In Chapter 4 we examined the considerable range of errors that occur when we make temperature measurements with spectral-band radiation thermometers. In many cases, we showed how to correct for these errors based on additional measurements or estimates of the quantities that influence the temperature determination. In other cases we sought to minimize the influence.
Ultimately, in any temperature assessment we arrive at a value that forms the basis of a decision. In a reformer, the decision may be to increase the firing in order to raise the tube temperature; it may be to decrease the firing or even shut down the reformer; or it may be to do nothing at all. Whatever decision we make, there is some risk that we do the wrong thing. The risk arises because we cannot be sure that we have accurately estimated all the errors in our temperature measurements. As a consequence, in order to make informed decisions, i.e., to quantify the risk, we need to measure the quality of our temperature measurements. This measure is known as the uncertainty.
Just as you wouldn't buy a used car without first seeing it or knowing something about it, you shouldn't accept the result of a measurement without an associated uncertainty. Uncertainty analysis is a complicated subject, involving at times serious mathematics, but is necessary for anyone equally serious about the results of their measurements. In this chapter we don't attempt to cover all aspects of uncertainty analysis, but provide enough background to enable you to assess the quality of your temperature measurements.
Online access to SPIE eBooks is limited to subscribing institutions.