1 documents found
Information × Registration Number 0212U004791, 0110U002169 , R & D reports Title Methods of the frequency-temporary analysis of vibration signals boht system engineering of automated diagnostigs and forecasting of development of defects of hydrounits popup.stage_title Head Kuharchuk Vasyl, Registration Date 26-01-2012 Organization Vinnitsa National Technical University popup.description2 Object of research are the hydrounits of a vertical type, which during commercial operation are by a source of vibration. The purpose of work is increase of reliability of diagnosing and forecasting of de-fects of hydrounits on the basis of application of wavelet-transformations and forecasting of their de-velopment neuronets by technologies. The analysis of object diagnostic of vibrations is carried out, and is established, that by the basic reasons of vibrations of hydrounits (on an example Dnestr hydroelectric power station-2) is: a unbalance of a rotor; infringement vibrations coaxis of shaft; infringements of rigidity of basic system; defects of persistent and turbine bearings; vibration of electrical machines of an electromagnetic origin; infringement hydro- and gasedinamics of a flow. The dependences of making frequency of vibration on defects of bearings are experimentally received. The dependences of strips of frequencies on kinds of defects and meanings of vibrations of electrical machines from a current of loading the hydrounits and from a level of water in reservoir are established. Is shown, that the vibration of an electromagnetic origin has common property: its level in steps falls at switching-off of the machine from a network at the moment of a beginning of a stop, the level of vibration grows at fall of a level of water. On the basis of the developed common mathematical models for the measurements, most widespread in practice, of vibrations approached mathematical and electrical models are received, which analysis has found out new sources of an error during transformation of not electrical size of acceleration to electrical capacity, as has allowed to specify known mathematical parities. In static mode of operations of the measuring channel of vibration the equations of transformation, equation of sensitivity, analytical dependences for estimat-ing of an additive and multiplicate making error are received, the static characteristic of the measuring channel is constructed. The algorithms of discrete wavelet-transformation for different parent wavelets and different factors of compression are developed, therefore the optimum kinds wavelets and meanings of factors are established. The method frequency identification of a matrix of wavelet-factors is developed and are received amplituda-frequency-times spectra signal vibrations of for different modes of operations of object. Is developed non-standard non-uniform neuronets, into which structure enters three-layer neuronets for a task of diagnosing of existing defects and two-layer neuronets for forecasting development of defect. Procedure of training and the self-training of a neutron network is supposed during research operation of system. The practical meaning of work is, that are developed hardware and software of system, the experimental sample is made and the experimental researches are carried out. The introduction of such system on hydroelectric power station will allow to raise an overall performance of hydrounits, by constant monitoring of their condition, to raise reliability of their work and to reduce to a minimum emergency of a condition of this equipment. Product Description popup.authors Ведміцький Юрій Григорович Кацив Самоїл Шулімович Кухарчук Василь Васильович. Мадьяров Вячеслав Губейович. Усов Віктор Васильович popup.nrat_date 2020-04-02 Close
R & D report
Head: Kuharchuk Vasyl. Methods of the frequency-temporary analysis of vibration signals boht system engineering of automated diagnostigs and forecasting of development of defects of hydrounits. (popup.stage: ). Vinnitsa National Technical University. № 0212U004791
1 documents found

Updated: 2026-03-22