Dual-rate Sampled Hammerstein Nonlinear System Identification

Nonlinear system identification has essential practical significance.But in actual industrial processes,due to factors such as hardware limitation,environmental impact and economic conditions,system samples at different sampling rates,the data obtained is often dual-rate sampled data.It is hard to identify by dual-rate sampled data directly.As the complexity of nonlinear system,there is still none universal model to describe nonlinear system.In recent years,many researchers use block-oriented models to describe and identify the nonlinear system.Among the block-oriented models,one model called Hammerstein which has become one of the hot topics in research.Thus,the identification of the Hammerstein system based on the dual-rate sampled data has essential theoretical and engineering practical effect.Aiming at the Hammerstein system,this thesis study the online identification algorithms with the dual-rate sampled data.The main work is summarized as follows:1.Aiming at Hammerstein OE(output error)system based on dual-rate sampling,a Hammerstein OE model with nonlinear part shown as dead-zone nonlinearity is established.Based on the coupling between parameter and the coupling between the parameter vector and the information vector in dead-zone nonlinearity,the recursive least squares algorithm based on auxiliary model is proposed;a Hammerstein OE model with nonlinear part shown as saturation nonlinearity is established.Considering the existence of the coupling between parameter and coupling between the parameter vector and the information vector in saturation nonlinearity,the recursive least squares algorithm based on auxiliary model is proposed to estimate parameter.Based on the system identification of two simulation examples,the effectiveness of the two proposed algorithms is proved.2.Aiming at Hammerstein OE model based on dead-zone nonlinearity,in order to reduce the influence of noise and improve the anti-interference ability during system identification,combine binary pseudo-random sequence signal with random signal based on dual-rate sampling.Separating nonlinear part and linear part from three kinds of states perspective,the influence from coupling is weaken.Based on the above work,the recursive least squares method based on auxiliary model with combined signal is proposed.The effectiveness of the proposed algorithm is verified through simulation.3.Aiming at Hammerstein OE model based on saturation nonlinearity,combine binary pseudo-random sequence with random signal based on dual-rate sampling.By processing input signal,nonlinear part and linear part are separated from four kinds of states perspective.Based on the signal processing,the recursive least squares algorithm based on auxiliary model with combined signal which estimate parameter by three stages is proposed.The effectiveness of the proposed algorithm is proved by simulation,and the anti-interference ability is improved.4.Aiming at the main exhaust valve and choke finger of the 2.4m wind tunnel system,the corresponding Hammerstein OE models of the main exhaust valve and the choke finger are established,respectively.Based on dual-rate sampled data,considering to apply the recursive least squares method based on the auxiliary model with combined signal to identify the Hammerstein OE models with dead-zone nonlinearity to prove the effectiveness of algorithm.After proposing the above algorithms and confirming them by corresponding simulations,thesis summarizes all of the work,and finally sorts out the directions worth studying and discussing in the future,and make prospect.