| Transmission line galloping is a kind of low-frequency,large-amplitude vibration of iced conductors with non-circular section under wind excitation.Due to the large amplitude,galloping will lead to fatigue and increased tension of conductors on the one hand,which may cause damage to the fittings and even tower collapse.On the other hand,phase conductors may be too close and cause flashover,tripping and other consequences,posing a serious threat to the safe operation of the power grid.At the macroscopic level of galloping spatial distribution,the galloping distribution map is an important guide for route planning and differentiated anti-galloping,so it is of important practical significance to study galloping spatial distribution model fusing multi-source micro-topographic and micro-meteorological information to obtain a more detailed galloping distribution map.At the micro level of the galloping process,for established lines in regions prone to galloping,the research on trajectory extraction algorithms based on video monitoring and early warning systems establishment can serve basic galloping data accumulation,mechanism and model verification,and operation status evaluation,which is of great strategic significance to galloping theory research and grid disaster prevention and mitigation.Aimed at solving problems in the current galloping spatial distribution model of State Grid(meteorological-geographical method),such as the lack of consideration of synchronous change of meteorological elements,ignoring the spatial configuration of observations in interpolation,and insufficient expression in specific terrain and landforms of obtained galloping distribution map,this paper proposed a Geostatistical Meteorology Model(GMM)method.In this method,optimal geostatistical interpolation models of daily minimum temperature,daily average relative humidity,daily maximum wind speed and corresponding wind direction were first constructed,then the galloping distribution map was inverted based on daily meteorological surfaces,and the ability of galloping distribution map to express micro-topography was thus improved by improving the interpolation accuracy of the galloping sensitive meteorological elements.Since the above meteorological elements to be interpolated involved extreme values(maximum/ minimum)and vector characteristics at daily scale,the optimal geostatistical interpolation model for each meteorological element was determined by fusing multi-source terrain and land cover information,synthesis after interpolation of vector components,considering the spatial non-stationarity of the regression,and introducing covariates.Refering to the micro-topography factors thataffecting icing and wind excitation,and combing general topographical and meteorological information of Henan Province,the following factors were introduced to explanatory variables for the first time: longitude difference,angle between slope aspect and the north direction,Relief Degree of Land Surface(RDLS),and area of water body and forest in the neighborhood.The mixed Geographically Weighted Regression(m GWR)model with local parameters varying with geographic location was established for trend,Ordinary Kriging(Ordinary Kriging,OK)or Co-Kriging(CK)interpolation was performed on the residuals using correlation of variables and crosscorrelation between variables,and the optimal(m)GWR(C)K model with the smallest Root Mean Square Error(RMSE)was therefore constructed.The meteorological data on which day large-scale galloping accidents occurred in Henan Province were used for accuracy validation of the optimal geostatistical interpolation model.The results showed that,compared with the models in improved meteorological-geographical method,the RMSE of the optimal geostatistical interpolation model was reduced by 22.227?48.974%.Furthermore,the meteorological data from 1998 to 2013 was used to invert the galloping distribution map according to the GMM method proposed in this paper.Compared with the 2013 version of the State Grid galloping distribution map,results of GMM method could better reflect the influence of micro-topography with canyon effect,alpine watershed and rich water vapor on meteorological elements and galloping,and could effectively deal with the problem of underestimating the galloping days at stations with missing or abnormal meteorological data.For existing conductor or spacer positioning algorithms in the video online monitoring method rely too much on manual operations and are not highly automated,a new algorithm for spacer centroid positioning based on Block Directional Field(BDF)and Normalized Rotation Auto Correlation(NRAC)was proposed in this paper according to conductor and spacer morphology.In this method,BDF was first used to express the linear characteristics of the conductor and perform conductor region segmentation,NRAC was then used to measure the rotational symmetry,and the centroid of the spacer was the local optimal center of rotational symmetry due to its regular polygon shape.For the situation without supplementary observation like distance and elevation angle,a monocular plane measurement scheme with the spacer structure points as the control points was further proposed according to its motion characteristics.The spacer structure points were first extracted through morphology,and the mapping from pixel coordinates to spacer plane coordinates was realized with the knowledge of spacer dimensions,thus forming a complete video monitoring-based galloping trajectory extraction solution.The online monitoring video of Jianshan fullscale transmission lines was used to verify the proposed algorithm of spacer centroid positioning,and the results showed that the BDF and NRAC based method could greatly reduce the degree of manual operations,and the RMSE of point position reduced by 5.900?34.079% compared to that of normalized cross-correlation template matching algorithm.With the results of manual control points as a reference,the plane measurement accuracy of spacer centroid in the continuous frames was verified.The RMSE of spacer centroid coordinates was between 2.760 cm and 9.521 cm,and the RMSE of the displacement(difference from sequence mean)in X and Y directions ranged from 0.220 cm to 7.090 cm.The centimeter-level accuracy was equivalent to that of existing monitoring methods of acceleration sensor and Real-Time Kinematic(RTK),which could meet the demands of galloping monitoring.In addition,the spacer centroid positioning and plane measurement results were checked by the galloping monitoring device verification platform,and the relative error of maximum displacement difference and galloping frequency is less than 5%.In response to the lack of unified galloping disaster warning indicators and grading standards at present,a "meteorological early warning + phase safety factor early warning" mode was proposed and a double-layer early warning system for the galloping possibility and galloping is constructed,by considering the icing and wind excitation conditions for transmission line galloping,the possible electrical failures,and meteorological disaster grading early warning mechanism in China.Meteorological early warning was to interpolate hourly meteorological elements before galloping,and give a 4-level early warning of galloping possibility based on reaching icing conditions(level IV),icing and wind conditions(level III),continuous and stable wind excitation(level II),vertical direction of the wind from lines(level I).Since the galloping amplitude was an important parameter representing the dancing intensity,a two-step extraction method of galloping characteristic parameters based on the frequency domain and time domain was proposed,the phase safety factor was introduced to describe the potential electrical performance threats,and a 4-level early warning of galloping intensity and potential electrical threat warning was realized.The online monitoring and early warning system for transmission line galloping was constructed by component geographic information system(GIS)development technology,covering all the models and algorithms in this paper.The thematic functions of geostatistical interpolation model of meteorological elements,spacer positioning based on video monitoring,plane trajectory measurement,galloping characteristic parameters extraction,and double-layer warning were embedded in GIS,realizing visualized data management,analysis and decision support for galloping. |
PDF Full Text Request