In order to use the properties of the orbits of shaft centerline effectively, the properties of the orbits of shaft centerline are studied. And the feasibility of improved back propagation algorithm which makes the c...In order to use the properties of the orbits of shaft centerline effectively, the properties of the orbits of shaft centerline are studied. And the feasibility of improved back propagation algorithm which makes the convergence faster is proved in the paper. The neural network classifier is constructed by using Fisher distinguishing principle. The problems of concentrated property of samples learning and sensitivity of given threshold are resolved effectively, so that the intellectual level of fault diagnostic system for turbo generator is improved.展开更多
文摘In order to use the properties of the orbits of shaft centerline effectively, the properties of the orbits of shaft centerline are studied. And the feasibility of improved back propagation algorithm which makes the convergence faster is proved in the paper. The neural network classifier is constructed by using Fisher distinguishing principle. The problems of concentrated property of samples learning and sensitivity of given threshold are resolved effectively, so that the intellectual level of fault diagnostic system for turbo generator is improved.
