ACCURACY

Accuracy is the life of analysis.

Cloudland Deconvolution Software can use GC data only to compute out some results not only deemed impossible, but also with their accuracies seemed unbelievable (especially for version 4.0). See some results listed on the table below.

REPRODUCIBILITY (RP)

ItemAKIIBP10%50%90%FBPVPAPI
Ver 2.00.272.72.72.53.54.8N/AN/A
Ver 4.00.181.41.31.41.63.00.130.26
ASTM RP0.67.34.66.610130.42-0.510.5
ASTM MethodD2699
D2700		D86D86D86D86D86D5191D287

Note: The ASTM RP is an average value in a range of the specification.

The RP on the table above were computed between the GC results and those by listed ASTM methods for more than a thousand of samples. As we know, most software used for petroleum testing computes results without comparing with ASTM results or with a very limit number of results. Look at the reproducibility of 1.4, 1.3, 1.4, 1.6, and 3.0 for IBP, 10%, 50%, 90%, and FBP, which are well below the corresponding ASTM RP and can be deemed close to a limit of accuracy by ASTM D86. The RP of 0.26 for API may be improved.

As known there is some software available on market to compute out distillation results using GC data. But some analysts of petroleum testing labs have experienced these results are not accurate especially for high FBP, which may be 40 – 50 ºF lower than the manual results for diesel contaminated gasoline.

As known, due to the technological development the repeatability by GC is great. When a sample is run several times by a GC with software, usually the range of the results is only 1 ºF for IBP, 10%, 50%, 90%, and FBP. Its repeatability is nearly perfect. However, this does not assure the accuracy by GC. We must compare the results by GC with those by other accurate methods.