Generator start-up is a pivotal step of transmission system self-healing after large-scale blackouts. Considering the uncertainty initial power situation blackouts and line restoration during restoration, an online generator algorithm based on Monte Carlo tree search (MCTS) sparse autoencoder (SAE) proposed for real-time decision making. First, support efficiency indicator involving total generation capability number restored lines are proposed. Then, SAE deployed to learn data relevant...

Pre-fault dynamic security assessment (DSA) of power systems needs to consider different fault locations. Existing data-driven DSA methods lack generalizable location features, failing accurately assess unlearned locations, i.e., locations not in the training set. To address this issue, paper proposes a continuous feature representation method based on electrical distance, aiming construct pre-fault model Firstly, is represented continuously as an coordinate with distance preselected...

Transmission network self-healing considering uncertain wind power becomes crucial with increasing penetration of power. A hybrid reinforcement learning (HRL) method combining offline self-learning online Monte Carlo tree search (MCTS) is designed to deal the strong uncertainty induced by restoration. The HRL trains a policy based on historical and transmission system data. It then applies guide MCTS realize step-by-step real-time forecast data in different scenarios. Besides, model...

The objective of power system restoration is to restore the load as soon possible, while successful start-up generator precondition recovery and emphasis restoration. In this paper, both sequence path are considered. A preference multi-objective model which considers difference objectives' importance proposed. To get most preferred alternatives, a non-r-dominance sorting genetic algorithm II (r-NSGA-II) used solve model. reference point determination method solution set scale controlling...

Considering the uncertainty of initial power network topology, restoration time lines and downtime generators during generator start-up, this paper presents an on-line decision making strategy start-up based on deep learning Monte Carlo tree search (MCTS). At first, in order to generate labeled samples atomically, a self-generated method is proposed. The sparse autoencoder (SAE) applied train establish value network, which used estimate optimal indicator under certain circumstances. Then,...

The additional power demand caused by cold-load pickup restricts the simultaneous restoration of all loads a substation due to excessive loading on network elements and violation limits. In proposed approach, load amount can meet system requirements based direct control TCLs during pickup. At first, post-outage is evaluated aggregated estimation method considering outage time. Then, transient safety constraints available recovery system, an optimization model established calculate maximum...

Load restoration coordinating transmission grid, distribution and microgrids is an effective measure that taken into consideration while improving the power system resilience in extreme weather conditions. An online decision-making method proposed to deal with unexpected nature of supply issues regarding re-energization grids. In this research work, multi-agent interaction technique used for coordinated load restoration. The main algorithm comprises two subsections, namely, a index...

Power network reconfiguration is a complex non-convex, nonsmooth and nonlinear optimization problem. A preference-based multiobjective evolutionary algorithm proposed to incorporate the preference for different objectives optimization. Three about generators, lines loads are establish model. To handle high discreteness of suggest model, discrete nondominated sorting genetic II (PD-NSGA-II) designed, with which solutions required quantity quality obtained. The simulation results demonstrate...

Extreme disasters seriously threaten power system security and stable operation. Power resilience is an important means to evaluate the response low probability high impact events. This paper proposes a enhancement strategy based on cyber physical cope with extreme natural restore system. At first, defense framework for disaster given show stages techniques used in strategy. Secondly, zone division flow adjustment method proposed adjust transmission failure probability, which can reduce risk...

An assessment method of cascading failure is proposed to acquire high-risk failures chains under high penetration renewable energy power system. At first, a dynamic simulation model considering the controller DFIG established, and simplified LVRT for response model. Then, risk index defined depth first tree search used in searching proceeding. Finally, based on flow presented, which can simulate interactive between grid failure. The effectiveness verified by simulations concerning modified...

Efficient power transmission system restoration after natural disasters contributes to resilience enhancement and reliable energy supply. The destruction of cyber physical equipment caused by leads complex unpredictable outages. To deal with outages damage, an online collaborative decision-making method coordinating damage repair dispatch is proposed for rapid restoration. Reinforcement learning mathematical programming are combined realize domain knowledge-guided intelligent search...

Line Commutated Converter based-high voltage direct current (LCC-HVDC) systems are an important support for large-scale renewable energy transmission. Compared with two-terminal LCC-HVDC, multi-terminal LCC-HVDC flexible in control and reliable power supply. However, the integration of brings great complexity to operation modes flow systems. Therefore, it is very resolve calculation problems AC/DC hybrid containing LCC-HVDC. This paper proposed a sequential algorithm Firstly, steady-state...

A rapid dynamic security assessment (DSA) is crucial for online preventive and restoration decision-making. The deep learning-based DSA models have high efficiency accuracy. However, the complex model structure training cost make them hard to update quickly. This paper proposes a partition method, aiming develop accurate incrementally updated with simple structures. Firstly, power grid self-adaptively partitioned into several local regions based on mean shift algorithm. input of algorithm...

Abstract Successive commutation failure seriously threatens the safe operation of large AC/DC power grids, which may lead to cascading failures across regional grids. Firstly, from perspective reactive exchange in systems, cause successive is analyzed. Then, taking a simplified three-zone interconnected system as an example, variation relative angle synchronous generator AC during process analyzed based on equal-area criterion. Finally, simulation analysis conducted changes when occurs at...

Load restoration fitting to the change of load amount, available power and outdoor temperature is an important measure enhance system resilience under extreme hot weather. A multi-time scale substation method considering active control air conditioners proposed for adaptive restoration. First, a framework formulated coordinate longtime whole process scheme making short-time control. Second, optimization model established make feeder switch schemes improve efficiency from perspective...