Direct current (DC)/DC converter fault diagnosis method and system based on improved sparrow search algorithm

Patent metadata

Field	Value
Publication number	US-11853898-B1
Application number	US-202218060582-A
Country	US
Kind code	B1
Filing date	Dec 1, 2022
Priority date	Jun 7, 2022
Publication date	Dec 26, 2023
Grant date	Dec 26, 2023

Abstract

A DC/DC converter fault diagnosis method based on an improved sparrow search algorithm, includes: establishing an simulation module of the converter, selecting a leakage inductance current of a transformer as a diagnosis signal, and collecting diagnosis signal samples under OC faults of different power switching devices of the converter as a sample set; improving a global search ability of a sparrow search algorithm by using a Levy flight strategy; dividing the sample set into a training set and a test set, preliminarily establishing an architecture of a deep belief network, and initializing network parameters; optimizing a quantity of hidden-layer units of the deep belief network by using an improved sparrow search algorithm, to obtain a best quantity of hidden-layer units of the deep belief network; and training an optimized deep belief network obtained based on the improved sparrow search algorithm, and obtaining a fault diagnosis result based on a trained network.

First claim

What is claimed is:

1. A DC/DC converter fault diagnosis system based on an improved sparrow search algorithm, comprising:
a data collection module configured to establish a simulation module of a DC/DC converter, select a leakage inductance current of a transformer as a diagnosis signal, code and classify a fault type based on open circuit (OC) fault states of different power switching devices of the DC/DC converter, and collect diagnosis signals of the DC/DC converter under different fault states as a sample set;
an algorithm optimization module configured to improve a global search ability of a sparrow search algorithm by using a Levy flight strategy;
a network optimization module configured to optimize a quantity of hidden-layer units of a deep belief network by using an improved sparrow search algorithm, and search for a best quantity of hidden-layer units of the network;
a network training module configured to set the quantity of hidden-layer units of the deep belief network as the best quantity of hidden-layer units, train the deep belief network by using the training set, and test a trained deep belief network by using the test set; and
a fault diagnosis module configured to input a newly obtained test sample into the trained deep belief network directly for fault diagnosis to obtain a diagnosis result.

Assignees

• Univ Wuhan

Inventors

• He Yigang
• Zhao Yingying
• Xing Zhikai
• Liu Xiaoyu
• Wang Xiao

Classifications

• G06N3/086 Primary

  using evolutionary algorithms, e.g. genetic algorithms or genetic programming · CPC title

• G01R31/2848

  using simulation · CPC title

• G06F30/27 Primary

  using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model · CPC title

• G06N3/0464 Primary

  Convolutional networks [CNN, ConvNet] · CPC title

• G06F30/38

  Circuit design at the mixed level of analogue and digital signals · CPC title