Research On Fault Characteristics And Diagnosis Method Of The Superconducting DC Current Limiter | | Posted on:2021-01-13 | Degree:Master | Type:Thesis | | Country:China | Candidate:J J Pan | Full Text:PDF | | GTID:2492306107492894 | Subject:Engineering (Electrical Engineering) | | Abstract/Summary: | PDF Full Text Request | | With the rapid development of superconducting technology,the superconducting DC current limiters have become one of the ideal devices to limit the short circuit current of the power grid.The superconducting DC current limiter has the characteristics of no resistance in the superconducting state and high resistance in the current limiting state.It can reduce the fault current to a safe level quickly,which greatly improves the stability and reliability of power system operation.The structures and operating conditions of the superconducting DC current limiter(-196℃ low temperature,hundreds of kilovolts)are complex.Its long-term operation failures have become the focus of attentions.At the same time,its fault characteristics are non-linear and dynamic coupling.There is a problem of highly overlapping features between some fault modes,which brings great difficulties to fault identification and location.The purpose of this article is to study the fault characteristics of the superconducting DC current limiters.A fault diagnosis method for the superconducting DC current limiter based on LM-BP neural network is proposed.This paper mainly includes the following contents:(1)The fault tree analysis models and typical fault modes of the superconducting DC current limiters are established.The functions and working principles of the basic components in the superconducting DC current limiter are analyzed.The defects and potential failures caused by its weak parts are studied.The fault tree models and typical fault modes of the superconducting DC current limiters are established.The relationships between the faults from top to bottom are obtained.The electrical parameters(voltage,current)and non-electrical parameters(liquid level,temperature,pressure)of the superconducting DC current limiter are determined as important monitoring parameters for comprehensive fault diagnosis by studying the typical fault correlation coupling system.(2)The effects of superconducting tapes defects on the current limiting unit are studied by experiments.The fault simulation model of the superconducting DC current limiter based on thermal equivalent method is proposed.The typical fault characteristics of the superconducting DC current limiter are analyzed.The critical current and quench resistance characteristics tests of superconducting tapes under normal condition,conditions of strain and crack defects are experimented.The effect of defective tapes on the current limiting unit are obtained.An analysis and calculation method of superconducting DC current limiters based on the heat accumulation equivalent model of superconducting tapes is established.The nonlinear and multi-factor coupling resistivity relationships are simplified into a single correspondence relationship of heat accumulation-resistivity based on tapes heat accumulation-resistivity experimental characteristics.The calculation programs of thermal power and terminal voltage are designed.The fault characteristics of the typical fault modes of the superconducting DC current limiter under different operating conditions are obtained.(3)A fault diagnosis method of the superconducting DC current limiter based on LM-BP neural network is established.The input layers(monitoring data)-output layers(fault information)of the typical failure modes are defined.A fault diagnosis model of the superconducting DC current limiter based on BP neural network is established by training fault characteristic parameters of typical fault modes.The fault diagnosis model can dig the fault features deeply.In view of the high similarity of some fault characteristics of superconducting DC current limiters,LM-BP optimization algorithm is introduced to solve the problem of slow convergence and error cannot converge to the required value in BP neural networks.It is found that the optimized diagnostic algorithm has high accuracy and high stability in the implementation process by performance verification.The typical failure modes are integrated in stages and a hierarchical protection scheme for superconducting DC current limiters is established. | | Keywords/Search Tags: | Superconducting DC current limiter, failure modes, failure characteristics, failure diagnosis, LM-BP neural network | PDF Full Text Request | Related items |
| |
|