In the context of Integrated Energy System (IES), accurate short-term power demand forecasting is crucial for ensuring system reliability, optimizing operational efficiency through resource allocation, and supporting effective real-time decision-making in energy management. However, achieving high forecasting accuracy faces significant challenges due to the inherent complexity and stochastic nature of IES’s short-term load profiles, resulting from diverse consumption patterns among end-users and the intricate coupling within the network of interconnected energy sources. To address this issue, a dedicated Short-Term Power Load Forecasting (STPLF) framework for IES is proposed, which relies on a newly developed hybrid deep learning architecture. The framework seamlessly combines Long Short-Term Memory (LSTM) with Temporal Convolutional Network (TCN), enhanced by an attention mechanism module. By merging these methodologies, the network leverages the parallel processing prowess of TCN alongside LSTM’s ability to retain long-range temporal information, thus enabling it to dynamically concentrate on relevant sections of time series data. This synergy leads to improved prediction accuracy and broader applicability. Furthermore, the integration of residual connections within the network structure serves to deepen its learning capabilities and enhance overall performance. Ultimately, results from a real case study of a user-level IES demonstrate that the Mean Absolute Percentage Error (MAPE) of the proposed framework on the test set is 2.35%. This error rate is lower than the averages of traditional methods (3.43%) and uncombined single submodules (2.80%).

The single pole-to-ground (SPG) fault is one of critical failures which will have a serious impact on the stable operation of the multi-terminal DC distribution network based on the modular multilevel converter (MMC). It is very significant to analyze fault characteristics for detecting faults and protection design. This paper established the DC SPG fault model, which showed that in the presence of a reactor, the short-circuit current was reduced from 2.3 kA to 1 kA at 6 ms after the fault. Then, a novel SPG fault protection strategy was proposed, which detected the current derivative in connection transformer grounding branch. When the value increases past the threshold of current derivative, small resistance was switched on to increase fault current. Thus, the reliability of differential protection was enhanced. Compared with the traditional protection method, the proposed method does not need communication, and improved the speed of protection. Finally, the simulation model was established in PSCAD/EMTDC. The model included three converter stations: T1, T2 and T3. Among them, T1 outputs power, and T2 and T3 receive power. The results of RTDS showed that the DC circuit breaker operated within 3 ms, the three-port circuit breaker worked within 50 ms, which proves that the proposed strategy was effective. At this time, the system switched from the T1–T2–T3 three-terminal networking operation mode to the T1–T2 two-terminal hand-in-hand operation mode. Since the T3 terminal no longer received power, the transmission power of the T1 terminal decreased, and the received power of the T2 terminal remained unchanged.

With the increased complexity of power systems stemming from the connection of high-proportion renewable energy sources, coupled with the escalating volatility and uncertainty, the key transmission sections that serve as indicators of the power grid’s security status are also subject to frequent changes, posing challenges to grid monitoring. The search method for key transmission sections based on an improved spectral clustering algorithm is proposed in this paper. A branch weight model, considering the impact of node voltage and power flow factors, is initially established to comprehensively reflect the electrical connectivity between nodes. Subsequently, a weighted graph model is constructed based on spectral graph theory, and an improved spectral clustering algorithm is employed to partition the power grid. Finally, a safety risk indicator is utilized to identify whether the partitioned sections are key transmission sections. Results from case studies on the IEEE39-node system and actual power grid examples demonstrate that the proposed method accurately and effectively searches for all key transmission sections of the system and identifies their security risks. The application in real power grid scenarios validates its ability to screen out some previously unrecognized key transmission sections.