English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

With the adjustment of China's energy structure and the development of smart grid technology, the ability to adjust the flexible load on the power demand side plays an important role in the future development of the power system. Therefore, this paper proposes a scaled flexible load scheduling platform architecture based on Spark technology. The architecture consists of five layers, which can not only support the business needs of huge real-time data, but also be compatible with existing computer resources. The validity and efficiency of the platform are verified by a case study. Introduction In recent years, in response to the demand side response requirements in the power system, combined with the “Internet +” concept, the concept of flexible load on the user side and the establishment of virtual power plants for power system load regulation are proposed. In order to effectively promote the deep integration of “Internet +” and smart grid, a large number of smart meters and related sensor devices and smart home devices are needed to measure the power consumption of users on the demand side and transmit them to the data center. Power System Traditional Data Processing Platform Architecture In the smart grid environment, the amount of data generated by China's power demand side has increased dramatically, and the generation of data has been distributed. Faced with these massive and distributed power load data, horizontal expansion of the system should be considered, such as distributed system expansion, distributing the data processing pressure of the system to multiple server nodes, and processing the data in parallel. User-Side Scaled Flexible Load Scheduling Platform Architecture and Basic Workflow Platform Architecture Aiming at the characteristics of large-scale flexible load on the power demand side, combined with the current situation of power system basic settings, and using the Internet of Things technology and Spark memory computing technology, a scaled flexible load platform architecture based on Spark technology is proposed. Flexible Load Scheduling Model The overall process of the flexible load scheduling of the scheduling platform is as follows: the user submits his own power load data and power consumption history data to a load aggregator, and the load aggregator preprocesses and aggregates load data of multiple users. The data in the standard data format is uploaded to the scheduling platform. Here, a large number of heterogeneous data streams are pre-processed and integrated, and the scheduling platform integrates the data and historical data generated by each aggregator into Hive and HBase by using sqoop extraction. Use Datanucleus to operate on the column store database and write online data generated by cloud-based applications to HBase. K-Means' clustering algorithm clusters the schedulable flexible load data of each user from the load aggregator, extracts the flexible load characteristics of each user and obtains the current time period schedulable according to the predicted load group power demand. The total amount of flexible load is fed back to the load aggregator. Definition The curve of the flexible load value submitted by the user at consecutive m time points in a day is called the flexible load characteristic curve, which is recorded as ) , , , ( 2 1 im i i i x x x K ... = , ik is the kth time point, which is the flexible load value at the kth time point, k=1, 2, ..., m. The flexible load curve for each user is a sample in the algorithm of this paper. Here the distance between the flexible load curves is defined as: m k x x D jk ik ij , , 2 , 1 |}, max{| ... = − = The distance between the two curves is the maximum of the difference in the value of the flexible load at each measurement point on the two curves. The smaller the distance ij D , the higher the similarity between the curve and the curve, and vice versa. Curves with similar precision within a certain range are classified into one curve class, that is, the same result class. In the flexible load dispatching process, the power supply capability data on the power supply side is wind power generation data. In this paper, the normal distribution is used as the distribution model of wind power power prediction error. The assumption that the 2 σ in the normal distribution error depends on the predicted time scale, then the supply side The power expectation is: ) 2 )) ( ( exp( 2 1 ) ( 2 2 σ πσ t P x t P f − − = (1) Here ) (t Pf is the predicted output power. Combined with the schedulable flexible load given by the load aggregator, the user's flexible load is divided into a transferable load, a translatable load, and a load reduction. Taking the response price as an example, the platform only considers the transferable load and the load reduction that are closely related to the user's life. The transferable load can be expressed as: )) ( ), ( ), ( ), ( ( ) ( 0 1 t v t k t p t P f t P shift shift shift shift ∆ = ∆ (2) ∑ = ∆ 0 ) (t Pshift The load reduction can be expressed as: )) ( ), ( ), ( ), ( ( 2 ) ( 0 t v t k t p t P f t P re re re re ∆ = ∆ (3) Under the premise of not affecting the user's electricity experience, the user's power consumption at the peak time is minimized. The objective function can be expressed as:

Research on Scaled Flexible Load Scheduling Platform Architecture Based on Spark Technology