Dynamic modeling algorithm based on time difference and fuzzy-tree model

In order to improve the dynamic modeling performance of fuzzy tree (FT) model, a dynamic modeling algorithm based on time difference (TD) method and FT model is proposed. TD method can indirectly utilize the historical information of system without increasing the input dimension of FT model, so as to improve the accuracy of dynamic modeling. Firstly, the influence of TD method on signal-to-noise ratio (SNR) was studied by theoretical derivation and simulation analysis. Then TD-FT algorithm was applied to Mackey-Glass chaotic time series modeling and dynamic modeling of denitrification system in coal-fired power plant under variable-load condition. The simulation results show that the modeling accuracy of TD-FT algorithm is higher for chaotic time series modeling, but it is susceptible to noise. For the dynamic modeling of denitrification system, the input dimension of TD-FT algorithm is lower, the modeling accuracy is higher and it is insensitive to noise.