Study On Cable Fault Detection Based On Traveling-wave Reflectometry

As the carrier of information transmission and energy supply of railway system,the reliability of railway cable is an important part of railway safety.Since the railway cables are buried deep underground,daily maintenance and repair are difficult to operate.Once the cable fails,it will have a great impact on train safety.Manual inspection is not suitable for underground cables,which is easy to waste manpower.The cable fault detection methods based on traveling-wave reflectometry are mainly studied in this thesis,which adopt single-end detection method to locate the fault of railway cable.A simulation model and a physical platform of Time Domain Reflectometry(TDR)are established combined with the study of cable fault types and cable structure,in order to study the theory and application of the traveling-wave reflectometry.The principle of Spread Spectrum Time Domain Reflectometry(SSTDR)which can be used online is further explored.Some important components of SSTDR such as the type of incident signals,related algorithms,and characteristic indicators are analyzed in detail to lay the foundation for the specific application of this methodSome improvements are proposed based on the SSTDR theory.Firstly,by adding quadratic correlation and generalized quadratic correlation on the basis of the one-time cross-correlation operation,the time delay algorithm can be improved;secondly,sign detector is added in the long-distance test.These two methods can respectively improve the anti-noise ability of the detection system and the recognition ability of correlation peak of the reflected signal.Cubic spline interpolation algorithm is used in SSTDR,increasing the number of sample points by interpolating the reflection correlation peak in order to reduce the error of the delay result.The accuracy of fault measurement results are improved without improving hardware devices or sampling frequency by using this method.Finally,a simulation model of SSTDR is built on MATLAB platform.Several simulation experiments are conducted under various signal-to-noise ratios to verify the above algorithms,and the relationship between the reflection peak identification ability and the sampling time and frequency is obtained through those experiments.An FPGAbased verification system is built in this thesis,and multiple sets of measurements of faulty cables of different lengths within 8 kilometers are completed.The results show that SSTDR can be used to locate cable faults.To sum up,in this thesis,there are 72 figures,4 tables and 56 references.