Power system in China has gradually stepped from the past era of data growing relatively slower in relatively simpler type into the era of big data that are massive, multi-source, heterogeneous, complex, under distribution control and within dynamic inline. As an important support for future smart grid construction,on the basis to combine with the existing building capability, big data platform fully integrates computing power, storage resources, data resources of enterprise internal system, etc., sharply raises operation capability of power system and maximize data value of enterprise. The construction of big data platform will help power enterprises to cope with the challenges of the big data era and provide strong support for the construction of big data of power enterprises.

At present, the ultimate goal of smart power grids in our country is to construct panoramic real-time system covering the whole production process of power system, including multiple links such as power generation, power transmission, power transformation, power distribution, power utilization as well as power dispatching, etc. The foundation to support safe and reliable operation of the smart grid is panoramic real-time data collection, transmission and storage of the power grid, as well as the rapid analysis of accumulated massive and multi-source data. Therefore, along with continuous and further progressive process of smart grid construction, the data, generated from power grid operation and equipment inspection/monitoring, gradually growing in exponential order, contribute to big data concerned by current circle of information science, which need corresponding storage and rapid processing technology as support.

The unified construction of big data platform can satisfy overall operation of the internal data of the state grid and analysis and exploration of laws in the professional field, and promote lean management; realize comprehensively monitoring internal and external management of power grid, improve insight into current operation situation of power grid, timely adjust strategic direction, and promote healthy and scientific development of the power industry. Big data platform construction is conductive to efficiently using historical data and real-time data of production and management accumulated by SGCC for many years, forming index system and dig deeper value, providing a large number of high value-added services, generating innovative business and meeting the demand of electric power business, improving the company's core competitiveness, so as to enhance influence of the company both inside and outside the industry.

The construction of base platform project of power big data that CEPRI promotes by efforts closely embraces application results of all kinds of big data advanced technologies, mainly completes module function optimization, improvement of platform depth and performance including function optimization such as unified foreign service interface, work flow control, etc., permissions deep integration such as access integration, component integration, etc., enhancement of performance and accessibility, etc, to provide public big data infrastructure services for various professions of CEPRI, such as data fusion, data analysis and access to big data applications system and so on. Specific requirements are as follows:

1. Uniform external service interface: provide standardized uniform external service interface encapsulation for developers, analysts and other users, including synchronous and asynchronous interfaces.
2. Authority management: achieve unified control over platform function and data permissions.
3. Work-flow: realize unified configuration and scheduling control of various tasks through work-flow operation mechanism.
4. Integration of components: realize unified authentication, unified authority management and unified installation and deployment for all components of the platform.
5. Data control: including such functions points as data audit, data quality management, data provenance, data security management, etc.

1. On the basis of the existing data processing ability, in aspects of data acquisition, storage, calculation, analysis, mining and visualization interaction, etc., the method of external service uniform interface encapsulation can meet usage requirements of different users such as developers, analysts, etc.
2. Realize unified control and scheduling for data collection, storage, calculation, analysis and mining, etc., through the approach of work flow driven.
3. Unified authentication, unified access management, unified installation and deployment for components within platform such as data acquisition, data calculation, analysis and mining, data presentation, data service, platform control, data marts, etc.
4. Further strengthen the application of data control through the construction of complete data audit, data quality management, data provenance and data security management and other aspects.
5. Realize the unified access management for platform access and data permissions.

Figure 1: Power big data platform of CEPR

CEPRI regards one-stop big data analysis platform launched by Yonghong Tech as one of the core components of power big data platform to carry out fast modeling analysis and visual presentation on the basis of integrated internal and external data, and the visually presented data report is still embedded in the framework of power big data platform to implement uniform experience on platform for users.

The following are two typical application scenarios of power data platform of CEPRI:

1. Weak environment identification of distribution network. For users and section with relatively low reliability, deeply analyze internal (net rack, equipment, technology, management) and external factors (climate, geography, humanities, transportation) to cause weak power grid, explore direct or indirect relations between these factors or between these factors and reliability index, and result in quantitative results, in order to make an important support for related decisions on distribution network.

2. Analysis of line reliability model. Generally, a certain number of lines with similar structural characteristics, electrical characteristics and power supply reliability, etc, exist in the distribution network of a same area, which can be sorted into several types of line groups by using classification method of "network reliability characteristic patterns" , and lines in same group can be considered to have same model of reliability characteristics, so as to analyze from the view of reliability model of power grid to provide important reference for power distribution planning.