University of Mohaghegh Ardabili Journal of Operation and Automation in Power Engineering 2322-4576 2025 11 27 Enhanced DC Microgrid Protection: A 2D Current Modeling and Deep Learning Approach 4262 10.22098/joape.2025.15874.2219 EN Behrooz Taheri Department of Electrical Engineering, Qazvin Branch, Islamic Azad University, Qazvin, Iran. Seyed Amir Hosseini Electrical and Computer Engineering Group, Golpayegan College of Engineering, Isfahan University of Technology, Golpayegan, Iran. Mostafa Sedighizadeh Faculty of Electrical Engineering, Shahid Beheshti University, Evin, Tehran, Iran. Journal Article 2024 09 18 This paper introduces a novel protection method for identifying and locating faults in DC microgrids, which is aimed at overcoming the challenges faced by modern power systems. A two dimensional current modeling technique is utilized to detect faults, in which even minimal changes in the sampled data result in rapid detection due to the model's sensitivity. Additionally, the method differentiates between transient and permanent faults and is robust against noise in sampled signals. Furthermore, a deep learning model based on long short term memory layers, optimized using the whale optimization algorithm, is applied for fault location. The deep learning model's layers are fully aligned with the data, and the optimization process enhances the model's accuracy. The proposed scheme operates without relying on extensive communication links, making it practical for real world applications. Comparative evaluations demonstrate that the system outperforms existing methods in terms of accuracy, speed, and reliability, confirming its effectiveness in DC microgrid protection. The deployment of the proposed method effectively identifies and pinpoints faults at various locations within the microgrid in as little as 1 millisecond and within PV and EV components in up to 11 milliseconds. This capability has been validated across a range of fault types and impedances. Additionally, the method has demonstrated reliable performance despite noisy conditions, maintaining accuracy with a signal to noise ratio of 40 dB. https://joape.uma.ac.ir/article_4262_3878a9fbeda4648f6057ec095af9133f.pdf