Multivariable large delay system neural network multi-step predictive dynamic modeling research

The paper mainly introduce the dynamic modeling method of multi-step predictive neural network in the large delay system. First proposed the improved LMBP algorithm, and then discussed the algorithm and structure of neural network and dynamic response characteristics of large delay system in detail. Finally based on single variable delay system modeling frame presented multivariable large delay system´s neural network multi-step predictive modeling new algorithm. And have done simulation to multi-step predictive modeling based on the data of the water purification coagulant sedimentation delay system.