Speaker online pure voice failure diagnosis method

Abstract

The invention discloses an online pure tone fault diagnosis method for speakers, which includes a training method and a test method. The test method includes: adopt jumping swept signals as input signals of failed test samples; acquire time discrete signal sequence of speaker pure tone fault; adopt short time Fourier transformation time signals; extract the abnormal area with speaker pure tone fault in time-frequency domain and form training sample set; adopt principal component analysis (PCA) method, extract the feature set of speaker pure tone fault and at the same time, obtain the principal component transformation matrix; adopt a support vector machine to perform classification training for the feature set, in order to obtain a systematic side and determine classification; the test method also includes: according to harmonic distortion (HD) of all frequency bands in frequency spectrum, extract the speaker pure tone fault sequence in the time-frequency domain signal sequence and form test samples; obtain the tested speaker feature sample; compare the classification sides obtained from the feature sample and training and obtain the classification fault diagnosis results.

Description

technical field [0001] The invention relates to an on-line fault diagnosis method, in particular to methods for on-line detection, fault diagnosis and pattern recognition of loudspeakers. Background technique [0002] The loudspeaker is a transducer device that converts electrical signals into acoustic signals, and the performance of the loudspeaker directly affects the sound quality. In the production process of loudspeakers, due to the many production steps, the general production line only conducts quality inspection on the loudspeakers after the loudspeakers are assembled, and unqualified products are classified and returned for processing. [0003] China is currently the world's most important loudspeaker producer, and the output of loudspeakers accounts for about 80% of the global annual output. In terms of online pure tone detection and pure tone fault diagnosis of loudspeakers, the subjective judgment method of manual inspection is mainly used in China, and there is...

AI Technical Summary

Problems solved by technology

In terms of online pure tone detection and pure tone fault diagnosis of loudspeakers, the subjective judgment method of manual inspection is mainly used in China, and there is no research on the online pure tone fault diagnosis method of loudspeakers abroad.

Images