouple cold junction not stable enough for the compensating temperature sensor. FWPRT with same diameter as thermocouple has a slower response time, but better absolute accuracy. Total hardware cost of less than 100 EUR per sensor. Acknowledgements Frank Beyrich, Udo Rummel (German Met Service, DWD) for the possibility to measure at the observatory in Lindenberg and making the data available.