Research Of Fuzzy-NN (Neural Network) Control On Structural Control

Ship-lifting whose main function is to lift the coming-and-going ships vertically to cut short the dam crossing time is a kind of important implements in comparison with the multilevel ship-strobe. The structure will bring us prominent economic profits and social benefits. However, due to the great pliability of the structure itself, the lateral stiffness of the cylindric structure at the underside of ship-lifting structure is a few hundred times as much as that of the single stratum workshop pillar at the upside of ship-lifting structure under the earthquake. The huge stiffness mutation will give rise to a violent whiplash effect on the single stratum workshop. Therefore, reducing whiplash effect is becoming a very critical issue at present This subject stems from " intellectual control of structure vibration of civil engineering " of the fund key item of national natural science (50038010) .According to the whiplash effect of the ship-lifting structure, analysis in theory simulation and experimental investigation on real-time simulation system are employed to carry out the research. In this paper the fuzzy control theory and neural network algorism are introduced into structural control field and fuzzy-NN (neural network) control algorism is proposed, which adopted fuzzy control algorism to control MR dampers, and a inverse neural network model is established for MR dampers in the control process. It can obtain continuous control for the system and improve the control effect of the system.Based on the three Gorges ship-lifting structure and MR damper model, the intelligent control simulational system on fuzzy-NN control algorism is designed. On the basis of the simulational research the intelligent control real-time simulational system for the three Gorge ship-lifting structure based on dSPACE is developed. The system adopted RCP (Rapid Control Prototyping) technology to control the Three Gorge ship-lifting structure model in real time.The results on real-time simulational experiment verify that the major advantages of the proposed fuzzy-NN control algorism include the quick and simple control without requiring exact models. The intelligent control algorism can control the MR dampers to obtain the perfect control effect. Comparing simulational results and experimental results, the validity of the numerical simulational method is good and the control effect of fuzzy-NN control algorism can also be proved.