Using BP neural network

and based on vast amount of coal quality analysis data

as well as thermobalance test data

a BP neural network model has been established under three situations

namely proximate analysis

relation of calorific value with ignition temperature

and with stability of ignition indices. Results show that the very high accuracy of prediction can be reached by using BP neural network

at the same time

the correlation between input and output can be compared. The obtained results of prediction can provide basis for purchasing coal

as well as provide reference for boiler operation.