| Hydropower units are an important part of the hydropower generation process,and their important task of energy supply and flexibility regulation leads to the need for frequent unit operation during transition.When a hydropower unit operates in the transition process,the stability problems such as vibration,oscillation and pressure pulsation of the unit are prominent,and the potential factors that induce unit failure will increase.In this paper,the scientific problem is to explore the fault feature extraction method of hydropower units and the risk quantification model of transition process,to study the comprehensive misalignment fault mechanism of hydropower units,to realize the dynamic monitoring of unit operation and the reliability assessment of transition progress,and the research has achieved the following results1.On the basis of the analysis of its motion mechanism,a nonlinear dynamics model of the integrated misalignment fault of hydropower unit shaft system is established,and the effects of unit speed,misalignment bias and misalignment bias angle on the singular spectrum entropy value of rotor response are investigated respectively,the singular value features of rotor vibration signal are extracted,and the dynamic singular spectrum entropy and principal component analysis(DSVDE-PCA)are used to establish the early warning index of hydropower unit fault and build the unit process diagnosis model.The results show that the singular spectrum entropy can intuitively describe the change of rotor operation status when the rotor response changes significantly with the change of unit operation parameters.Specifically,when the rotational speed ω is 8 rad/s,17 rad/s,27 rad/s and 47 rad/s;the misalignment deflection d is 0.17 mm,0.22 mm,1.2 mm and 1.3 mm;and the misalignment deflection angle α is 9.6×10-4,the rotor operating state changes abruptly,and the response of the corresponding rotational speed in the singular spectrum entropy graph has a corresponding abrupt increase and decrease;When the speed ω is in the range of 17～27 rad/s,the rotor vibration amplitude has obvious fluctuations,and the corresponding speed range in the singular spectrum entropy diagram also shows drastic fluctuations.The actual data of the power station were used to verify the validity of the DSVDE-PCA model,and the unit fault warning indicators were constructed,including the Hotelling statistical indicator T2 and the squared prediction error SPE,whose warning values were 8.642 and 14.1974,respectively.When the monitoring model was run to 1558 monitoring points,the Hotelling statistical indicator T2 value exceeded the warning value,and it was 9 sampling points later than the actual When the monitoring model runs to 1558 monitoring points,the Hotelling statistical index T2 value exceeds the warning value and is delayed by 9 sampling points compared with the actual fault point;when the monitoring model runs to 1568 monitoring points,the prediction error squared value SPE exceeds the warning line and is delayed by 19 sampling points compared with the actual fault point,the results show that the process diagnosis model proposed in this paper can quickly and accurately discover the abnormal operating condition of the unit.2.The problem that the current fault diagnosis methods cannot effectively detect the early abnormal state of the unit and the single feature quantity of the unit fault data cannot effectively contain all the fault characteristics information is addressed.In this paper,we propose a dynamic singular spectral entropy unit abnormal state classification model and a multi-feature vector fault diagnosis method based on variational modal decomposition and support vector machine.When the rotor operating state changes at monitoring points 1448,3496 and 5544,the model of this paper shows abrupt changes at the corresponding monitoring points,and the results show that the proposed method can detect the abnormal operating state of hydropower units in a timely and accurate manner,which can provide a good basis for the next premaintenance and fault diagnosis.The results show that the proposed method can timely and accurately detect abnormal operating conditions of hydropower units and provide a theoretical basis for the subsequent pre-surge and fault diagnosis.The proposed fault diagnosis method is verified by using the rotor test bench data and the actual measurement data of the power station,and the two diagnostic methods of extracting fault feature sets directly without signal processing and extracting information entropy features by empirical modal decomposition(EMD)are compared and analyzed with the proposed fault diagnosis method.The overall diagnostic accuracy of the proposed method is 98.57% when using the rotor test bench data,which is 15% more accurate than the unprocessed direct extraction of fault feature set and 35% more accurate than the EMD decomposition of signal to extract information entropy features;the overall diagnostic accuracy of the proposed method is 98.57%when using the measured data of the power station.The overall diagnostic accuracy of the proposed method is 98.57%,which is 18% higher than that of the direct extraction fault feature set diagnosis method without signal processing,and 10% higher than that of the signal EMD decomposition extraction information entropy feature diagnosis method,and the results show that the proposed method can accurately diagnose the fault category of hydropower units.3.In order to accurately quantify the operating risk rate of a hydropower unit during the transition process,define the operating zone of the unit and realize the dynamic reliability assessment of the unit operation.In this paper,the operating zone division method based on the improved approximation ideal method(TOPSIS)and the reliability assessment model based on the improved proportional covariance(PCM)are proposed based on the dynamic balance experimental data of a conventional hydropower unit,and the results show that the proposed operating zone division method can accurately quantify the risk contribution rate of each evaluation index under different operating conditions of the hydropower unit,for example,the risk contribution rates of water pressure at the worm shell inlet For example,the risk contribution rate of the inlet water pressure of the worm gear is 0.0272,0.0268,0.027,0.046,0.062,0.057,0.056,0.06,0.056,0.04 and 0.046 in 10～110MW,and the risk ranking of the evaluation indexes in the medium-high risk operation zone is given,for example,more attention should be paid to the operation of the worm gear,top cover,tail pipe,upper and lower For example,in the high-risk operation zone,more attention should be paid to the operation of the worm shell,top cover,tail pipe,upper and lower frame and stator,and the risk affiliation degree of the unit under different operating conditions should be quantitatively described.When the unit is operated to 100-110 MW,the unrecommended operation zone,which originally ranges from 0-120 MW,can be reduced to 0-100 MW and the regulation capacity of the unit can be increased by 20 MW on the premise of paying more attention to the operation of the upper and lower guide bearings and upper frame.In addition,the risk assessment model based on improved PCM proposed in this paper has good evaluation accuracy,and its evaluation error parameters are: coefficient of determination is 0.99,root mean square error is0.0012,mean absolute percentage error is 0.266,and mean square error is 0.00245.The results show that the risk assessment model can dynamically update the real-time risk rate of hydropower units according to the real-time operation monitoring data of the units,and ensure the safe and stable operation of hydropower units.ensure the safe and stable operation of hydropower units. |
PDF Full Text Request