Vibration Analysis And Reliability-based Robust Design Of Large-scale Compressor Rotor Systems With Multi-failure Modes

Rotating machinery is one of the most widely adopted and most essential core mechanical devices. Centrifugal compressor, a typical rotating machine, plays a more and more important role in fundamental industries, such as energy, chemical, aerospace and automotive industry.According to the problems of localized centrifugal compressor set, a new approach that adopts modern design methodologies such as reliability-based sensitivity design and reliability-based robust design in the processes of a product’s design and manufacturing is provided, aiming to design a product which can meet the requirement with high performance and high reliability under variation of processing and using conditions. To achieve this goal, this thesis concentrates on the following fields:1. The dynamic theory of random complex rotating systems and technology system of reliability-based robust design are elaborated, and on the basis of those theories, vibration analysis and reliability-based robust design of large-scale compressor rotor are presented in this thesis.2. A comprehensively analyzing approach for3D (3-dimensional) dynamic response of large-scale rotor systems based on3D geometric modeling software, finite element analyzing software and multi-body dynamic simulation software are put forward to establish the dynamic model of parameterized rigid-flexible nonlinear rotor system, and then rigid-flexible hybrid dynamic simulation analysis is then proceeded after proper restrains which can affect system’s reliability such as oil-film force, torque, working speed et al being properly simplified. Finally, the main reasons leading to system’s failure according to the simulation results are obtained, and the dynamic stress response distribution of the rotating shaft is obtained.3. A general transverse-rotational-axial-swinging coupled dynamic model of helical gear is developed by the rotor dynamic theory, a new type of coupling modeling method is introduced and different kinds of finite element model of rotor system are set up according to different coupled forms, different excitation forms and different coupling models. On the basis of dynamic characteristics of PTA-set geared-coupled compressor rotor system, the effect tendencies of many parameters such as the direction of rotation shaft, the selection of driving shaft, the helix angle, meshing stiffness, meshing damp, bearing stiffness, to the dynamic characteristics of system are discussed. The main reasons leading to system’s failure are obtained; the rubbing failure mode and resonance failure mode are put forward to establish reliability probability model, which will be used in the reliability and reliability-based robust design.4. The artificial neural network (ANN) technique is integrated into reliability design, and the response function derived from ANN model is more efficient than the finite element model of large-scale rotor system for reliability design. Based on the random perturbation method, reliability and reliability-based sensitivity theory, the reliability-based dynamic stress model is established, and a new method to compute the reliability of complex system which has strong nonlinear limit state function is put forward. And then the effect of designed parameters on the reliability of rotor system is studied, the sensitivity of reliability to the mean value and variance of original random parameters is obtained. Finally, the influence degree of all the original random parameters on the reliability is derived, which provides theoretical guidance for the product designing process.5. The logic formula and relative theory of reliability-based optimization design and robust design are elaborated. Based on the dynamic response of multi-shaft gear rotor system, the parameters which have greater influence on the dynamic characteristics are determined and two failure modes, rubbing failure and resonance failure are established, then the reliability-based probability model is set up for reliability design and reliability-based sensitivity design with dimensionless method. The influence degree of every parameter is derived and greater-influence parameters are obtained. On condition of certain given reliability, the reliability-based robust optimization design of random system is analyzed to reduce the sensitivity of those parameters to system’s reliability, which improves the stability of system and provides theoretical guidance on the basis of reliability theory.6. Reliability design and reliability-based sensitivity design of geared rotor system which has relative parameters are studied. On the basis of documents available, correlation system reliability is obtained by the narrow reliability boundary theory, and sensitivity design of reliability to the mean value and variance of original random parameters is developed to solve the complex correlation system.