ASTM D6621 Performance Testing of Process Analyzers for Aromatic Hydrocarbon Materials
9. Analyzer Parameter Adjustment
9.1 Adjust analyzer response parameters only when the analyzer's control sample SPC chart indicates one or more rule violations, and after investigation, causes have been found and eliminated. Then, if in the opinion of the operator the response accuracy of the analyzer has changed, parameter adjustment may be performed.
9.2 Methods for adjusting analyzer parameters depend on the type of analyzer, and are specifically available from the analyzer vendor. For this reason, analyzer parameter adjustment guidelines are beyond the scope of this practice.
9.2.1 It may be necessary to revalidate the analyzer instead of adjusting its parameters, depending on the reasons for analyzer accuracy changes. If the basic setup or operation of the analyzer have changed, it should be revalidated.
9.2.2 A rule violation is NOT an automatic signal for parameter adjustment! Most rule violations will not signal a need for changing the analyzer's parameters. Rather, they are indicators of change that must be investigated and used, with the operator's knowledge of the process and analyzer, to decide further action. Doing nothing, adjusting analyzer parameters, performing analyzer maintenance, or revalidation are options.
9.3 Whenever analyzer parameters are changed, all information relating to it must be recorded in a log book or computer file for the analyzer. Records should include current analyzer settings, changes in analyzer settings, maintenance service performed, any causes of change and repair, or any other relevant information.
9.4 A spot sample, an identical material analyzed both in the analyzer and in the laboratory by the primary test method, may be analyzed at any time to check analyzer accuracy. However, these results generally should not be used to determine the need for analyzer parameter adjustment. Rather, if the laboratory and analyzer results differ by more than test method reproducibility, a subsequent analyzer performance test should be performed. If that run produces a control chart rule violation, the cause is identified and corrected. If then the operator feels that analyzer accuracy has changed, only then should analyzer parameter changes be considered (see 9.2).
9.5 If a statistical bias between laboratory and process analyzer results has been previously well-characterized during analyzer validation, and it has been monitored continuously by SPC (for example, from an SPC chart of the difference between the laboratory and analyzer on the check sample), a spot sample result may be used to trigger analyzer parameter changes. (Warning - the laboratory and process analyzer must be analyzing the same material as the spot sample). Under these conditions, if the difference between the spot sample and laboratory results are statistically different (for example, differ by more than test method reproducibility), then the operator may consider changing analyzer parameters appropriately (see 9.2).
10. Keywords
10.1 analyzer; on-line; performance testing; process stream analyzer
