Distributed photovoltaic (PV) systems often lack adequate measurements due to cost considerations, which makes it very difficult to predict them accurately. Here, an approach is proposed for behind‐the‐meter (BTM) PV power prediction using attention‐LSTM neural network and transfer learning. First, the weather is classified into four types based on the deviation ratio β. Second, the correlation analysis algorithm identifies the weather factors that contribute the most to PV power generation as GHI, DNI, humidity, and temperature. Then, attention‐LSTM pre‐trained model is constructed, in which the LSTM network fully extracts the temporal characteristics of PV power generation, while the attention mechanism enhances the attention to the important information in the input. Finally, a novel BTM PV short‐term power prediction method using transfer learning that the upper layer parameters of the pre‐trained model were frozen, and a small amount of DP‐NB data was used to fine‐tune the model. The proposed methodology has higher prediction accuracy compared with other benchmark methods under four weather types, and the time cost is saved by freezing the parameters of the first layer of the network by 37.2%.

This paper proposes a dual-loop control strategy to address the trajectory tracking problem of differential wheeled mobile robots (WMRs). First, the kinematic model of the WMR is established, and the dynamic model including the actuators is derived. To tackle the issue of y-axis direction divergence in existing methods, a predefined-time velocity control law based on intermediate variables is proposed. By introducing the y-axis error term into the angular velocity control, the ability to rapidly track the target trajectory is enhanced, providing a reliable velocity tracking target for the dynamic controller. Furthermore, a predefined-time nonsingular fast terminal sliding mode controller is designed, which combines a nonsingular fast terminal sliding surface with predefined-time stability theory to overcome the singularity problem in existing approaches, achieving fast and accurate tracking of velocity errors. Additionally, to improve the system’s disturbance rejection capability, a nonlinear extended state observer (NESO) is proposed to estimate external disturbances and provide feedforward compensation to the dynamic controller. Experimental results demonstrate that the proposed strategy outperforms existing methods in terms of trajectory tracking accuracy and robustness, providing an effective solution for the high-performance control of WMRs.

This paper proposes a bi-level optimization scheduling strategy for integrated photovoltaic (PV) and energy storage systems (ESS) to meet electric vehicle (EV) charging demands while reducing charging costs. First, a battery degradation cost model is developed in order to convert the long-term costs into short-term costs for real-time operation. The upper layer of ESS and power grid operation strategies are obtained by minimizing costs associated with battery degradation and distribution grid costs. The lower layer considers the PV uncertainty and the error caused by the upper layer operation strategy, and obtains the lower layer operation strategy by adding a penalty function to minimize fluctuations in power. Second, the author proposes a global optimization algorithm that combines Particle Swarm Optimization (PSO) and Sequential Quadratic Programming (SQP) in order to solve the above-mentioned models, effectively combining the global search feature of PSO with the local search capability of SQP. Finally, the bi-level optimization scheduling strategy is obtained by solving the model through the algorithm. Simulation results verify the practicality of the scheduling strategy and the effectiveness of the proposed algorithm.

A short-term photovoltaic power forecasting method is proposed, integrating variational mode decomposition (VMD), an improved dung beetle algorithm (IDBO), and a deep hybrid kernel extreme learning machine (DHKELM). First, the weather factors less relevant to photovoltaic (PV) power generation are filtered using the Spearman correlation coefficient. Historical data are then clustered into three categories—sunny, cloudy, and rainy days—using the K-means algorithm. Next, the original PV power data are decomposed through VMD. A DHKELM-based combined prediction model is developed for each component of the decomposition, tailored to different weather types. The model’s hyperparameters are optimized using the IDBO. The final power forecast is determined by combining the outcomes of each individual component. Validation is performed using actual data from a PV power plant in Australia and a PV power station in Kashgar, China demonstrates. Numerical evaluation results show that the proposed method improves the Mean Absolute Error (MAE) by 3.84% and the Root-Mean-Squared Error (RMSE) by 3.38%, confirming its accuracy.