ABSTRACT
The monitoring of leakage current (LC) and voltage characteristics in transmission line insulators is regarded as a good technique for anticipating the physical state of in-service insulators. In the current work, the temporal and frequency characteristics of LC and voltage under various situations were derived for assessing the health condition of porcelain, glass, and silicone rubber insulators. The contamination severity indicated by soluble deposit density, wetting level (Wt), non-soluble deposit density, and uneven pollution distribution (Pu/PL) were chosen as the environmental factors that impact the insulators. Six criteria were utilized to evaluate the physical state of the insulators, with four of those derived from the LC signal in the time domain, namely, the LC signal peak (C1), the phase shift between applied voltage and LC (C2), the LC signal slope between two consecutive peaks (C3), and the crest factor (C4). The remaining two indices, namely, the total harmonics distribution (C5) and the harmonics ratio indicator (C6), were obtained from the frequency domain of the LC signal. In addition, the flashover voltage index (C7) was also employed. The LC indicators were then classified based on the laboratory test results to reflect the physical state of the insulators. The findings revealed that the proposed indicators had an important impact in determining the physical state of the insulators. Furthermore, a confusion matrix was created for the test and prediction data using the suggested indicators to determine the effectiveness of each indicator.
ABSTRACT
Early drone anomaly inspection is vital to ensure the drone's safety and effectiveness. This process is often overlooked, especially by amateur drone pilots; however, some faulty conditions are difficult to notice by the naked eye or discover, even though the drone inspection process has been conducted; therefore, a real-time early drone inspection approach based on vibration data is proposed in this study. Firstly, the reliability of several microelectromechanical systems (MEMS) sensors, namely the ADXL335 accelerometer, ADXL 345 accelerometer, ADXL377 accelerometer, and SW420 vibration sensor in detecting faulty conditions, were tested and compared. The experimental results demonstrated that the vibration parameter measured using ADXL335 and ADXL345 accelerometers are the best choice as most of the faulty conditions can be detected, in contrast to other MEMS sensors. The output produced from the anomaly inspection algorithm is then converted to the "Healthy" or "Faulty" state, which is displayed in a mobile application for easy monitoring.
