Abstract: The present invention provides a hydraulic turbine cavitation acoustic signal identification method based on big data machine learning. According to the method, time sequence clustering based on multiple operating conditions under the multi-output condition of the hydraulic turbine set is performed by utilizing an neural network, characteristic quantities of the hydraulic turbine set under a steady condition in a healthy state is screened; a random forest algorithm is introduced to perform feature screening of multiple measuring points under steady-state operation of the hydraulic turbine set, optimal feature measuring points and optimal feature subsets are extracted, finally a health state prediction model is constructed by using gated recurrent units; whether incipient cavitation is present in the equipment is judged.

Abstract: A three-dimensional reconstruction method based on half-peak probability density distribution, including: slicing three-dimensional point cloud along Z-axis direction to obtain N spatial layers; extracting the scatter information in i-th spatial layer and projecting information to Zi plane; constructing membership function of each grid and scatter in the Zi plane and drawing a three-dimensional probability density plot; making a plane parallel to XOY plane through half-peak wmax/2 of three-dimensional probability density plot, parallel intersecting a three-dimensional probability density plot to obtain a contour LXY; superimposing radioactive source reconstruction contours corresponding to N spatial layers sequentially to obtain a three-dimensional reconstruction model of a radioactive source.

Abstract: The present invention discloses an optimization design method for structural parameters of biomass boiler economizers and belongs to the field of big data learning models. In the present invention, a sample database is established by utilizing historical operating big data of biomass boiler economizers, a heat exchanger residual self-attention convolution model is established based on a CNN and a self-attention mechanism, a plurality of target parameters to be optimized are quickly predicted through machine learning, and multi-target optimization of structural parameters to be optimized in the economizers can be performed in combination with an iterative optimization algorithm.

Abstract: The present invention discloses an optimization design method for structural parameters of biomass boiler economizers and belongs to the field of big data learning models. In the present invention, a sample database is established by utilizing historical operating big data of biomass boiler economizers, a heat exchanger residual self-attention convolution model is established based on a CNN and a self-attention mechanism, a plurality of target parameters to be optimized are quickly predicted through machine learning, and multi-target optimization of structural parameters to be optimized in the economizers can be performed in combination with an iterative optimization algorithm.

Abstract: The present invention provides a hydraulic turbine cavitation acoustic signal identification method based on big data machine learning. According to the method, time sequence clustering based on multiple operating conditions under the multi-output condition of the hydraulic turbine set is performed by utilizing an neural network, characteristic quantities of the hydraulic turbine set under a steady condition in a healthy state is screened; a random forest algorithm is introduced to perform feature screening of multiple measuring points under steady-state operation of the hydraulic turbine set, optimal feature measuring points and optimal feature subsets are extracted, finally a health state prediction model is constructed by using gated recurrent units; whether incipient cavitation is present in the equipment is judged.

Abstract: The present invention discloses a three-dimensional reconstruction method based on half-peak probability density distribution, comprising the following steps: slicing three-dimensional point cloud along Z-axis direction to obtain N spatial layers; extracting the scatter information in i-th spatial layer and projecting information to Zi plane; constructing membership function of each grid and scatter in the Zi plane and drawing a three-dimensional probability density plot; making a plane parallel to XOY plane through half-peak wmax/2 of three-dimensional probability density plot, parallel intersecting a three-dimensional probability density plot to obtain a contour LXY; superimposing radioactive source reconstruction contours corresponding to N spatial layers sequentially to obtain a three-dimensional reconstruction model of a radioactive source.