Risk and impact-centered non-stationary signal analysis based on fault signatures for Djibouti power system

Power system engineers’ intention is to produce power, transport it, and finally distribute it to customers under safe and reliable operating conditions to provide continuous and stable electrical energy. However, this goal is often hindered by unexpected faults that can lead to system breakdown. As a result, power system modeling is an effective technique which helps the analysis of power systems. Besides this technique, mathematical methods provide comprehensible information about the system's state. Furthermore, various studies have employed different techniques to detect electrical faults. In this paper, electrical faults are selected using a risk and impact approach, and fault characteristics are found using the Short-time Fourier transform. The case study is the Djibouti power network, and it is modeled with all real parameters using MATLAB-SIMULINK software. Following that, several fault scenarios were run, and the analysis was conducted using the proposed mathematical method. Ultimately, the simulation results indicate that the most critical faults are single-line-to-ground and double-line-to-ground. Hence, the extraction of signal features for these fault types is carried out. These faults have a high short circuit current, which can cause damage to the electrical network, and while clearing the fault, the oscillatory transient state appears with low-frequency components. These frequency components significantly affect power quality, thereby reducing the system’s performance.