| Autor: |
Cheng, Min, Zhang, Dan, Yan, Wenlin, He, Lei, Zhang, Rongkui, Xu, Mingyu |
| Předmět: |
|
| Zdroj: |
International Journal of Emerging Electric Power Systems; Feb2023, Vol. 24 Issue 1, p91-102, 12p |
| Abstrakt: |
The energy crisis is a problem that countries all over the world pay more and more attention to, and a series of ecological problems caused by it have become increasingly prominent. It is difficult for traditional fossil fuels to maintain a healthy and coordinated sustainable development of society and economy. The establishment of a sustainable energy system has become the development trend of various countries to solve energy problems. Electric energy is a secondary energy that all primary energy can be converted into, and an irreplaceable consumable for all industrial technologies and people's lives. Electric power data has the characteristics of large rate span, numerous data sources, complicated interaction methods, and various types of data. The existence of abnormal data in the power system will greatly reduce the accuracy of the system state estimation and the state estimation convergence rate. This paper introduces the power grid industrial control system, combines the data flow of power big data, and analyzes the abnormal information detection process in detail. It takes the data stream acquired by the acquisition unit PMU of the wide area measurement system as the research object. The rapid development of the Hadoop big data platform provides important technical support for the research of power grid big data. Based on the Hadoop platform, the clustering algorithm is used to complete the anomaly detection of real-time data. The LOF algorithm has poor performance when dealing with a large amount of high-dimensional data, and has high time and space complexity. In order to make up for the shortcomings of the LOF algorithm, this paper uses the K-means clustering algorithm to propose an improved algorithm K-LOF of the density-based local abnormal factor detection algorithm LOF, and optimizes the neighborhood query process. It is verified by experiments that the K-LOF algorithm can effectively reduce the time complexity of the anomaly detection algorithm and improve the detection accuracy by 2–4.2%. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
