Invited talk on calibration:Invited talk on calibration:

Univ.-Prof. Dr.-Ing. Otto HeuneckeUniv. Prof. Dr. Ing. Otto HeuneckeInstitute of Geodesy, University of the Bundeswehr Munich

The geodetic calibration line of the UniBw Munich Conception and implementation– Conception and implementation

11Wednesday, September 15, 2010

Outline

1. The new geodetic calibration line at NeubibergObj ti l i & d i- Objectives, planning & design

- Construction & side impressions

2. Determination and documentation of measurand „length“- Instruments & procedures& p

- Influence factor “meteorology”

- Calibration of low-cost meteo sensorsCalibration of low cost meteo sensors

- Use of geo sensor networks

Uncertainty of measurement: principles & implementation- Uncertainty of measurement: principles & implementation

3. Conclusion and outlook

22

1. The new geodetic calibration line at Neubiberg

Campus U i i f hUniversity of the

Bundeswehr Munich

530 m

33Further information on the University please see: www.unibw.de

Objectives

For the calibration of EDMs, tacheometers, levelling instruments, gyros etc.certified Geodetic Laboratories are obliged to: - Document and guarantee the traceability of the measurands (e.g. by intercomparison programmes);

- Document the obtained measurement uncertainty u of etalonsDocument the obtained measurement uncertainty u of etalons.

Geodetic calibration line UniBw Munich:

Primary measurands: Length and azimuth

Goals: udistance < 0.3 mm for distances up to 1100 m(better for shorter distances,see comparator of Geod.Lab.)

uscale < 0.2 ppm for distances up to 1100 m

uazimuth < 1 ´´

S C ff44

Secondary measurands: Coordinate differences

Design of calibration line

Schwendener Design 7 stations + 1 additional apart (P8)[m]

Shortest distance A 18.7Longest distance 1100.01st design parameter B 31,67 =1/15*(C-6A-U)Total length 7 stations C 590,3 =6A+15B+36D2nd design parameter D 0,08333 =1/36*UEDM half wave length U 3

P1 P2 P3 P4 P5 P6 P7 P8

55

Distribution of distances over length

0

Distribution of possible distances over total baseline

0

0

0

0

0

Min. 18.7 m Max. 1100.0 m

0

00 200 400 600 800 1000

0

0

0

Distance [m]

2 1 + 7 = 28 measureable distances

66

Distribution of distances over EDM half wave length

m[m/s] MHz MHz

[ / s]M

MM

cU ff

15 1002 2 1

EDM half wave length

Examples:fM = 50 MHz fM = 30 MHz

Distribution over U = 3 m (design parameter)

Distribution over U = 5 m

0

0

0

0

0

0

0

0

0

0

0

0,0 0,5 1,0 1,5 2,0 2,5 3,00

0

0

0

0 1 2 3 4 5

0

0

0

0

0

0

0

0

77

0

Half wave length [m]0

Half wave length [m]

Construction of measurement pillars

88

Construction details

Container for t

Measurement pillar No. 12 pillars for

additional

permanent measurement setups

(230 V, Internet)

99

equipment

Impressions from construction works in 2008

1010

Impressions from recent status

1111

Measurand “azimuth”

Leica TCM 1800Gyro azimuth

Gyromat 2000

Presentation is focused on measurand length“ only!System „Icarus“

Presentation is focused on measurand „length only!

- Astronomical azimuth (pole star)- Deflection of the vertical

1212GNSS azimuth

Equipment of Geodetic Laboratory

30 m interferometric longitudinal comparator:- Interferometer HP 5507B- Measurement uncertainty u=(0.6 +1.0L[m]) m

Climate chamber:

[m]- Air-conditioned environments

1313

- Temp. band -25°C – +45°C- Aligned with comparator

Instruments for high quality distance measurements

Leica TDA 5005Kern Mekometer 5000 Leica AT901-LR

L i AT401Leica System 1200 Leica AT401

1414

2. Determination of measurand „length“ by EDM

*Y

Meteorology

*Y

Y* : Uncorrected raw“ observation

Niemeier, W.:Zur Zuverlässigkeit geodätischer Systeme – Problemformulierung und Lösungsansätze.

Y* : Uncorrected „raw observation

1515

Zur Zuverlässigkeit geodätischer Systeme Problemformulierung und Lösungsansätze.Wiss. Arbeiten der Fachrichtung Vermessungswesen, Nr. 153, Hannover, 1989

Functional and stochastical model of evaluation

* *,1 1

p q

j jj j

Y Y c r f Y

XFunctional model:

Each correction is a function of some influence factors Xi , i = 1, … N , e g temperature and adds a respective percentage to the variance of the resulte.g., temperature and adds a respective percentage to the variance of the result. Even “centering” etc. can be seen as a correction cc with E(cc) = 0 and c

2.

2raw

XXN N

O

O

Stochastical model: Co-variance matrix of influence factors

,N N

*,2 2 TY XX

f Y

XF F FVariance of result Y :

1616

,Y raw XX F F F

XVariance of result Y :

Guide to the expression of uncertainty in measurement

Produced by:Joint Committee for Guides in Metrology

Download of documents see:

1717

www.bipm.orgBureau International des Poids et Mesures

Basic terms and definitions – GUM in brief

Uncertainty of measurement u:Parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measuranddispersion of the values that could reasonably be attributed to the measurand.

Uncertainty of measurement comprises many components. Some of these components may be evaluated from the statistical distribution of the results p yof series of measurements and can be characterized by experimental standard deviations. The other components, which also can be characterized by standard deviations, are evaluated from assumed probability distributionsby standard deviations, are evaluated from assumed probability distributions based on experience or other information.

Principles to obtain u :Principles to obtain u :- Analytical deduction (law of error propagation, Bayes theorem);- Numerical studies: Monte-Carlo-Simulation (see JCGM 101:2008);Numerical studies: Monte Carlo Simulation (see JCGM 101:2008);- Intercomparison programmes and their combined evaluation.

1818

Essential: Consideration of influence factors Xi , i = 1, … N ; N : complete.

Graphical illustration of values, error, and

t i tuncertaintySee JCGM 100:2008,Annex D, p. 53Annex D, p. 53

C l t ltComplete result:

;f f 2

Y k u

1919

: factor, often 2k k

Influence of meteorological correction on EDM

ìïïïïí

dt : variation of temperature [°C]

dp : variation of air pressure [hPa] 610 1 00 0 28 0 04d dt d d íïïïïî

de : variation of partial water vaporpressure [hPa], with

[%]rF

. . .610 1 00 0 28 0 04dn dt dp de

î

rF : rel. humidity, E : saturation vapor

100[%]rFe E=

y, p

Demand for 0.1 ppm:ith 0 1°C ith 0 3 hP F ith 10% t ti f i l th t d!t with 0.1°C, p with 0.3 hPa, rF with 10% representative for signal path requested!

1 2 3 4 5 6 7 8

2020

Alignment of 20 units for integral determination of t , p and rF

Principle of geo sensor networks

2121

Crossbow mote with sensor board MT420

Mote with integrated devices:- 1 Dual axis accelerometer- 1 Light sensor www xbow com1 Light sensor- 1 Barometer- 1 Temperature- und humidity measuring unit

1 L t GPS i

www.xbow.com

- 1 Low-cost GPS receiver

- 64 Kbyte EEPROM, TinyOS & ISM radio moduleapprox. 3 x 5 x 4 cm

Light sensor

- Type TAOS TSL2550D

Accelerometer

- Type ADXL202JE (MEMS Technology)

- Range 400–1000 nm- 1 channel for IR- 1 channel visible light

- Range ±2 g - Resolution 2 mg

Rate 60 Hz - 1 channel visible light- 12 bit resolution

„Lux to bit“

- Rate 60 Hz- Nonlinearity 0.2%- “Zero g bias level 2.0 mg/°C from 25°C“

t i

22222222www.analog.com

www.taosinc.com

Meteo sensors at Crossbow board MT420

Barometer

- Type Intersema MS5534AM

Temperature- and humidity meas. unit

- Type Sensirion SHT11- Range 300 – 1100 hPA- Resolution 0,01 hPA- Accuracy ± 1 5% at 25°C“ ca. 7,5 x 5 x 2,5 mm„Accuracy ± 1.5% at 25 C

Temperature:Range 40 to + 80°C

ca. 7,5 x 5 x 2,5 mm

- Range -40 to + 80 C- „Accuracy ± 0.5°C at 25°C“

Humidity: ca. 3 mm

y- Range 0 – 100% Rel. humidity (RH)- Resolution 0,3% RH

Absolute RH accuracy ± 3 5% RH“www.intersema.ch

- „Absolute RH accuracy ± 3.5% RH“

www.sensirion.com

23232323

Location unit at Crossbow board MT420

- GPS 16 channels L1 C/A Code- Rate 4 Hz

www u-blox com- NMEA output, support DGPS & WAAS- External antennaAccuracy Position 2 5m CEP according

www.u-blox.com

- Accuracy Position 2.5m CEPDGPS 2.0m CEP

according μBlox

Ca. 17 x 22.4 x 3 mm

Accuracy specifications of Crossbow meteorological units:t with 0.5 °C, p with 1.5 hPA, rF with 3.5% at 25°C

Low-cost sensors for high accuracy application?Goal: Enhancement by factor 5 by calibration with respect to t and p .

2424

Goal: Enhancement by factor 5 by calibration with respect to t and p .

Recording of reference values for calibration

Quartz thermometer 2804A

t

Universal meteo unit Almemo 2590-3S

t Ft, p, rF

Climate chamber

Thommen-Meteo-Station HM30

t, p, rF Digiquarz

B t 760Barometer 760p

25252525Reference instruments by factor 10 better then crossbow MT 420 test items

Investigations with respect to t

35Temperaturverhalten zu Referenztemperaturen

1

2

T = 1 min.

25

30

2

3

5

6

20

n °C

8

9

10

11

Motes

10

15

Tem

pera

tur i

12

13

14

15

= 0.5 °C= 0.1 °C

5

10 15

18

19

20

Reference

5

071

Referenz

26262626

0 100 200 300 400 500 600-5

Zeit in min

Investigations with respect to p

966

968Druckverhalten zu Referenzdruck

1

2

T = 1 min.

962

964

3

5

6

8

= 0.3 hPA = 1.5 hPA

958

960

ck in

hP

a

9

10

11

12

Reference

954

956

Dru

c

13

14

15

18

950

95219

20

71

Referenz

0 100 200 300 400 500 600948

Zeit in min

2727

Pressure units depict a certain dependence to temperature.

Investigations with respect to rF

90Feuchteverhalten zu Referenzfeuchte

1

2Reference

T = 1 min.

70

80

2

3

5

6

8

Reference

= 10 % = 3.5 %

60

hte

in %

9

10

11

12

40

50Feuc

h 13

14

15

18

19

30

40 19

20

71

Referenz

0 100 200 300 400 500 60020

Zeit in min

Rel humidity at climate chamber varies due to regularization of temperature

2828

Rel. humidity at climate chamber varies due to regularization of temperature.

Noticeable time lag of humidity units.

Calibration results of Crossbow meteo sensors

- Several calibrations performed (t = 0 – 30 °C) to evaluate long term stability.

- 3 of 20 motes with bad and unreliable results.

- Determination of individual response curves using regression analysis.

30

35Korrrigierte Temperaturen zu Referenztemperaturen

1

2

958Korrigierter Druck zu Referenzdruck

1

2

3

90Korrigierte Feuchte zu Referenzfeuchte

1

2

15

20

25

30

empe

ratu

r in

°C

3

5

6

8

9

10

11

12

13

14

956.5

957

957.5

Dru

ck in

hP

a

3

5

6

8

9

10

11

12

13

14 50

60

70

80

Feuc

hte

in %

3

5

6

8

9

10

11

12

13

14

0 100 200 300 400 500 600-5

0

5

10

Zeit in min

T

15

18

19

20

71

Referenz

0 100 200 300 400 500 600955

955.5

956

Zeit in min

15

18

19

20

71

Referenz

0 100 200 300 400 500 60020

30

40

Zeit in min

15

18

19

20

71

Referenz

RMS (simple regression)

cal. temp. vs. true cal. pressure vs. true cal. rel. humidity vs. true

Temperature °C

PressurehPa

Rel. humidity%

0.23 0.11 3.85

29292929

*

* : Improved later by a polynomial regression

Determination of reference distances by ME 5000

Kern Mekometer ME 5000Kern Mekometer ME 5000- Variable frequency 460 – 510 MHz - U = 0.3 m (half wave length)A 0 2 0 2

Set up ofCrossbow motes

3030

- Accuracy: 0.2 mm + 0,2 ppm- Range: 5000 m

Temperaturfeld korrigiert

28

30

T = 1 min.

100018

20

22

24

26

erat

uren

in °

C

distance in m

200

400

600

800

8009001000

10

12

14

16

Tem

pe

Spatial temporal determination of

00100200300400500600700800900000

Entfernung in mZeit in min time in minDruckfeld korrigiert

temperature and pressure field

970

during ME 5000 measurements

800

1000

955

960

965

Dru

ck in

hP

A

distance in m

313131310

200

400

600

01002003004005006007008009001000

950

Entfernung in mtime in min

Uncertainty budget ME 5000

Influence factor XiAssumption for

distribution Influence Measurement uncertainty u

Addi i l l f i l dditi 0 14Additional value of instrument normal additive 0,14 mm

Scale factor of instrument normal proportional 0,50 ppm

Meteorological correction normal proportional 0,14 ppm

Eccentricity of prism rectangle additive 0,02 mmEccentricity of prism g ,

Resolution of instrument rectangle additive 0,02 mm

Inclination of prism rectangle additive 0,03 mm

Misalignment of prism normal additive 0,005 mm

Uncertainties of additional value uc and scale factor us are from the old calibration line.

Uncertainty of scale factor is dominant (and must be improved)!

3232

Influence of meteorological correction relatively low compared to uc and us .

Determination of reference distances by AT901-LR

Absolute distance measurement (ADM), interferometric distance measurement (IFM)

S ifi d t t i t ADMSpecified measurement uncertainty ADM: Distance: U = 10 m + ppm?Modulation frequency: 900 MHzMax. range: 80 m (volume 160 m)

- New and certified instrument- Re-calibrated according to NIST at Geodetic LabRe calibrated according to NIST at Geodetic Lab.

From a Leica brochure:

Leica laser tracker AT901-LR

3333

MPE: Maximum permissible error

Approx.: MPE = U = 2 u

Possible configurations for AT901-LR

Simplified functional model:

2 1Y Y Y

1 2Y Y Y 1 2Y Y Y

Use of intermediate points if Yi > 75 m (out door situation):

3434

Uncertainty budget AT901-LR

Influence factor XiAssumption

for distribution Influence Measurement uncertainty u

Additional value of instrument normal additive 5 µm

Scale factor of instrument normal proportional ???

Meteorological correction normal proportional 0,14 ppm

Production accuracy corner cube:circularity rectangle additive 2 µmcircularity

Production accuracy corner cube:surface rectangle additive 3 µm

Error of alignment rectangle additive 5 µmStability set up rectangle additive 6 µmg

Meteorological conditions are most crucial for measurement uncertainty of AT901-LR.Uncertainty budget much better than ME 5000.

3535

Scale factor of instrument uncertainty actually not considered.Unfavorable error propagation for set ups with intermediate points.

Preliminary results

DistanceEstimated value

[m]

Measurement uncertainty u

ME 5000 [mm]

Measurement uncertainty u

AT901-LR [mm][ ] ME 5000 [mm] AT901-LR [mm]

1-2 18,78??? 0,295 0,016

2-3 82,45??? 0,295 0,023

3 4 146 14??? 0 300 0 0293-4 146,14??? 0,300 0,029

4-5 177,97??? 0,305 0,063

5-6 114,29??? 0,298 0,026

6-7 50,62??? 0,297 0,017

7-8 509,69??? 0,432 0,120

3636

, , ,

3. Conclusion and outlook

- Results obtained so far at the calibration line UniBw Munich are only preliminary.

- Intercomparison programmes with participation of several geodeticinstitutions ongoing and open to any other interesting party.

W k h th fi l lt d ibiliti f th lib ti- Workshop on the final results and possibilities of the new calibration line scheduled for Spring 2011.

- Documentation of complete results according to GUM is obligatory forGeodetic Laboratories – and also any other metrology institutions.

- Principles of GUM are known in Geodesy since years, but a full adaption is still outstanding.

C id ti f ll l t i fl f t th i t ti- Consideration of all relevant influence factors on the interestingmessurands is essential in today’s metrology; see meteorology on EDM’s.

3737

Thank you very much for the invitation!Thank you very much for the invitation!