Research On Characteristic Analysis And Optimized Control Of Offshore Wind Turbine In Complex Environment

Compared with conventional coal-fired generating units,the working conditions of gas generating sets are more stringent.In order to improve the safety,reliability,service efficiency and service life of gas turbine operation,avoid causing major equipment accidents and affecting power supply,it is very important to carry out continuous monitoring and protection of the operating status of the equipment and timely fault diagnosis and identification.Due to the energy conversion of gas turbine through high temperature and high pressure air flow,the alternating mechanical stress and thermal stress of each component structure on the rotor system is more complex,and it is also impacted by micro particles in the air flow.as a result,the operating conditions of the structure are more extreme,and the probability of abnormal states such as blade scaling,corrosion,clearance change,high temperature fatigue and creep is higher,and even blade fracture fault occurs.It causes secondary fault loss,which seriously affects the safety and reliability of its operation.Aiming at the multi-disk tie rod structure and multi-stage turbine blades of gas turbine rotor,this paper studies the dynamic response characteristics of the rotor system under two special fault conditions:tie rod loosening and blade breaking,and deeply analyzes and understands the particularity of the dynamic problems of multi-disk tie rod rotor.it can provide technical support for rotor design.It can also accumulate typical cases and provide technical methods for operation condition monitoring and fault diagnosis and early warning of actual gas turbine rotor system,which has strong engineering application value.The main research contents and achievements are as follows:(1)Modeling and analysis of nonlinear dynamic characteristics of pull rod rotor considering interdisk relaxation:The looseness of the pull rod caused by long-term operation of the multi disc pull rod structure of a heavy-duty gas turbine is a special failure mode In this paper,the rotor dynamics under the condition of axial tie rod loosening is simulated,and the nonlinear dynamic model of tie rod rotor is established.based on the fractal contact theory of rough surface,the relationship between contact stiffness and contact load is studied,and the influence of contact stiffness on displacement bifurcation and spectrum component of tie rod rotor system is discussed.Then,the effects of uniform and non-uniform relaxation of pretension of circumferential tension rods on the dynamics of rotor system are studied.The results show that there is a nonlinear relationship between the contact load and the distance between the two planes,the contact stiffness and the contact load is weakly nonlinear,and the bending stiffness decreases with the decrease of the preload.The influence of the contact stiffness between the plates on the dynamic characteristics of the tie rod rotor system is different at different speeds,which is mainly related to the motion state of the system.The uniform relaxation and non-uniform relaxation of the pretension of the disk show different nonlinear response characteristics.(2)Modeling and analysis of dynamic characteristics with distributed blade defects on multi-disk pull rod rotors:Blade fault is the main fault mode of gas turbine.The fault forms and distribution parts are various,which indirectly shows that the axial Mass distribution of the rotor system changes.In this paper,a simulation study is carried out on the multi disk rotor dynamics problem under the condition of uneven axial Mass distribution,and a four disk simulation rod rotor structure dynamics model is built.Based on the nonlinear characteristics of disk contact stiffness and oil film bearing stiffness,the blade failure is simulated,and the nonlinear dynamic response characteristics of the rotor are studied and analyzed.Research has shown that when a rotor blade malfunctions,different positions will affect the dynamic characteristics of the rotor system,and the frequency doubling component,amplitude,axis trajectory,and phase difference of the bearing’s vibration spectrum will vary with the angle between the unbalanced eccentricity of the fault,and have certain regularity.(3)Experimental study on fault diagnosis of multi-stage turbine blades:A gas turbine experimental test system was designed and built,and a typical blade fault diagnosis experiment was carried out based on multi-source dynamic information.Several typical blade fracture faults are simulated on a multi-stage turbine experimental unit.Transient signals of gas path pressure,cylinder vibration and rotor vibration under different operating conditions of the turbine are analyzed by classical and modern signal analysis methods.Support vector machine is used as a classifier.The effect of single sensor type and multisensor information fusion on blade fault classification and recognition is studied.A sensor information fusion analysis method based on "coherent cepstrum" analysis is developed,which can extract characteristic values that are better for distinguishing different blade states and improve the accuracy of blade fault classification and recognition.(4)Blade fault diagnosis method combined with feature selection and deep classification network:The feature extraction and compression process in intelligent fault diagnosis play a key role.The depth of the neural network technology developed in recent years,although the relative to the requirement of the input characteristic value is reduced,but for small sample classification problem,the choice of input feature vectors are still play an important role,based on this consideration,this paper proposes a feature vector based evaluation combining modern intelligent classifier of fault diagnosis method of turbine blades,Will first to extract the characteristic value of different kinds of monitoring signal,composed of high-dimensional feature matrix,and then use a variety of characteristics of the engineering method for feature selection or compression,realize the feature dimension reduction,due to the different characteristics of the engineering method of low dimensional feature vector is different,this article introduced a feature vector of PFE assessment method,the classification of the different characteristic vector can be effect gives the quantitative evaluation,On this basis,feature vectors with good classification effect were selected to train the intelligent classification model,and the classical SVM classifier and modern LSTM classifier were used to diagnose and identify the blade fracture faults respectively,which verified the effectiveness of the proposed method.