Data-Driven Power System Analysis

For centuries, Electrical Power System has been the largest and most complicated man-made machine ever invented by humans, which is always sending electricity,  the essentials of human's production and life, to the almost every corner of the world all the time. The scientist and engineers have been studied the system to make sure it can operation in a safe, stable and economic manner from aspects of generation end, transmission system, distribution and user end, from system modelling, equipment and its manufacture, and control/protection with its own exclusive communication system, from aspects of computer simulation and high voltage testing of it. From all the aspects, we have achieved our goal so much that we all take the continuous and high/quality electricity supply for granted.

It is truly a glorious history of Electrical Power System until decades ago the internet has been developed. Then the top intelligent people on the earth turns to be focusing on big data, data mining, internet of things, machine learning, deep learning and finally artificial intelligence.

We all know that the usage of electricity can provide tons of data which can show human behaviour and habit. It has been long that we are thinking about how we could possibly take advantage of it. Therefore the insiders of the electrical power engineering and computer science at the beginning was looking at smart grid，which for a decade meant smart meters.

The real excitement comes from the most recent development of data-driven power system analysis. Thanks to the seminar given by Dr Ning Zhang, our Chinese collaborator at Tsinghua University, in Beijing. It subversively gives up modelling driven method for analyzing the electrical power system with hundreds of nodes, but take advantage of the AI technology to discover the whole system characteristics according to its historical or even real-time date of operation and then modelling and simulation of it without knowing the structure. Their research shows the machine learning increasing successful applications in power system analysis including power flow calculation, topology identification and operation mode analysis.

The online seminar was organized by the North America Chinese Power Professional Association. More than 100 Electrical Power Professionals participated and many constructive comments and discussion have been raised. There will be a series of online seminars held by the Association to share the cutting edge of electrical power engineering development relevant topics.

Lei Zhang, Ph.D. (lzhang@utec.edu.pe)

Energy Engineering Department