| Description |
Spread spectrum time domain reflectometry (SSTDR) has found application in detection and localization of intermittent faults in live electrical cables. These sporadic faults can be open circuits, short circuits, or resistive changes, all of which preserve the original shape of the SSTDR correlated waveform. However, things are very different when SSTDR encounters a complex impedance discontinuity such as a capacitor or inductor. In this case, the reflection is a function of frequency and changes the shape of the SSTDR signature. So, in this case, we would need advanced signal processing algorithms to increase the accuracy and resolution of sampled SSTDR measurements. Algorithms with the best accuracy have high complexity, typically O(N2), which limits real-time fault location in embedded applications. So, the first part of this thesis deals with introducing three algorithms with a time and space complexity of O(N) to improve the accuracy of the SSTDR. We then evaluate these algorithms for accuracy on the practical case of calculating the velocity of propagation and the characteristic impedance of a photovoltaic (PV) cable. The second part of the thesis discusses the SSTDR response to single capacitive loads, capacitors connected in series, and simple curve fit algorithms to diagnose capacitive changes in the network. The simulations and measurements of these tests are provided in the thesis. Finally, this thesis presents preliminary SSTDR analysis for finding faults in live PV arrays. SSTDR was applied to a combination of PV modules, and responses are recorded and analyzed for various discrete discontinuities. An internal complex circuit diagram of a PV module is presented, for a better understanding of the effect of SSTDR signal inside the PV modules. |
