Search

Non-intrusive load monitoring (NILM) has significant supporting effect on demand response. However, the mini-watt equipment, continuous operation equipment and variable state equipment have brought great difficulties to the identification accuracy. This paper proposes a method to identify load components based on the data correlation. This method is constructed via three steps: first, a physical electrical constraints model is applied to abstract the relationship between load customers and distribution network; second, a stepwise method with cross validation is used to select the most relevant features; and third, the load components are explained by the characteristics model and solved by combinatorial optimization method. Finally, the accuracy of the proposed algorithm is verified by comparison against the NILMTK algorithms through detailed simulations over high frequency engineering data.

with the development of information technology, automation and intellectualization of distribution network, distribution network has gradually developed into a typical cyber physical system (CPS). In distribution network CPS, the information flow and energy flow is mixed, the discrete state and the continuous process interweave, and the dynamic process of conduction mechanism between cyber system and physical system are complex, these brings difficulty for simulation of distribution network CPS. In this paper, the coupling relationship between physical system and cyber system in active distribution network CPS is analyzed. A distribution network communication service model of communication node, link and process based on OPNET are constructed. A framework of information flow and energy flow fusion calculation is proposed, and a multi-rate cooperative timing simulation method based on trigger synchronization is adopted. A prototype system based on distribution network simulation module and OPNET is realized. The effectiveness and correctness of the model and algorithm are verified by a practical example from a CPS demonstration project.

Distribution system lean management (DSLM) is becoming an essential element of many distribution applications, the distribution network measurements are serving as the core data foundation of numerous DSLM. Considering the communication cost, the most valuable measurement data should be selected and uploaded. This paper proposes a data-upload strategy (DUS) based on distribution system state estimation (DSSE), which considers both the network observability and the mesurements' contribution. The measurements are first sorted, and then the most valuable data set is selected by using a particle swarm optimization (PSO) and a greedy algorithm. The method is tested on IEEE 4 node test feeder, different data sets are studied under several given communication cost constraints, its results are always in the top 5% group, and the DUS’ result is also used on a period of time containing multipy time points, and is shown to have good stability. The proposed method is also applied to a IEEE 123 test system and works well.

Capacitors and taps of MV distribution systems are usually discrete optimization objects, a control sequence is needed to respond to the randomness of renewable distributed generations (DG), and many existing algorithms are trapped into the dimension house. A control strategy is proposed to minimize network losses, switching costs, generation wastes of DGs and voltage deviation, a weighted sum method is used to reflect the preferences on the mutually conflicting control objectives. For the mixed integer nonlinear programming (MINLP) problems of capacitors and taps control, a cooperative particle swarm optimization (Co-PSO) method employing cooperative behavior to significantly improve the optimizing ability is used. Finally, the correctness, convergence and robustness of the proposed algorithm are verified by comparison against existing popular schemes through detailed simulations over 24h on the IEEE 13-bus unbalanced distribution system.

Frequency converter occupies a large proportion in production and daily life, in order to enable it to effectively take part in the process of system frequency regulation, a control strategy based on the virtual synchronous motor technology (VSM) is proposed in this paper. In the inverter part, VSM control strategy is used, using the mechanical inertia and electrical characteristics of the synchronous motor, the power control and frequency control are researched and designed respectively. The frequency converter are simulated under environment of Simulink/Matlab. Simulation results show that the proposed control strategy can make the frequency converter follow the characteristics of VSM, can change the inverter’s output voltage and frequency with the change of motor load.

To effectively clean the large-scale, mixed and inaccurate monitoring or collective data, reduce the cost of data cache and ensure the consistent deviation detection on timing data of each cycle, a big data online cleaning algorithm based on dynamic outlier detection has been proposed. The data cleaning method is improved by local outliner detection upon density, sampling cluster uniformly dilution Euclidean distance matrix retaining some corrections into next cycle of cleaning, which avoids a sampling causing overall cleaning deviation and reduces amount of calculation within data cleaning stable time, enhancing the speed greatly. Finally, the distributed solutions on online cleaning algorithm based on Hadoop platform.