The 2003 Australian Wool Innovation On-Farm Fibre Measurement Instrument Evaluation Trial. Part 1: Accuracy and Precision Trials

Abstract

The 2003 Australian Wool Innovation Ltd (AWI) On-Farm Fibre Measurement (OFFM) Instrument Evaluation Trial was a direct outcome of market research commissioned by AWI, which indicated an increasing need by wool producers for an independent and objective assessment of the commercial application of the two current OFFM technologies (OFDA2000 and Sirolan Fleecescan™). AWI established an Expert Advisory Group to provide assistance and guidance with the design and implementation of the trial and the analysis and interpretation of the trial results.

The trial design included sampling the whole fleece (Fleecescan) and two sampling sites (OFDA 2000 for the traditional midside and the less common pinbone) on-farm followed by laboratory testing to Australia/ New Zealand Fleece Testing Standards.

With respect to this report, the key objectives were to determine the Confidence Limits, relevant to the whole fleece rather than simply the sampling site, for the different fleece measurement systems as they are used commercially. The measured characteristics were: Mean Fibre Diameter (MFD); Standard Deviation of fibre Diameter (SDD), Coefficient of Variation of Fibre Diameter (CVD), Comfort Factor (CFR) and Mean Fibre Curvature (MFC).

The Overall Confidence Limits are presented below (Note: for a given parameter, values with the same letter are considered to be statistically equivalent (p>0.05)):

The above Confidence Limits relate to the procedures that were common commercial practice in 2003.  Where required, they can be improved, at an increased cost, by increasing the sampling and or testing regimes.

The key conclusions from this component of the trial were that:

• Midside sampling and laboratory testing gave the best and equivalent confidence limits of ±1.04µm for Laserscan and ±1.05µm OFDA100.; and
• The Fleecescan (±1.17µm) and OFDA2000 midside sampling (±1.24µm) were considered to be equivalent in precision.

The Overall Means and Biases (referenced to the Whole Fleece Average Value) are presented below (Note: for a given parameter values with the same letter are considered to statistically equivalent (p>0.05)):

The different measurement systems produced different biases compared to the whole fleece average. In the case of MFD, pinbone sampling was biased coarse (+0.6µm) and midside sampling was biased fine (-0.3µm). OFDA2000 underestimated the SDD (-0.2µm) and the CVD (-2.3%). The biases in MFC reflect the current lack of a calibration procedure for MFC.

The key conclusions from this component of the trial were that:

• For on-farm MFD measurements, OFDA2000 measured on midsides and Fleecescan measured on fleeces gave equivalent accuracy compared with the whole fleece average.
• For MFD, the pinbone samples measured on-farm or in a laboratory provided significantly higher results (average +0.6 µm) than the whole fleece average. The pinbone as a sampling site does not give as accurate a measure for MFD for the whole fleece. 

Citation

"The 2003 Australian Wool Innovation On-Farm Fibre Measurement Instrument Evaluation Trial. Part 1: Accuracy and Precision Trials", J.W. Marler and P. Baxter, Report CTF01, Commercial Technology Forum, IWTO Evian Meeting, May 2004