Study On Modeling And Control Technology Of Air-Magnetic Active-Passive Hybrid Vibration Isolation For Floating Raft System

It has the extremely vital significance to the Chinese navy to improve the performance of naval vessels acoustic stealthy.Vibration isolation technology has become a key technique in the field of vibration and noise reduction of ships and widely used to reduce vibration transmission to the base.As a passive vibration isolator,air spring has many advantages such as strong bearing capacity,low natural frequency,easy installation height adjustment,excellent high-frequency vibration isolation performance,which has been gradually applied to the field of vibration isolation of marine main engine,and its vibration isolation performance have been greatly improved.Therefore,the air spring is applied to the floating raft isolation which the advantages of"mass effect","mixing effect"and"tuning effect",so as to improve the high frequency vibration isolation effect.However,the inherent shortcomings of poor low-frequency vibration isolation performance and amplification for system resonant frequency cannot be completely overcome by air spring just as other passive vibration isolators.Active unit(such as:active vibration isolator,etc.)is introduced to solve the problem of low frequency and resonant frequency near the vibration and noise reduction by dynamically adjusting the system support characteristics of the parameters.Thus,the theory,methods and techniques between active and passive vibration isolation are combined to effectively isolate the submarine mechanical vibration,reduce submarine mechanical noise and improve the submarine sound stealth ability.The principle and structure of the air spring and the electromagnetic actuator for the experimental platform are roughly introduced in the first part.Then,the relationships of height-bearing capacity-stiffness for air spring and voltage-current-electromagnetic force/acceleration are studied by theoretical analysis and experimental methods.The results verify the characteristics of the airbag isolator and the electromagnetic actuator and meet the vibration isolation reqirments of floating raft system.It proposes to introduce air springs and electromagnetic actuators to construct a magnetic-air active-passive hybrid single-layer vibration isolation system,which will achieve the full frequency domain isolation of the marine main engine eventually.The dynamic model of single-layer vibration isolation system is established.The dynamic model of flexible floating raft system without active isolator is established with power flow as evaluation criterion.The influence factors on the output power flow of the flexible raft system are analyzed by using ADAMS analyzing software.Influence factors include rigid/flexible raft,the motor mass,the stiffness and damping of the isolator,the number of the isolators and the arrangement of the isolators and so on.According to the results,the optimization influence factors are used for the desigen of floating raft experimental platform.OKID(Kalman filter system identification)isintroduced to construct a mathematical model of the floating raft system directly from input-output measurements in the presence of noise.The core equations of OKID are derived and the system identification algorithms based on OKID are formulated.The SISO(single-input single-output)and MIMO(multi-input multiple-output)models of floating raft platform are obtained by OKID and the robustness of OKID is superior to PEM.The results are verified by identified models from different proportions of noise in input-output data.The accuracy,stability and transient response of identified model from OKID/ERA are tested by fitting degree of frequency response function,zero pole diagram and impulse response graph respectively.Then,the high order model indentified from OKID/DP for MIMO floating raft system is used to reduce the order to obtain the low order mathematical model by using unit impulse response as evaluation criterion.The comparison between reduced model and full model is conducted in the impulse response and frequency response function to verify the feasibility of the approach.The reaserch of system identification for the excitation source difficult to measure is carried out in the floating raft system.The excitation source consisting of a variety of motors for the floating raft system can be considered a white noise signal.The stochastic problem can be translated into deterministic problem by using O3KID.The system matrices can be obtained by deterministic identification methods(ERA?ERA-DC?DPi?DIi).Futher,the estimation of natural frequency and damping factor can be obtained from identified system matrices.This identification method verifies the feasibility of the motor as an input signal for system identification,which provides a new way and field of view for obtaining the parameters of floating raft system without input data for system identification.The normalized FXLMS algorithm is one of feedforward control algorithm and applied to SISO and MIMO control with single excitation source.The designed normalized FXLMS algorithm is proved to have a better convergence rate and vibration isolation effect than fixed step size FXLMS algorithm for the single frequency disturbance.In order to solve the multi-frequency perturbation problem for Futher,the H_?controllers for SISO and MIMO are designed to suppress the multi-frequency perturbation for the floating raft system.The identified models from OKID and H_?controllers are taken into the Simulink software to get the perfect performance of suppressing the disturbance main frequencies.The active-passive flexible floating raft platform with electromagnetic actuator and air spring is established,the identified model and controller is applied to the active vibration isolation experiments for single motor and sveral motors.The vibration level difference of acceleration is selected as evaluation criterion,the comparison experiments between designed controller and fuzzy controller are conducted for the floating raft platform.The results present that the performance of the new solution(normalized FXLMS algorithm with OKID model)is compared with previous work based on a fuzzy controller and a dynamic model identified via OKID.The new H_?/OKID solution proved to outperform the previously developed fuzzy control/solution.Both simulated and experimental results confirm the validity and the benefits of the approach.