| Rubs between rotor and stator is a kind of common faults occurred in large capacity and high speed rotating machineries, such as turbine generator sets. To satisfy the requirements of high speed and high efficiency on modern machineries, the gaps between rotor and stator become smaller and smaller, which leads to more rubs. Rub is a kind of very complicated dynamic behavior, the rubbing process is strong nonlinearity and non-smooth, which will cause some faults, such as seal wear, blade bowing or breaking, shaft thermal bending, rotor system being unstable, even leads to serious running accidents. It will help us to reveal the running rules of rotor system and improve the dynamic characteristics by building dynamic models for rubbing rotor system and analyzing their vibration characteristics, bifurcation and chaos. The vibration signals of rubbing system are unstable, nonlinear and time-varying. It is the key for fault diagnosis to analyze the rubbing vibration signals rationally, extract rubbing fault feature efficiently and locate the rubbing positon accurately, which is very important for rotor system running safely and reliable.Models of rubs are analyzed. Dynamic models and equations of rubbing system are built. Using numerical methods, the dynamic characteristics of rubbing rotor system is analyzed. Rubbing fault feature extraction and axial location of radial rubbing are studied. The main research work and results are as followings.(1) Radial rubs between rotor and stator are analyzed. Radial rubbing models are classified into physical models rubbing force models and system dynamic models according to system structure, rubbing force style and system dynamic characters. There are rigid support models and elastic support models in physical models; linear spring models, non-linear models based on the Hertz theory and non-linear spring-damping models in normal rubbing force models based on dynamic contacting theories; linear coulomb force models, non-linear coulomb models, non-linear Hertz models, Mane models and LuGre models in tangential friction force models; and constraint differential models, piecewise smooth models in dynamic models of rubbing rotor system. The theories, expressions and applications of above models are analyzed, which will provide theoretical basis for dynamic analysis in rubbing rotor systems.(2) The stress distribution of journals in sliding bearings is analyzed. Dynamic equations are set up for rotor system with rubs considered non-linear oil-film force and non-linear friction force, in which capone non-linear oil-film force model is used. The numerical analysis results shows that the velocity influence coefficient in tangential friction force is less effect on system when the rotational speed is lower, but with its increasing, system vibration amplitude will increase. When the rotational speed reaches to the system critical speed, the velocity influence coefficient will influence the system dynamic characteristics most apparently. At that time, a very smaller velocity influence coefficient will cause system unstable.(3) The characteristics of rubbing rotor system with asymmetric stiffness at the same time and respectively in shaft and supports are analyzed. The dynamic equations are set up considering the non-linear oil-force and asymmetric stiffness. The numerical analysis results shows that the asymmetric stiffness in two main directions of shaft and supports will affect the stability of the rotor system, and the stiffness ratio in two main directions of the shaft and supports will affect the severity of rubbing. Therefore, when designing rotor systems, stiffness differences in two main directions should try to be reduced.(4) Two-span rubbing rotor-bearing system is analyzed. The dynamic equations of rotor-bearing system with rubs is set up considering non-linear oil-film force, eccentric mass of double shafts and coupling stiffness between two shafts. The numerical analysis results shows that the coupling of eccentric mass and rubbing will lead the system to produce a large chaotic motion area, and the size of chaotic region is related to the eccentric mass and severity of rubbing. When the mass eccentric directions of two rotors are same, the vibration displacement and the size of coupling rigidity is in a inversely-proportional relationship. When the mass eccentric directions of two rotors are opposite, the chaotic region of system response is the biggest. When there are rubs in rotor system, smaller coupling stiffness will cause chaos. Two-span rotor-bearing system test rig with rubs is built up. The influence of eccentric-mass-angle difference between two discs on dynamic characteristics of rubbing rotor system is verified through experiments. The experimental results are consistent with the results of numerical analysis basically. Therefore, eccentric mass of rotors will try to be reduced and coupling stiffness will be selected rationally when designing and assembling rotor system.(5) Rubbing fault feature extraction methods are studied. Owing to its self-adaptability and demodulation of modulated signal, Local Mean Decomposition (LMD) is suitable for analyzing non-linear, non-stationary and time-varying vibration signals. The smoothing methods of local mean and local amplitude are investigated. A new end-point extension method are presented based on difference, which is used to before LMD decompositions of multi-frequency simulation signals, rubbing simulation signals and testing signals from experiment rig of rubbing rotor system. The rubbing features are extracted successfully from above signals by improved LMD.(6) Recognition techniques for rubbing position of rotor system are studied. A new method of locating axial positon of radial rubs on multi-span rotor system based on LMD energy feature extraction and Support Vector Machine (SVM) is presented, which is verified by experimental data from the test rig of two-span rubbing rotor system. The experimental results show that the recognition ratio is to 97.33%, so it can be used to the location for axial positon of radial rubs on multi-span rotor systems. |
PDF Full Text Request