Fan fault diagnosis method based on extreme learning machine

Abstract

This paper discloses a fan fault diagnosis method based on an extreme learning machine, which includes the following steps: using the time-domain characteristic parameters of the fan bearing vibration signal as a sample feature vector to form a training set and a test set; respectively using 1, 2, 3, and 4 Identify the categories in the training set and test set; learn the ELM classification model on the training set: select the activation function, and obtain the ELM parameters by improving PSO combined with CV optimization; substitute the samples of the test set into the classifier to verify its category. Strengthening the fault diagnosis of wind turbine bearings is of great significance to reduce the downtime of wind turbines and improve the economic benefits of wind farm operation.

Description

technical field [0001] The invention relates to the field of fan fault diagnosis, and more specifically relates to a fan fault diagnosis method based on an extreme learning machine. Background technique [0002] With the continuous increase of the world's population and the continuous development of society, human beings' demand for energy is increasing day by day, while the reserves of traditional energy sources such as oil and coal are decreasing sharply. Therefore, it is particularly important to vigorously develop new energy sources. Wind energy is a non-polluting, renewable new energy source, so wind power generation has received more and more research and development around the world in recent years. However, wind turbines are mostly installed in very harsh natural environments, and are easily affected by wind impacts with variable speeds and directions and erosion by temperature differences, so wind turbines are more prone to failure. Bearings are crucial transmissio...

AI Technical Summary

Problems solved by technology

The BPNN algorithm can realize nonlinear complex mapping and has good adaptive ability; however, the number of hidden layers of the algorithm is difficult to determine, and there is "overfitting", the convergence speed is slow, and it is easy to fall into the problem of local optimum, which affects the model predictive ability; the SVM algorithm does not have the problem of the BPNN algorithm; but it needs to obtain the support vector with the help of quadratic programming, and the constraint condition is an inequality constraint, which affects the classification accuracy; the LS-SVM algorithm is improved on the basis of the SVM algorithm, and the minimum The square linear system is used as a loss function instead of quadratic programming, and equality constraints are used instead of inequality constraints; however, the LS-SVM model does not have sparsity, and all training samples need to be used as support vectors for the classification of unknown samples, resulting in the algorithm the training speed is slower

Images