The present application relates to a kind of positioning identification method and device for
power equipment abnormal sound monitoring, the method first defines
spherical coordinate system, determines the number of
microphone in spherical
microphone array and the coordinate position of each
microphone thereof;Then construct the ideal sound
signal model of spherical
microphone array, and construct the
transfer matrix of focusing surface and array surface based on this in turn obtain
noise source intensity solving equation, since focusing surface division point number is much more than array microphone number, so the equation is underdetermined equation group;Finally, construct block sparse bayesian learning optimization framework and solve to obtain source
intensity coefficient vector, finally achieve the reliable monitoring and accurate positioning of
power equipment abnormal sound.The present application compared with prior art, overcome the deficiency that conventional
planar array is difficult to realize flexible autonomous positioning of three-dimensional space
sound field, can effectively improve the positioning accuracy of space
sound field, and still can guarantee good anti-interference performance in low
signal-to-
noise ratio environment.