Development of Fast Real-time Online Dynamic Security Assessment System

 By Mike Zhou, XueWei Shang, Lin Zhao, JianFeng Yan, DongYu Shi, and Ying Chen

After a decade of development, online dynamic security assessment (DSA) analysis has been widely used in the power dispatching control centers in China. The current online DSA system is a near real-time analysis system with speed in the order of minutes. The development of a fast real-time DSA system with speed in the order of seconds has been started with the focus on helping the operator to perform DSA analysis in real-time. An end-to-end optimization approach is used to minimize the DSA system overall round trip time, including data acquisition, data processing and computation. The transient stability analysis is performed by searching of the existing simulation study cases in the knowledge base, rather than the step-by-step time-domain simulation method. Big Data technologies, including in-memory computing, complex event processing, search and pattern-matching, and deep machine learning, will be introduced to the power grid dispatching control centers in China in the near future.

After a decade of development, online dynamic security assessment (DSA) analysis has been widely used in the power dispatching control centers in China. In an online DSA system, power grid operation measurement data is processed first by supervisory control and data acquisition (SCADA) and then by State Estimation to estimate the state of the network and establish a power flow study case from the redundant telemetry measurements before starting the DSA analysis. DSA analysis currently used in the dispatching control centers includes feature:

a) monitoring and early warning;
b) stability margin analysis; and
c) control/adjustment decision support.

The state grid of China currently is modeled using a 40000-bus power network model. A round DSA trip, including SCADA, State Estimation and DSA analysis, currently takes about 10 minutes. Therefore, the current online DSA system, when used to handle the large-scale power network model, is a near real-time analysis system with speed in the order of minutes. Online DSA analysis is currently performed in the dispatching control centers periodically at an interval of 15 minutes.

With the development of the ultra-high voltage AC/DC national grid in China, we have foreseen the need for a fast real-time online DSA system with speed in the order of seconds. Power grid dynamic process, due to disturbance, develops at a speed of seconds. The fast real-time online DSA system will help the operator to perform real-time analysis and support their decision making process, while the system dynamic process is developing, rather than after-the-fact analysis. A new fast real-time online DSA development project, sponsored by the State Grid of China, has been started. The primary goal is to reduce the online DSA system overall round trip time, from data acquisition to complete the analysis, from the current proximate 10 minutes to less than 60 seconds.

A distributed in-memory data grid based dataspace technology is used to speed up the data acquisition process, including SCADA and State Estimation. A real-time power network model, which runs inside the dataspace and subscribes to the change event published by the SCADA/WAMS (Wide Area Measurement System) and updates itself, is used to track the power grid operation state in real-time. A complex event processing engine listens to power grid state change events, published by the real-time power network model, to perform situation awareness analysis through event-driven rule evaluation and pattern-matching to identify potential grid operation dynamic security risk. If significant operation condition change with regarding to the security is detected, a fast DSA analysis is performed to identify the potential risk, and if necessary warn the operator and provide assistance to the operator in decision making to mitigate the risk.

At an interval of 15 minutes, the current DSA system running in the dispatching control centers performs the online DSA analysis, including step-by-step time-domain transient stability simulation, of the current operating point (model-driven approach). The new DSA implementation is based on search of existing power grid simulation cases (data-driven approach), produced in the off-line studies and the on-line simulation, and stored in a Hadoop based power grid simulation study case knowledge base. Currently, the State Grid dispatching control center of China produces per year 20,000 off-line and 25 million on-line study cases in their power grid simulation. These simulation study cases are collected and stored in the knowledge base. Big Data technology and Deep Machine Learning method are used to process the huge volume of the simulation data and make them ready for the search and pattern-matching based new DSA system. As a comparison, Google's AlphaGo has about 30M Go moves (equivalent to the power grid simulation study cases) collected from expert human players in its knowledge base to beat the humankind in Go games.

In summary, the development of a fast real-time DSA system with speed in the order of seconds has been started with the focus on helping the operator to perform DSA analysis in real-time. An end-to-end optimization approach is used to minimize the DSA system overall round trip time, including data acquisition, data processing and computation. The transient stability analysis is performed by searching of the existing simulation study cases in the knowledge base, rather than the step-by-step time-domain simulation method. Big Data technologies, including in-memory computing, complex event processing, search and pattern-matching, and deep machine learning, will be introduced to the power grid dispatching control centers in China in the near future.

For a downloadable copy of  June 2016 eNewsletter which includes this article, please visit the IEEE Smart Grid Resource Center

Contributors