6.9.
Conducting a meaningful test with the purpose of assessing the integrity or functional capability of an item requires that the conditions existing for this item in the plant during an earthquake are correctly or conservatively simulated or that any departure from these conditions will not significantly influence the results. Among these conditions, the most important are:
6.10.
In a test, the item should be subjected to conservatively derived test conditions in order to produce effects at least as severe as those of the design basis seismic event concurrently with other applicable operating or design conditions. Deviations should be evaluated on a case by case basis.
6.11.
Tests may be conducted either on the site or in a laboratory. On-site testing of equipment and components should be limited to a few qualification aspects since it proves expensive and often conflicts with accessibility. It represents a reliable strategy for the evaluation of real support, boundary conditions and ageing effects. On-site testing of structures is often the only means of capturing the actual properties of materials, global structural seismic behaviour and the effects of soil–structure interactions, and it should be carried out whenever feasible to provide results as a reference for similar structures.
6.12.
The input motion should be consistent with the seismic categorization of the item to be tested, to provide confidence in the required safety margin.
6.13.
The functional testing and integrity testing of complex items such as control panels containing many different devices should be performed either on the prototype of the item itself or on individual devices with the seismic test input scaled to allow for the location and attachment of the device within the item or on the item (via the in-cabinet transfer function).
6.14.
Account should be taken of ageing effects, which may cause deterioration or otherwise alter the characteristics of the item during its service life.
6.15.
Seismic tests may be performed on the item itself or on a full scale model or, where appropriate, on reduced scale models. However, for qualification purposes, the component itself or a full scale model should be tested without any simplification; if there is no other practical alternative, the careful use of a reduced scale model may be permitted for qualification purposes. Such tests include:
6.16.
When reduced scale testing is performed, the setting of similarity criteria associated with indirect methods of seismic qualification should be considered.
6.17.
Any test result should be accompanied by a detailed evaluation of the reliability of the measurements (usually obtained by means of statistical analysis), evaluation of the signal to noise ratio and sensitivity evaluations, with a clear identification of the numerical (from data processing) and physical (from assumptions in the modelling) sources of uncertainties.