Kalman Filter-Based Reconstruction of Power Trajectories for IoT-Based Photovoltaic System Monitoring

This paper presents the reconstruction of signal paths acquired from a power electronics system for energy conversion and management. This reconstruction is performed using the Kalman filter (KF) for monitoring photovoltaic (PV) systems enabled for Internet of Things (IoT) systems. This proposal is motivated by the fact that the global energy transition towards renewable sources makes PV systems a crucial alternative. To guarantee the efficiency and stability of these systems, monitoring critical electrical parameters using IoT technology is essential. However, the measurements acquired are frequently corrupted by stochastic noise, which obscures the true behavior of the system and limits its accurate characterization. Based on this problem, the main objective of this work is explicitly defined as evaluating the effectiveness of the KF as a power-path reconstruction method capable of recovering accurate electrical trajectories from noisy measurements in IoT-monitored photovoltaic networks. To achieve this goal, the system is modeled as a discrete-time stochastic process and the KF is implemented as a real-time estimator of power flow behavior. The experiment was conducted using real-world generation and consumption data from a proprietary two-layer IoT platform: an Edge Layer (acquisition with ESP8266 and PZEM-004T-100A sensors) and a Cloud Layer (visualization on Things-Board). To validate the results, quantitative metrics including the mean squared error (MSE), statistical moments, and probability distributions were computed. The MSE values were found to be nearly zero across all reconstructed power-paths. The statistical moments exhibited near-perfect agreement with those of the actual power signals, approaching 100% correspondence. Additionally, the probability distributions were compared visually and assessed statistically using the Kolmogorov–Smirnov (KS) test. The resulting KS values were very low, confirming the high accuracy of the reconstruction for all power-paths. The proposed research concluded that the KF successfully reconstructed the power trajectories, demonstrating high agreement with the measured steady-state behavior. This study thus confirms that integrating Kalman filtering with IoT monitoring delivers a practically viable and statistically accurate method for power trajectory reconstruction, which is fundamental for enhancing the observability and reliability of photovoltaic energy systems.