The design of ferritic steel welding alloys to fit the ever expected properties of newly evolved steels is not a very easy
task. It is traditionally attained by experimental trial and error, changing compositions and welding conditions until a sufficient
result is established.
Savings in the economy and time might be achieved if the trial process could be minimised. The present work outlines the use of
an artificial neural network to model the charpy toughness of ferritic steel weld deposits from their chemical compositions,
welding conditions and heat treatments.
The development of the General regression neural network (GRNN) models is explained, as is the confirmation of their
metallurgical principles and precision
