The networking of multiple DC microgrids under DC distribution network will become an effective means to accept largescale distributed energy and load with DC characteristics. The stability research of multivoltage levels DC distribution system is an important issue in its design and operation. Therefore, a stability analysis method suitable for multi voltage levels DC distribution system with multi DC microgrids was proposed. Firstly, the equivalent admittance of each microgrid subsystem is derived according to the control mode of the interconnected converter, and the equivalent openloop gain when each microgrid operates alone is used to judge whether there are right half plane poles in its equivalent admittance; Secondly, the multi voltage level DC distribution system is further simplified into a single voltage level DC system containing only medium voltage bus, and the stability of the medium voltage side subsystem is judged by the equivalent impedance ratio. When and only when the equivalent openloop gain of each microgrid and the equivalent impedance ratio of the medium voltage side of the system meet the Nyquist criterion, the system can operate stably; Finally, based on PSCAD / EMTDC simulation platform, a multi voltage level DC distribution system including two DC microgrids is built for verification. The simulation results show that the proposed stability analysis method can accurately determine the stability of multi voltage DC distribution system with multi DC microgrid.

With the deregulation of the power sector, the power exchange between regions of the power system becomes more frequent, and the stability of the new energy power system is also affected. In order to solve the regional power market power trading and denial of service attacks on the interference of wind power system, a datadriven load frequency control method is designed in this paper. An improved modelfree adaptive control algorithm is designed by collecting the input and output data and estimating the pseudopartial derivative using the compact form dynamic linearization, the antisaturation compensation control algorithm is introduced to solve the output disfollowing problem caused by input saturation, and the stability of the control algorithm is proved theoretically. Finally, taking threeregion wind power system as an example, the validity of the proposed algorithm is verified by MATLAB simulation.

Droop control is a common current sharing method for parallel converters in DC microgrid. However, due to the inconsistency of line parameters and the sampling error of sensors, the current distribution accuracy of traditional droop control is low. In order to solve this problem, this paper proposes a parallel current sharing strategy based on AC signal injection. Firstly, by superimposing an AC voltage small signal on the output voltage of the converter, the droop characteristic between the frequency of the AC voltage and the output current of the converter is constructed, and the reactive power generated by the signal and the feedback mechanism are used to realize the accurate current sharing of the converter. Secondly, the compensation virtual resistance is introduced to improve the stability of the system when the load changes greatly. Then, the design mode switching link stops the injection of AC signals in the steady state, so that the power quality can be improved. Finally, the effectiveness of the proposed control strategy is verified by simulation results.

With the largescale integration of clean energy sources such as photovoltaics and energy storage into the power grid, grid type control technology has obvious advantages in dealing with voltage stability issues in new energy power systems that lack synchronization. However, how to adaptively control the parameters of grid type photovoltaic storage inverters to maintain voltage stability even when the impedance of the power grid changes is an urgent problem that needs to be solved. Based on this, a method for optimizing the control of optical storage grid inverters using a convolutional neural network optimized by the gazelle algorithm is proposed. Firstly, build a control model for grid type inverters and analyze the stability of output voltage; Secondly, based on the convolutional neural network, an inverter parameter control model is established, and the Gazelle optimization algorithm is utilized to optimize the hyperparameters of the convolutional neural network with strong optimization ability and fast search speed, improving the model's feature learning ability and outputting inverter control parameters; Finally, a certain photovoltaic power generation area was selected for simulation verification. The experiment showed that the proposed grid type photovoltaic inverter control method can adaptively optimize control parameters based on realtime changes in grid impedance, achieve stable voltage output, and have strong practical engineering significance.

In Active Distribution Network (ADN), the penetration rate of Renewable Energy Sources (RES) is continuously increasing, leading to more complex and uncertain operational scenarios. This complexity introduces significant risks in the daily operations of ADN. This study proposes a collaborative configuration of distributed power sources within ADN to enhance the absorption capacity for renewable power. The proposed model thoroughly considers the variability of RES, the characteristics of adjustable demand response resources, the bidirectional flow of ADN, and the constraints of safe operation. To address the contradiction between the effective absorption of renewable energy and the economic operation of ADNs, this paper introduces a multiobjective Bayesian optimization algorithm based on hyperspace indicators (EBO). This method probabilistically models multiple objective functions, effectively balancing the exploration of solution space and the unidirectionality of optimization. Moreover, its computational efficiency surpasses traditional heuristicbased multiobjective planning algorithms.

With the development of new power system construction, various energy sources are integrated into the distribution network in the form of clusters and networks, making the distribution network exhibit active characteristics. Fully leveraging the complementary characteristics of multiple energy sources can alleviate the uncertainty of the distribution network while providing sufficient demand for electricity, heat, gas, cooling, heating, and other resources on the demand side. The existing distributed coordinated control methods only consider the gas network, heat network, and power network, without considering the hydrogen energy network. Provide a comprehensive and distributed coordinated control method for active distribution network, heating network, natural gas network, and hydrogen energy network, as well as energy interconnection. Firstly, the economic goal of minimizing operating costs was established, and the constraints of each network were provided; Secondly, a hierarchical control strategy is proposed for the active distribution network, heating network, natural gas network, and hydrogen energy network, and an improved consistency algorithm and particle swarm optimization method are used for solution comparison; Finally, a simulation example of an actual power grid is given, and it is found that the distributed comprehensive energy interconnection scheduling control method can achieve the same results as traditional methods, and the effectiveness of the proposed scheduling control strategy is verified.

Aiming at the impact of thermal stratification on storage, the Fourzone storage model is proposed and Helmholtz energy equation is used to calculate the thermal properties of hydrogen in the study. According to different experimental conditions, the Nusselt number correlation formula is proposed, and the heat transfer coefficient of condensation surface is modified, and the modified Fourzone model is built. Compared with the experimental data, the error of the modified model is less than 3%, which has higher accuracy. On this basis, the influence of different filling ratio on selfpressurized storage is studied. The result shows that with the increase of filling ratio, the selfpressurized rate of storage tank first slows down and then becomes faster. There is an optimal filling ratio when liquid hydrogen is stored at low temperature, which makes the safe storage time the longest.

To achieve efficient advanced nitrogen and phosphorus removal from wastewater while simultaneously recovering energy, denitrifying phosphorus removal and electrogenesis (DPRE) system was constructed to research the influence of mass concentration of organic substance on nitrogen and phosphorus removal and electricitygenerating performance using synthetic domestic wastewater. The results showed that the influent mass concentration of organic substance had a minor effect on COD removal but significantly influenced nitrogen and phosphorus removal and power generation. When the influent COD concentration was 200~300 mg/L, the DPRE system achieved optimal pollutant removal, with the removal rates of COD, NH4+N and PO43P reaching 86.07% ~86.22%, 83.94%~85.10% and 80.29%~83.38%, respectively. When the influent COD concentration was 300~400 mg/L, the system exhibited the best electricitygenerating performance, with an average power density of 36.49~39.47 mW/m². When the influent COD concentration was 300 mg/L, the DPRE system could simultaneously obtain highefficiency removal of nitrogen and phosphorus and electricity generation.

In this study, a multiphysical field coupling model of metal hydride hydrogen storage reactor (MHHSR) based on cylindrical heat exchanger was established. The influence of the geometric shape and position of the cylindrical heat exchanger on the hydrogen absorption performance of the reactor was investigated, and the mathematical model was developed. The optimal position of the heat transfer structure was obtained, and the characteristics and intrinsic mechanisms of heat and mass transfer in the alloy bed during the hydrogen absorption process were explored. Additionally, based on the area of the temperature differential zones among different layers, the uniformity of heat transfer in multilayer beds was analyzed. The research results showed that when the embedded heat transfer ring was located at 0.62R of the alloy bed, the hydrogen storage reactor achieved 90% hydrogen capacity within the shortest time. By comparison to the central heat exchange tube structure and the external heat exchange jacket structure, there was a time reduction of 76.3% and 60.7%, respectively. Different types of heat exchanger structures caused differences in the thermal mass transfer characteristics of the alloy bed, which changed the evolution modes of the bed's reaction interface area and moving speed, ultimately affecting the reactor's hydrogen absorption performance. When multiple independent reaction bed layers existed in the reactor, a smaller temperature difference region area among different bed layers resulted in more uniform heat and mass transfer and higher energy efficiency of heat exchanger structures.

Carbonbased nanofluid has a higher heat transfer coefficient than other nanofluids and is relatively cheap. Graphene nanofluid has a high heat transfer coefficient, and reduced graphene oxide nanofluid has good stability. So, this article developed a new type of carbonbased composite nanofluid (waterbased graphene/reduced graphene oxide composite nanofluid) as the heat transfer medium in PV/T collectors by twostep method. The radiation intensity, flow rate, and composite nanofluid were analyzed. The effect of proportion on the water temperature at the outlet of the collector was studied, and the proportion of waterbased graphene/reduced graphene oxide with better heat transfer performance was obtained in different compound proportion ranges of the experiment. Experimental results show: the higher the radiation intensity, the faster the flow rate, the higher the water temperature at the outlet of the PV/T collector, the better the heat transfer effect of the nanofluid; when the composite ratio is 8:2, the heat transfer effect of the nanofluid is the best, the thermal efficiency and electrical efficiency of the PV/T system are 23.49% and 20.18% respectively. Compared with waterbased graphene and reduced graphene oxide, nanofluid has better heat transfer effect. The research of this article provides reference for the practical development of carbon nanofluids in heat transfer media.