Howling Quantitative Evaluation Method Based on Orthogonal-Pairwise Comparison Test and SVR

Abstract

Based on the orthogonal-paired comparison test and the SVR quantitative evaluation method of howling, firstly combine the orthogonal test and the paired comparison test to obtain the test samples used for training the model, and then carry out the subjective evaluation of the measured audio samples, and conduct different pure tone pairs Quantification of howling abnormal noise contribution, combined with the idea of ​​exclusion-comparison and grade scoring method, and finally using support vector regression to build a multi-pure tone howling quantitative evaluation model, and carried out effectiveness testing; characteristics, can greatly reduce the number of tests; pairwise comparison has the characteristics of easy operation, simple and fast, does not require high professional knowledge, and can better reflect the target characteristics; the invention realizes the contribution of the howling frequency in the audio signal to the total noise Quantitative evaluation of size.

Description

[0001] Technical solutions [0002] The invention relates to the technical field of whistling and abnormal sound evaluation of decelerators, in particular to a quantitative evaluation method of whistling based on an orthogonal-pair comparison test and SVR. Background technique [0003] The speed reducer is a mechanism used to change the rotational speed and torque from the engine. It can realize the output of different rotational speeds. It has the advantages of high efficiency, simple structure and convenient use. It is widely used in ships, vehicles and other fields. The gear system of the speed reducer is generally composed of a constant meshing gear pair, a gear meshing gear pair and an idler meshing gear pair. During the production process of the gear, due to the lack of precision of the processing equipment and unreasonable assembly operation, the gear tooth surface appears microscopic. Problems such as uneven appearance and local defects, which in turn lead to harsh whi...

AI Technical Summary

Problems solved by technology

Judging howling by directly analyzing the sound pressure level of the center frequency by 1 / 3 octave band only considers the energy of the noise signal, ignoring the influence of comprehensive factors such as human psychology, physiology, and environment, and cannot give howling noise based on the subjective feeling of the human ear. Quantitative index called abnormal noise

There are usually multiple pure tone components in the noise signal of the gear reducer. The TNR index in the ECMA standard calculates the signal-to-noise ratio in the critical frequency band, which can indicate the contribution of the pure tone to the howling to a certain extent, but ignores the psychological and physiological factors. , environment and other factors, it is impossible to give the contribution percentage of each pure tone component to the abnormal howling sound

Images