Research On Fault Feature Identification And Diagnosis Of Gearbox Of Doubly-fed Wind Turbine

In recent years,the rapid development of wind power generation has effectively reduced the use of fossil fuels.The development of wind power generation is of great significance for alleviating energy crisis and coping with environmental pollution.As a typical rotating machine,the parts of the wind turbine are impacted by non-uniform load during operation,which is easy to cause failure,thus affecting the stability of power generation.Maintenance work may solve the minor failures.However,when it comes to the serious failures,shutdown maintenance or component replacement will be necessary,which will cause huge economic losses.Therefore,the fault diagnosis of the fan becomes the prerequisite to ensure the running state.As the important part of the wind turbine to transmit motion and power,the fault diagnosis of the gearbox is particularly important.This thesis takes the gearbox of a 1.5MW doubly-fed wind turbine as the research object.Because the first-stage planetary structure of the gearbox bears the largest torque,it’s easy to damage the its sun gear.And,at the high-speed end of the gearbox,the rolling bearing is responsible for carrying and transmitting power,which is prone to failure under the impact of alternating loads.Therefore,the fault diagnosis,identification and classification of the first planetary structure of the planetary gearbox,the sun wheel and the high-speed end of the rolling bearing are studied.Firstly,the failure mechanism of gear and bearing is analyzed and studied.The influence of the two faults on the box is analyzed,which provides a theoretical basis for the study of fault diagnosis and identification classification.Secondly,according to the actual working conditions of the planetary gearbox,the nonlinear vibration signal of the first planetary structure of the planetary gearbox is collected.In order to extract effective fault features and process nonlinear signals,a new fault diagnosis method is established.The sparrow search algorithm was introduced to optimize the variational mode decomposition,and the combination of kurtosis and cross-relation number was used as the fitness function to realize the adaptive decomposition of vibration signals.The method is verified by using the data of the simulated gear signal and the measured gear signal,and the fault frequency and frequency doubling are obtained.The fault diagnosis of fan gearbox under variable working conditions is realized.Finally,aiming at the continuous collection of vibration signals,how to identify and classify the signal features quickly and accurately is another research focus of the thesis.Wavelet packet decomposition is used to extract the energy characteristics of each frequency band of the signal.At the same time,the chaotic sequence,gaussian mutation and sparrow search algorithm form the chaotic sparrow search algorithm,and support vector machine is combined.Energy characteristics are used as input parameters for adaptive fault classification,the chaotic perturbation and gaussian mutation ensure that the algorithm can escape local constraints and improve the overall search accuracy.It is verified by bearing data from public database and experiments.The results show that the classification effect is good and have a high recognition rate.The fault identification and classification of planetary gearbox under variable load condition are realized,which has practical significance for ensuring the normal operation of wind turbine.