With the purpose of realizing online monitoring for the unit economic performance

the Elman neural network model was built up for the steam turbine performance calculation.A 330 MW steam turbine unit was taken as an example to examine this model

and the model output was compared with that calculated by the steam turbine matrix equations.The results show that

the established Elman neural network model can reflect the nonlinear relationship between inputs and outputs

and it has lots of advantages such as short period of training

high calculation speed and adjustable inner structure.The average relative error of the results between the Elman neural network model and the matrix equation is within 1%

indicating this Elman neural network model is an effective way for steam turbine performance online calculation.