| With the rapid development of science and technology,the demand for vibration environment simulation is increasing in the fields such as aerospace,marine,road transport,earthquake prevention when considering the reliability of equipment or structural components working in vibrating environment.The vibration environment in natural world is more common with random vibration.A large number of observed phenomena indicate that the characteristics of random vibration can be described by the power spectral density(PSD),which reflects the distribution of the signal power in the frequency domain.Electro-hydraulic servo system can be widely used to simulate random vibration environment because of its advantages of strong bearing capacity,high control precision,fast response speed and so on.Electro-hydraulic servo system has a wide range of non-linear factors such as pressure flow characteristics,servo valve control dead zone and friction,which reduces the performance of the linear control system.At the same time,due to the influence of the dynamic characteristics of electro-hydraulic servo valves,hydraulic cylinders and other components,the bandwidth of the system is limited.Besides,the system is susceptible to coupling vibration because of the influence of the load environment characteristics.The above factors explore the common reasons that the system output is not good enough to follow the system input.Therefore,it is necessary to carry out the research on the wide-band random vibration control of electro-hydraulic servo system in order to simulate the random vibration environment.This paper studies the eletro-hydraulic servo systems with random vibration control.System linear modeling based on transfer function and system nonlinear modeling based on physical relationship are separately developed to accurately reproduce the reference acceleration PSD with 5?500Hz bandwidth range.Although the transfer function model is more commonly used,it can not accurately represent the influence of the system's nonlinear characteristics,while the physical relationship model can fully consider the nonlinear factors,but the solution is more difficult.For the above two models,the constant gain state feedback of the linear model is designed by linear quadratic regulator(LQR),and the variable gain state feedback of the nonlinear model is obtained by solving the state dependent Riccati equation(SDRE)online.The design system has good frequency response characteristics and stability.Meanwhile,in order to compensate for the OdB gain fluctuation and bandwidth shortage of the system frequency response function,the input signal correction is completed in the frequency domain in conjunction with the PSD equalization control technology.For the identification of the system frequency response function during the equalization,the linear model uses an LMS autostepping algorithm with a smooth recognition effort,and the nonlinear model uses a more stable HI method for identification.Finally,random vibration control simulations for different models under different equilibrium conditions are performed.The results show that(1)The reference PSD can be accurately reproduced by the response PSD with the increase of the number of equilibrium.(2)The number of equalization on response PSD reaching stable fluctuations can be reduced when increasing the equalization feedback gain.(3)The appropriate equalization feedback gain is between 0.2 and 0.5,and the steady fluctuation range of PSD is ± 1 dB.(4)For the linear equilibrium form,nonlinear factors of the system can be offset when increasing the equalization feedback gain.(5)For the logarithmic equilibrium form,the number of steady-state fluctuations on PSD in the nonlinear model is more than that in the linear model under the condition of the same feedback gain.The main contents of this paper are as follows:In Chapter 1,the electro hydraulic servo vibration system is introduced and then the reference PSD reproduction of random vibration control is derived.The research status of servo control and random vibration control at home and abroad is summarized.Finally,the significance and main research contents are presented.In Chapter 2,the scheme of electro hydraulic servo system with random vibration control is designed.According to the design parameter,linear modeling based on transfer function and nonlinear modeling based on physical relationship are performed on the system.Finally,model is verified through the dynamic characteristic experiment.In Chapter 3,the research on system state feedback is carried out.State feedback of linear system is designed by means of LQR theory,and state feedback of nonlinear system is gained by solving SDRE online.Good frequency response characteristics and stability can be obtained for the design system,which reducing the pressure of PSD equilibrium during the process of random vibration control.In Chapter 4,the research on system PSD equalization is introduced.The input signal is modified in the frequency domain.For the identification of the frequency response function in the equalization process,system identification simulation based on auto-step algorithm of LMS is adopted to gain smooth and steady identification effect.Estimation method of AR spectral model is used to estimate PSD because of high precision.Finally,random signal of time domain is generated by randomization both in frequency domain and time domain.In Chapter 5,random vibration control simulation is conducted.Random vibration control system is designed in Simulink.The random vibration control of the linear model and the nonlinear model is compared with different equalization feedback gains and equalization forms.The results show that increasing the feedback gain can reduce the number of equilibrium,and equalization form of linearity and nonlinearity can both reproduce the reference PSD well.Control effect of acceleration PSD reproduction is verified by the hardware-in-loop simulation in linear system model and monitoring random vibration state variable of hydraulics model in nonlinear system model.In Chapter 6,the main research work and conclusions of this paper are summarized,and future work and direction are also discussed and suggested. |
PDF Full Text Request