Back Propagation Neural Networks of Base Bleed Propellant Burning Rate under High Pressure Condition

To investigate the combustion characteristics of base bleed propellant moving with the projectile in the gun bore, the closed bomb semi-physical and experimental simulation technology was employed. The combustion property under the condition of simulative high pressure in the gun bore was studied. Two average pressure-time curves corrected by heat loss were obtained under different charge density of the closed bomb. The correlation data between burning rate and pressure (8－150 MPa) were processed by smoothing, filtering and data transformation. The burning rate model adopting back propagation (BP) neural networks describing, which is suitable for high combustion pressure condition, was built based on training data samples. Comparing with the exponential burning rate model, the BP neural networks burning rate model of base bleed propellant has higher fitting precision and stronger robust.