We propose a runnable feedforward neural network dedicated to detonation Chapman-Jouguet velocity calculation. You can assess a demo below on the cloud. It runs on organic formulation up to four components, energetics or not.
How to make a calculation?
The interface below allows to enter HE formulations up to 4 components. The first 3 lines can be filled with data from TNT,RDX,NQ and polyethylen. The last line is free if you have the a molecule’s formula and the Enthalpy of formation (in kcal/mol).
The amount of each component is in the last column (%). For instance, the data for 40% TNT, 50%RDX and 10% Polyethylen are
A token is required because I want to limit the access to professionals to avoid overload on my machine. You can ask for a token in the contact page or you can run a lighter/slower free instance limited to TNT/RDX formulae.
The trained neural network is a feedforward version of 50 neurones.
The results are displayed as a table of detonation velocities vs density and a graph showing them among 4 experimental data for pure HE:
Run the detonics neural network
The boot may require a while.. So please wait and refresh if it doesn’t start..