Denoising of Online Resistance Measurements of Power Connectors for IoT Applications

The increasing development of the Internet of Things (IoT) allows the acquisition of key parameters to develop health monitoring strategies for power transmission systems. These IoT devices need to incorporate sensors for data acquisition. However, noise is often present in the signals, especially in AC systems, making it a challenging task to obtain an accurate response from the deployed sensors. This paper develops simple moving average filter (SMAF), kalman filter (KF), and sum of sine filter (SoSF) on desktop platform, and analyzes these three filtering algorithms used to reduce signal noise from an IoT device mounted in a high-current laboratory. The IoT device can estimate the electrical contact resistance (ECR) in real time using smart sensors. Then, the most suitable algorithm is selected and implemented in the IoT device to facilitate the real-time signal processing task and obtain accurate ECR measurements. The experimental results have proven the feasibility and suitability of the proposed approach as it allows to extract useful information from the signals, thus facilitating the predictive maintenance task. The experimental results show that with the help of filtering algorithms, the error of raw signal is significantly reduced. For instance, when 312Arms is injected into the electrical loop, measurement error is 12.1% with raw signal and 1.9%, 3.4%, 2.1% using SMAF with moving window equal to 8, KF and SoSF, respectively. Furthermore, it is shown that with the approriate application of the filter, the IoT device is able to obtain highly accurate ECR measurements even at relatively low current.