Project number: PRA497-1920
The Biased position testing for verification of structural softwood project explores the use of biased position testing as a cost-effective method for verifying the quality of structural softwood timber in Australia. This innovative approach can unlock significant cost savings and improved efficiency whilst maintaining rigorous verification of structural softwood timber strength and stiffness properties.
When it comes to quality control and product performance, random selection is traditionally used for sampling structural timber. However, biased selection can also be used. Random selection is like drawing names from a hat; every item has an equal chance to be tested, giving a statistically clear picture of overall quality. Biased selection, however, is based on picking samples (i.e. with known defects) which give specific insights, but not about the whole group.
Biased ratios are essentially formulas that help translate results from that second biased sample into the results expected, as if the whole sample set was randomly tested. So, biased selection testing allows for a reduction in testing costs by utilising a smaller sample size while achieving the same level of confidence in the verification results. This cost-saving measure is particularly advantageous for sawn timber products such as MGP12, MGP10, and F5.
There is no universal Biased Ratio that can be used to directly relate biased position tests to random position tests for structural softwood timber in Australia. Instead, a process was developed to establish appropriate Biased Ratios for quality control testing. This process involves converting biased position test results to equivalent random position test results. The data from the 2023 In-Grade Study, which analysed different products from various mills, was evaluated to assist with transitioning to biased QC testing.
The study found that a biased position testing approach can indeed lead to significant cost savings for mills. For F5 and MGP10, the number of verification test samples can be approximately halved, while for MGP12 and MGP15, the number of samples can be significantly reduced to a minimum of 10. This reduction in sample size can be achieved without compromising the confidence in the verification results.
The report also discusses the proposed method of verification using biased position testing and its integration with the AS/NZS 4490 standard. It provides guidelines for determining the sample size for biased position testing and emphasizes the importance of testing the effectiveness of biased position testing for verification.
One key finding of the study is the derivation of biased strength factors for individual mills and grades. Three different methods were compared for deriving these factors, and it was found that all three methods yielded similar results for strength. This suggests that the chosen method can be tailored to the specific needs of each mill and grade.
The project report also addresses the issue of establishing Biased Ratios for new grading systems and checking these ratios in the Annual Check on Verification. It provides a method for deriving correction equations and simulating their effectiveness.
