Version 1003
State of the art of indoor calibration of pyranometers and pyrheliometers
2Indoor calibration
Main points
• Most field pyranometers are calibrated indoors
• Many procedures for indoor calibration
• Not all optimally connected to ISO 98-3 GUM
• Industry requires straightforward approach
3
Industry
• Meteorology - Solar renewable energy • Site assessment• Installation performance• Professionalisation / IEC
4
Future
• A few high accuracy outdoor calibrations
• A lot of indoor facilities• Accredited labs
5
Conclusion
• Points for discussion• Normal Incidence NI calibration is
preferred (Diffuse Sphere Source DSS not)
• Uncertainty & accuracy of reference can be optimised
• Indoor calibration complies with GUM• Pyrheliometer indoor calibration must
be allowed by ISO
6
Myself
• Kees VAN DEN BOS• Director / owner Hukseflux Thermal
Sensors• Last 20 years sensor design
7
Hukseflux DR01 pyrheliometer
• Founded 1993
• Thermal sensors
• 15 employees
• 5 radiometry
8
9Hukseflux 2010
10Reolith thermal properties on moon rover
11Snow thermal conductivity
My interest
• Hukseflux company cannot work with outdoor calibration
• Our customers want a understandable accuracy statement
• Feedback• More questions than answers
12
Background
• Most pyranometers and pyrheliometers have indoor calibration
• Exception: highest accuracy (BSRN, outdoor)
• Exceptions on national level: Japan, China, … (outdoor)
13
Background
• Cost, time, weather; outdoor calibration is unacceptable to industry
•DISADVANTAGE: Indoor methods only work with reference type = field type
14
Present status (excerpt)
• Eppley, US Weather Bureau: indoor integrating diffuse source
• Kipp, Hukseflux: indoor normal incidence
• EKO: outdoor tracker with collimation tube
• KNMI: indoor (network) and outdoor (BSRN)
15
16ISO 9060
17ISO 9060
Background
• Measurement uncertianty is a function of:
• Characterisation / class • Calibration (+characterisation and
class)• Measurement & maintenance
conditions• Environmental conditions
(+characterisation and class)
18
Background
• Indoor calibration covered by ISO 9847
• Present ASME: “Indoor Transfer of Calibration from Reference to Field Pyranometers”
19
20ISO 9846
21ISO 9847 also indoor
22ISO 98-3 GUM
Hierarchy of Traceability
• A: Reference calibration (uncertainty)• B: Correction of reference to indoor
conditions (uncertainty)• C: Indoor calibration of field
instrument (uncertainty)• Indoor calibration uncertainty
estimate (A+B+C)• Field measurement uncertainty
estimate
23
Hierarchy of Traceability
• A: Reference calibration (uncertainty)• B: Correction of reference to indoor
conditions (uncertainty)• C: Indoor calibration of field
instrument (uncertainty)• Indoor calibration uncertainty
estimate (A+B+C)
24
25ISO 98-3 GUM
26Hierarchy of traceability
27
28Indoor calibration Normal Incidence NI
29ISO 98-3 GUM
30
Hierarchy of Traceability
• KNMI TR 235 "uncertainty in pyranometer and pyrheliometer measurements at KNMI in De Bilt".
31
Hierarchy of Traceability
• A: Reference calibration (uncertainty)• B: Correction of reference to indoor
conditions (uncertainty)• C: Indoor calibration of field
instrument (uncertainty)• Indoor calibration uncertainty
estimate (A+B+C)• Field measurement uncertainty
estimate
32
33
34ISO 98-3 GUM
NI Hierarchy of Traceability
• A: Reference calibration (uncertainty) (conditions and class)
• B: Correction of reference to indoor conditions (uncertainty)
• C: Indoor calibration of field instrument (uncertainty)
• Indoor calibration uncertainty estimate (A+B+C)
• Field measurement uncertainty estimate (conditions & class)
35
Strange…
• Errors in reference calibration re-appear in measurement errors
• Counted double• At least systematic errors (Zero offset
A and directional errors) can be avoided.
36
One step back
• Calibration with restricted conditions results in lower uncertainty
• See yesterday’s presentation by Ibrahim Reda
37
One step back
• Present reference works well if calibrated pyranometers are used:
• Outdoor / unventilated• At same latitude
38
39
One step back
• Present approach does NOT work well calibrated if instruments are used:
• As indoor reference• At different latitudes• Ventilated
40
Typical secondary standard calibration• Irradiance 800 W/m2
• 40 to 60 degrees angle of incidence, + / - 30 degrees azimuth• Zero offset A: -9 +/- 3 W/m2 (larger
than ISO9060)• Directional: +/- 10 W/m2 @ 1000
W/m2 , now estimated +/- 5 W/m2
41
Typical calibration
• PMOD specified uncertainty +/- 1.3%• Systematic error -1%? Type B.
42
NI reference improved
• Restricted conditions• Zero offset A: -9 +/- 3 W/m2 (larger
than ISO9060)• Directional: +/- 10 W/m2
• Solution 1: ventilation• Solution 2: single angle of incidence
43
For consideration
• Japanese collimated tube with tilt correction and ventilation
• Tilted sun-shade method
44
45
Diffuse Sphere Source DSS
• Uniformity of sphere top-edge (experimental -13%)
• Weighing for non uniform source requires weighing of reference with source
• Diffuse sphere: weighing requires weiging of field instrument with source. Complicated!
• Normal incidence: weighing of field instrument is not necessary
46
DSS Hierarchy of Traceability
• A: Reference calibration (uncertainty) (conditions and class)
• B: Correction of reference to indoor conditions (uncertainty)
• C: Indoor calibration of field instrument (uncertainty) (conditions and class)
47
DSS Hierarchy of Traceability
• Indoor calibration uncertainty estimate (A+B+C)
• Field measurement uncertainty estimate (conditions & class)
• Additional uncertainty under C compared to NI calibration
• Bottom line: DSS has less restricted conditions than NI
48
49
Conclusion
• Indoor calibration offers only acceptable solution for manufacturers and “general users” in solar industry
• Indoor calibration fits within ISO 98-3 GUM
• detailed statements about field measurement still need to be agreed upon
50
Conclusion
• Indoor calibration: Normal Incidence calibration is preferred (Diffuse Sphere Source is not)
• Accuracy and precision of reference can be optimised to serve as indoor calibration reference (restricted: single angle, ventilated)
• Pyrheliometer indoor calibration must be added /allowed by ISO
51