Online coal analysis testing compares the laboratory results from a conventional sampling process, to the analyzer results using PGNAA and the current calibration. The objective is to statistically compare the two readings for ash, sulfur, and btu/lb and determine if the calibration linear equation is confirmed.
There are two types of tests that can be performed. A static test of coal or prepared sample tubes is used to confirm factory calibration after installation. This checks the repeatability of the device and software to initial manufacturers specs.
The next type of test is more dynamic. Coal is flowing on the belt in batches, whose mass matches the desired response time, say 5 or 10 minutes. This mass of coal is then sub-sampled by stopped belt increment or a sweep sampler. The analysis of about 30 batches are compared by the Hotelllings t2 statistic, a multivariate student’s t test. The results should be within 10% relative variance compared to the average of the reference, the conventional coal sample. This shows why accurate coal sampling is a must for good analyzer calibration.