In recent years, the peak–valley differences in urban power loads have been increasing. It is difficult to maintain the real-time balance of a power system by relying solely on the generation-side resources. As a typical flexible load, an air conditioning load can balance the supply and demand of a power grid by adjusting power using the thermal inertia of buildings. From the perspective of a load aggregator, this study models and aggregates the dispatch of a single inverter air conditioner distributed in a region to determine the adjustment potential of an air conditioning cluster. Then, according to the demand response capacity requirements, an optimal strategy for the aggregate dispatch of an inverter air conditioner considering incentive compensation measures is proposed with the objective of maximizing the load quotient economic benefit. The sensitivity analysis of the compensation factor for temperature rise is also performed. The results show that 3000 inverter air conditioners in the load quotient dispatch area participate in the demand response for 4 h, with a load reduction of 1.267 MW and a net income of RMB 14,435.97. Secondly, an increase in the temperature rise compensation factor will reduce the cost of temperature rise compensation by the loader to the user, but it will also reduce the load reduction and the net income of the loader. This study has practical significance for load aggregators to formulate compensation strategies and improve the economic benefits of participating in demand response.

With the goal of achieving “carbon peak in 2030 and carbon neutrality in 2060”, as clearly proposed by China, the transportation sector will face long–term pressure on carbon emissions, and the application of hydrogen fuel cell vehicles will usher in a rapid growth period. However, true “zero carbon” emissions cannot be separated from “green hydrogen”. Therefore, it is of practical significance to explore the feasibility of renewable energy hydrogen production in the context of hydrogen refueling stations, especially photovoltaic hydrogen production, which is applied to hydrogen refueling stations (hereinafter referred to “photovoltaic hydrogen refueling stations”). This paper takes a hydrogen refueling station in Shanghai with a supply capacity of 500 kg/day as the research object. Based on a characteristic analysis of the hydrogen demand of the hydrogen refueling station throughout the day, this paper studies and analyzes the system configuration, operation strategy, environmental effects, and economics of the photovoltaic hydrogen refueling station. It is estimated that when the hydrogen price is no less than 6.23 USD, the photovoltaic hydrogen refueling station has good economic benefits. Additionally, compared with the conventional hydrogen refueling station, it can reduce carbon emissions by approximately 1237.28 tons per year, with good environmental benefits.

With the rapid advancements in AI technology, UAV-based inspection has become a mainstream method for intelligent maintenance of PV power stations. To address limitations in accuracy and data acquisition, this paper presents a defect detection algorithm for PV panels based on an enhanced YOLOv8 model. The PV panel dust dataset is manually extended using 3D modeling technology, which significantly improves the model’s ability to generalize and detect fine dust particles in complex environments. SENetV2 is introduced to improve the model’s perception of dust features in cluttered backgrounds. AKConv replaces traditional convolution in the neck network, allowing for more flexible and accurate feature extraction through arbitrary kernel parameters and sampling shapes. Additionally, a DySample dynamic upsampler accelerates processing by 8.73%, improving the frame rate from 87.58 FPS to 95.23 FPS while maintaining efficiency. Experimental results show that the 3D image expansion method contributes to a 4.6% increase in detection accuracy, an 8.4% improvement in recall, a 5.7% increase in mAP@50, and a 15.1% improvement in mAP@50-95 compared to the original YOLOv8. The expanded dataset and enhanced model demonstrate the effectiveness and practicality of the proposed approach.