Soiling in photovoltaic systems is a recurring problem that reduces energy generation and demands efficient operation and maintenance (O&M) strategies. In this context, this paper proposes a machine learning-based approach to identify dirt levels and generate cleaning alerts using operational and weather data. Initially, the models were evaluated with a decision threshold ranging from 0.5 to 0.7, using only operational features. Subsequently, the inclusion of weather features was tested, which improved the models’ performance and enabled the selection of the best models for the exhaustive features search step. The models analyzed in this step were Extra Trees, Histogram-based Gradient Boosting, Extreme Gradient Boosting, and Random Forest. Exhaustive analysis further improved model performance, as indicated by global metrics and ROC curves. The Extra Trees model with a threshold of 0.5 showed the best performance and was selected as the final configuration, achieving an accuracy of 0.9884 and an AUC-ROC of 0.9957. Finally, the selected model was applied to determine daily soiling levels and trigger alerts based on temporal persistence, indicating its potential to support predictive O&M decisions and cleaning actions in PV systems.

The DLMS (Device Language Message Specification) standard is widely adopted in smart metering systems for utility services, as it defines the mechanisms that enable the standardized and interoperable exchange of data between metering devices. Wi-SUN FAN (Wireless Smart Ubiquitous Network Field Area Network) is a wireless communication standard that has gained increasing attention in the field of smart metering networks, due to its capability to operate in mesh topologies and support multi-hop communications in an efficient and scalable manner. To date, no studies have been reported that evaluate the combined performance of the DLMS and Wi-SUN FAN standards. In this context, this work evaluates the performance of the DLMS standard over a Wi-SUN FAN network through detailed simulations conducted using the Contiki-NG/Cooja platform. The study implements the DLMS data transmission process and evaluates key performance metrics such as Packet Delivery Rate (PDR) and latency, considering both linear and mesh network topologies. The results demonstrate that a Wi-SUN FAN network can provide efficient and reliable communication services, achieving PDR values above 86.02% in both topologies, thereby confirming its feasibility for DLMS-based smart metering applications.

The Wi-SUN FAN (Wireless Smart Ubiquitous Network Field Area Network) standard facilitates large-scale connectivity among smart devices in utility networks and smart cities. Specifically designed for Low-Power and Lossy Networks (LLNs), Wi-SUN FAN supports the formation of multiple Personal Area Networks (PANs) and mesh topologies with multi-hop transmissions. However, the node association process, divided into five junction states, often results in prolonged connection times, particularly in multi-hop networks, thereby limiting network scalability and reliability. This study analyzes the factors affecting these delays, with a particular focus on Join State 1 (JS1), which relies on PAN Advertisement (PA) packets that use asynchronous communication and the trickle timer algorithm, frequently causing significant delays. To overcome this challenge in JS1, we propose the Parallel Rendezvous (PR) strategy, which forms synchronized clusters of unassociated nodes and leverages the standard’s PAN Advertisement Solicit (PAS) packets to rapidly disseminate network information. The proposed algorithm, PR Wi-SUN FAN, is evaluated through simulations in various network topologies, demonstrating notable improvements in linear, fully connected, and mesh scenarios. The most significant gains are observed in the linear topology, with reductions of up to 71.22% in association time and 59.56% in energy consumption during JS1.