EU LARGE-SCALE REFERENCE FIRE TESTS 

Introduction
The European Commission has declared that 'real fire experience' should be the basis for adopting new fire tests to underpin the EU Fire Classification systems. The best way to test a product or system to reflect 'real fire experience', and collect relevant data, is to use an appropriate large-scale fire test.

However large-scale fire tests are usually expensive to perform and not in widespread use at all test laboratories - so a system using less expensive small-scale tests is to be preferred. But such a system must incorporate the ability to favourably compare ranking data from the selected small-scale tests, to the ranking data obtained from the larger 'reference' test, so that a form of correlation can be created between the two test formats, large and small scale.

In practice this meant that it was necessary to select large-scale fire tests that could act as EU 'Reference Tests' for fire performance. These 'reference tests' would provide information typical of a real fire experience. The test data could be used when a referral /appeal process is deemed necessary to compare information obtained in smaller-scale fire tests. For example the ISO 9705 Room Corner reference test [RCT] has been chosen as the reference test for the Euroclass classification system for reaction to fire performance of room lining products such as plywood, MDF, insulation materials and plasterboard products [The RCT test will be standardised as prEN 14390 - Full scale room test for surface products].

Sometimes, correlations between large and small test data can be difficult to establish since full scale building systems may be more susceptible to fire attack than small scale tests might indicate. For example, fire performance at joints and fixings is not challenged in a small-scale surface spread of flame test.

In the following illustration, compare the fire growth rates FIGRA in the two tests, ISO 9705 Room Corner reference test [RCT] and the smaller SBI test for lining products identified in the diagram as products a to j. The results are quite different, but follow a correlated trend for Euroclasses A1, A2 to F. Other small-scale tests may be needed in the classification process to tighten any correlation.