New method for determination of heat release rate has been developed at the university from Slovakia and now is ready for license agreement.

Heat release rate is fundamental fire characteristic of substances, materials and products. Three methods for heat release rate were known until now. The first is based on measure of temperature of combustion products. Main disadvantage of this method is complicated calibration (given by dependence of heat capacity on temperature). The second method is based on measure of mass loss and multiplication (of obtained value) by effective heat release rate. Main disadvantage of this method is chance of the effective heat of combustion (of most organic polymers) with time (volatile compounds burn first and carbon residue subsequently). The third method is based on measure oxygen consumption (during perfect combustion 13.1 MJ of heat is released per 1 kg oxygen consumed). Main disadvantage of this method is the need for a desiccant of combustion products (which often contains a carcinogenic moisture indicator). A common disadvantage of all described methods is that they are able to be used only in the laboratory (not during real fire). The research team from technical university in Bratislava with the background in material technology has created a method that measures heat release rates of substances, materials and products from photographs of their flames. The sample with the known reaction to fire class is tested and photographs of its flame are taken during this test. Taken photos are used for (convolutional) neural network training in the first step. After training the (convolutional) neural network is used for checking declared reaction to fire test of tested sample (from photos of its flame taken during second step) in the second step. The institution is seeking an industrial partner for licensing for the further development and commercionalization of mentioned technology.