Research On Forward And Inverse Problem Of The Total Electric Field And The Radio Interference For HVDC Transmission Lines

HVDC transmission technology, which can resolve contradiction among rapideconomic development, environmental pollution and unbalanced distribution ofresources, has been applied to west to east transmission projects due to its safety,stability and economy. However, the extensive attention has been paid to theenvironment effects of HVDC transmission lines, whose construction costs depend ondistributions of the total electric field and RI (radio interference). Great efforts havebeen made to characteristic analysis of the total electric field and RI from the testlines or corona cages but lack of the analysis of operation power lines. Meanwhile, thecurrent empirical formula such as EPRI cannot calculate the total electric field and RIfor the given temperature, humidity, etc.In order to overcome the technical bottleneck from theoretical calculation andexperiment means of the total electric field and RI on condition of changedenvironmental factors. The site measurement of practicaltransmission lines has beenimplemented. Furthermore,the novel theories and methods of the prediction algorithm,theforward and inverse problem of the total electric field and RI have been studied ina deep-going way. Moreover, the site distribution characteristic, the mathematicalmodel of theforward and inverse problem computation, iteration algorithm, theexamples application effect aredescribedindetailinthispaper.Theemphasisandachievementofthe paper mainlymanifestsinthefollowingaspects.(1) The site measurement for the total electric field and RI of UHVDCtransmission lines from Yunnan to Guangzhou (YG) projecthas been proposed torealize the idealization model of the factors such as the terrain for the test lines orcorona cages. The distribution of the total electric field and RI has been achievedthrough the further research on the characteristic of the different operation mode. Theerror between the measurement values and the regulationrecommendedsimulationvalues has been analyzed, whose reason is that the effect factors such as the altitudeand the terrain has been neglected by the regulationrecommended. Meanwhile, themeasurement vales can provide the prediction algorithm, theforward and inverse problem of the total electric field and RI for the fundamental data.(2) The control equations and boundary conditions for the corona andnon-corona region of the forward problem of the electric field have been presentedbased on the Poisson equationsof electric field, the ion flow diffusion equations andcurrent continuity equations, which can take the effect of the corona region and theenvironment factors into account. And the meshless flux corrected transportalgorithmand wavelet transformation have been employed to solve the control equations of thecorona region. Finally, the numeric examples has demonstrated that the proposedalgorithm has good convergence, solve the iteration divergence problem due to thegrid distortion for the finite element method, improve the computation accuracy anddecreased the calculation storage.(3) The corona current can be achieved by integration method based on the DCtransmission lines corona model, which can compute the corona current for the givenenvironment factors and overcome the constrained calculation from the excitationfunction because of the dimension and environment factors for the test lines of coronacages. Furthermore, the admittance matrix and the impedance matrix with highfrequency parameters can be obtained from the open circuit and short circuitimpedance measured by two phase power resources for HVDC transmission line,which can solve the low calculation accuracy problem due to the effect the earthconductivity and skin effect. Moreover, the superposition principle can be proposed toobtain the RI when the phase-mode transformation is applied to decouple the coupledrelative between bipolar lines. Finally, the method for high frequency parameters hasbeen verified by the equivalent length1000km of HVDC transmission lines mode.Meanwhile, the comparative analysis between simulation calculation andmeasurement value for full-voltage of negative pole, full-voltage of bipolar hasverified the accuracy reliability.(4) The optimization function of potential error between the site measurementvalues, conductors surface potential, the ground potential and the calculation resultsfrom the simulation charge has been presented to realize the model establish for thetotal electric field and RI inverse problem. Furthermore, the detailed solving procedure of the trust region regularization genetic algorithm has been proposed,which can optimize the location of the simulation, overcome the error increasingproblem due to the large solving iteration step and realize the iteration step constrainand its direction dynamic adjustment. Consequently, the several examples fromhalf-voltage of positive pole and full-voltage of negative pole, full-voltage of negative,full-voltage of bipolar have demonstrated that the solving algorithm has goodconvergence, stability and accuracy reliability.(5) The fuzzy clustering algorithm has been proposed to select the similarsamples data with the environment permanents of the prediction points. Furthermore,the distribution of the total electric field and RI can be simulated by the least squaressupport vector machine method (LSSVM) based on the similar samples data.Moreover, the penalty coefficient and the scale factor in LSSVM can be optimized bythe particle swarm algorithm, which can improve the computation efficiency for thetotal electric field and RI. Finally, the higher accuracy prediction results of the totalelectric field and RI can be achieved through the comparative between predictedresults and measurement values for different operation mode and environment factors.