More and more distributed power generators (DG), e.g., photovoltaic (PV), and various energy storage (ES) equipment are integrated into distribution network (DN). Integrated sources enable improved power supply reliability. After the upstream device fails, the islanding scheme can restore power to the downstream DN. However, in the process of formulating the islanding scheme, complex constraints should be considered, such as ES sequential characteristics and secondary outage constraints for load requirements. These constraint couplings make power recovery complex in DN. A method for DN island partition with ES is proposed. The islanding problem is a hybrid integer backpack model that include two key steps: the largest island is quickly identified by checking whether the power supply is sufficient, and the identified islands are corrected by checking whether the power supply meets real-time balance. Inspired greedy algorithms solve previous sub-problems. The latter sub-problem can be converted to a mixed-integer linear planning model for easy solution. By solving these two sub-problems alternately, it can quickly formulate islanding schemes with clear physical significance. A case study is carried out on an improved PG&E 69-bus system. It has been verified that the method can significantly reduce calculation time under the premise of the island partition benefit.
The modular multilevel converter (MMC) HVDC grid is a prospective technical route to unravel the high permeability of the renewable energy grid. The protection issues of MMC-HVDC grids are the main obstacle to its further development. In particular, an accurate and fast fault location scheme is essential to speeding up the time of fault clearing and power supply interruption. In this paper, a non-unit fault location scheme, that can be employed as a distance protection criterion for MMC-HVDC grids, is proposed based on the geometric features of the fault voltage. First, the fault transient characteristics are analyzed by solving the built equivalent circuit of MMC-HVDC grids after DC line faults. The multi-refection and refraction of the fault voltages between the line terminal and the fault point are evaluated. A time domain multi-resolution morphological gradient algorithm (MMGA) is utilized to obtain the geometric characteristics of the fault waveform. The change direction and the corresponding time of the initial two heads of the fault voltage can be recorded, and together with the estimated wave velocity, the search region of the target wave head with definite change direction can be obtained. Therefore, the proposed scheme is almost immune to the velocity of fault waves, fault resistances, and fault positions. A four-terminal +/- 500 kV MMC-HVDC grid built in PSCAD/EMTDC demonstrates the higher performance of the proposed scheme under different fault conditions, size of fault current limiting reactor, sampling frequency, and noise environments.
This paper proposes a novel method for depicting the dynamic characteristics of a hybrid AC/DC power distribution system with a high proportion of renewable energy sources (RES). It describes its reachable regions under multiple uncertainties. By modeling multiple uncertainties based on zonotopes and deriving differential-algebraic equations of the system, a reachability analysis method for the hybrid AC/DC power distribution system is realized. The effectiveness of the proposed method is verified by comparing the results with those from Matlab/Simulink time-domain simulations and Monte Carlo simulations. It can obtain all possible reachable regions of the key state variables under uncertain disturbances in one calculation, thus quickly determining the dynamic characteristics of the system without numerous sampling calculations.
This paper presents an adaptive staged overcurrent protection system based on predicted post-fault Thevenin equivalent parameters. The presented protection scheme can be used in a distribution system with more than one Inverter Interfaced Distribution Generators (IIDGs). To implement the protection system, equations of pre-fault and post-fault Thevenin equivalent parameters are obtained and relationship between them is revealed. Then protection's stage I and stage II trip currents are calculated by an easy algorithm based on that relationship. What's more, a sensitivity verification is applied to the Stage II operating current, enabling both Stage I and Stage II to protect the entire line, similar to conventional overcurrent (OC) protection. To verify equations and ensure protection system working normally, a distribution system is built in PSCAD/EMTDC. Simulation and calculation results show that the sensitivity and the selectivity of the protection system satisfy the need of an OC protection.
In order to relieve the greenhouse effect caused by SF6, using SF6/N-2 mixture gas instead of pure SF6 as the insulation medium of GIS is an effective technical means. However, there is still a lack of research on the over-thermal decomposition characteristics of SF6/N-2 at present, let alone its over-thermal decomposition mechanism. Therefore, in this paper, to explore the decomposition characteristics of SF6/N-2 at different fault temperatures, a large number of partial over-thermal fault experiments are conducted in the range of 550 degrees C to 650 degrees C. The experiment results show that the initial formation temperature of NF3 is 575 degrees C, while S containing products (SO2, SOF2, SOF4 and SO2F2) are generated in large quantities under the same conditions, and nitrogen oxides (NO and NO2) are already generated at 550 degrees C. When the temperature is between 550 degrees C and 650 degrees C, around 1.3%-4.0% of the SF6 decomposed, while the decomposition percentage of N2is only 0.003 to 0.017. In addition, according to calculated thermodynamic and kinetic parameters of the reactions, which are related to the thermal decomposition of SF6/N-2, it can be found the orders of magnitude of the reaction rate constant of reactions between SF4, SF2 and impurities O-2, H2O are higher than 10(20), and the same holds true for the hydrolysis reaction of SOF2. It explains why yields of SO2 and SOF2are large during the experiment. As for nitrogenous products, the reaction rate constant of the formation of NOx is much higher than NF3, which reveals why the concentration of (NO+NO2) is larger than that of NF3 in the experiment.
