智能电网大数据流式处理方法与状态监测异常检测

发布时间：2018-06-17 20:52

本文选题：智能电网 + 大数据　；参考：《电力系统自动化》2016年14期

【摘要】：针对智能电网大数据流的实时性、易失性、无序性等特点,提出智能电网大数据的实时流处理框架,实现数据收集、数据缓冲与流式计算,满足状态监测异常检测与用电数据分析等快速处理需要。通过采集系统节点监听数据源变化并实时收集数据,利用消息订阅模式对数据进行缓冲,解决数据采集与流式计算速度不一致的问题。提出一种基于Storm的状态监测数据流滑动窗口处理方法,在规定时间内分批处理状态监测数据流,保证数据的连续计算,通过阈值判断进行异常检测。实验结果表明,在集群规模一定的条件下,适当地改变工作进程数以及执行器线程的并发数设置,可以增大滑动窗口的元件吞吐量,提高状态监测异常检测的实时处理效率。
[Abstract]:In view of the characteristics of real time, vulnerability and disorder of large data flow in smart grid, a real-time flow processing framework for smart grid large data is proposed, which realizes data collection, data buffer and flow calculation. It meets the needs of rapid processing of state monitoring anomaly detection and electrical data analysis. The data source changes and real-time is monitored through the collection system nodes. Collecting data, using message subscription model to buffer data to solve the problem that data collection is not consistent with flow computing speed. A sliding window processing method based on Storm for state monitoring data flow is proposed. In the specified time, the data flow is monitored in batch, continuous calculation of data is guaranteed, and the anomaly is judged by threshold. The test results show that, under the condition of a certain cluster scale, properly changing the number of work processes and the setting of the concurrent number of the actuator threads can increase the throughput of the element in the sliding window and improve the real-time processing efficiency of the state monitoring anomaly detection.
【作者单位】：华北电力大学控制与计算机工程学院;
【基金】：国家自然科学基金资助项目(61074078)~~
【分类号】：TM76

【相似文献】