As a basic building block of graphene-based materials, graphene oxide (GO) plays an important role in scientific research and industrial applications. At present, numerous methods have been employed to synthesize GO, there are still some issues that need to be solved, thus it is of importance to develop a green, safe and low-cost GO preparation method. Herein, a green, safe and fast method was designed to prepare GO, namely, graphite powder was firstly oxidized in a dilute sulfuric acid solution (H₂SO₄, 6 mol/L) with hydrogen peroxide (H₂O₂, 30 wt%) as oxidant, and then exfoliated to GO by ultrasonic treatment in water. In this process, H₂O₂ was the only oxidant, and no other oxidants were used, thus the explosive nature of GO preparation reaction in the conventional methods could be completely eliminated. This method has other advantages such as green, fast, low-cost and no Mn-based residues. The experimental results confirm that obtained GO with oxygen-containing groups has better adsorption property compared to the graphite powder. As adsorbent, GO can remove methylene blue (50 mg/L) and Cd²⁺ (56.2 mg/L) from water with removal capacity of 23.8 mg/g and 24.7 mg/g, respectively. It provides a green, fast and low-cost method to prepare GO for some applications such as adsorbent.

A transformer is an important part of the power system. Existing transformer fault diagnosis methods are still limited by the accuracy and efficiency of the solution and excessively rely on manpower. In this paper, a novel neural network is designed to overcome this issue. Based on the traditional method of judging the ratio of dissolved gas in transformer internal insulation oil, a fast fault diagnosis model of a transformer was built with an improved probabilistic neural network (PNN). The particle swarm optimization (PSO) algorithm was used to find the global optimal smoothing factor and improve the fault diagnosis accuracy of PNN. The transformer fault diagnosis model based on improved PNN not only eliminates the influence of human subjective factors but also significantly improves the diagnosis speed and accuracy, meeting the requirements for real-time application in practical projects. The feasibility and effectiveness of the method proposed in this paper are illustrated by a case study of actual data. Through analysis and comparison, the diagnostic accuracy of the proposed method is 10% higher than that of the general BPNN and 5% higher than that of the traditional PNN on the premise of ensuring the efficiency of the solution.

Flexible resources such as adjustable load widely participate in interaction with power grid, which can effectively promote renewable energy consumption. In previous studies, researchers generally focused on industrial and household users, but usually ignored the manufacturing load. Therefore, in this paper, an day-ahead intelligent energy management strategy for manufacturing load is proposed. Firstly, we analyze the power demand behavior of manufacturing load in detail, and describe the energy flow and material flow of manufacturing load through state task network (STN) method and mixed integer linear programming model. Then, the conditional deep convolution generative adversarial networks (C-DCGAN) algorithm is used to describe the uncertainty of new energy and construct a set of scheduling scenarios. Finally, case study shows that the proposed method can effectively improve the regional renewable energy consumption level and economic benefits of enterprises.

A DC distribution network is an effective solution for increasing renewable energy utilization with distinct benefits, such as high efficiency and easy control. However, a sudden increase in the current after the occurrence of faults in the network may adversely affect network stability. This study proposes an artificial neural network (ANN)-based fault detection and protection method for DC distribution networks. The ANN is applied to a classifier for different faults on the DC line. The backpropagation neural network is used to predict the line current, and the fault detection threshold is obtained on the basis of the difference between the predicted current and the actual current. The proposed method only uses local signals, with no requirement of a strict communication link. Simulation experiments are conducted for the proposed algorithm on a two-terminal DC distribution network modeled in the PSCAD/EMTDC and developed on the MATLAB platform. The results confirm that the proposed method can accurately detect and classify line faults within a few milliseconds and is not affected by fault locations, fault resistance, noise, and communication delay.