| 引用本文: | 陈碧云,李弘斌,李滨.基于数据挖掘和CAPSO-SNN的电力作业风险态势感知[J].电力自动化设备,2020,40(1): |
| CHEN Biyun,LI Hongbin,LI Bin.Power operation risk situation awareness based on data mining and CAPSO-SNN[J].Electric Power Automation Equipment,2020,40(1): |
|
| 摘要: |
| 随着电力作业安全监控技术的不断发展，电力作业全过程在线信息采集成为可能。以电力作业数据为基础，提出一种基于数据挖掘和云自适应粒子群优化脉冲神经网络(CAPSO-SNN)的电力作业风险态势感知方法。该方法首先从电力作业事故事件数据库中提炼出所有作业风险影响因素以构建风险影响因素体系，然后通过主成分分析法从中挖掘出作业过程中应重点关注的风险关键要素，再以风险关键要素作为输入参数，通过云自适应粒子群优化脉冲神经网络进行作业风险态势感知的训练和预测。最后，以某电网公司的实际历史作业事故事件为样本，展示了所提方法的应用过程。算例结果表明，该方法不仅适用于分析电力作业的风险组成，还可以在作业过程中有效地感知风险状态信息，跟踪风险发展趋势，有助于实施电力作业风险的全过程精细化态势利导管控。 |
| 关键词: 数据挖掘 态势感知 云自适应粒子群优化 脉冲神经网络 态势利导 电力作业 |
| DOI：10.16081/j.epae.201912027 |
| 分类号:TM734 |
| 基金项目:国家自然科学基金资助项目(51767002) |
|
| Power operation risk situation awareness based on data mining and CAPSO-SNN |
|
CHEN Biyun, LI Hongbin, LI Bin
|
|
Guangxi Key Laboratory of Power System Optimization and Energy Technology, Guangxi University, Nanning 530004, China
|
| Abstract: |
| With the continuous development of power operation security monitoring technology, online information collection in the whole process of power operation becomes possible. On the basis of power operation data, a power operation SA(Situation Awareness) method based on data mining and CAPSO-SNN(Cloud Adaptive Particle Swarm Optimization-Spiking Neural Networks) is proposed. Firstly, this method extracts all the operation risk influencing factors from the power operation accident events database to cons-truct the risk influencing factors system, and then excavates the key risk factors which should be focused on in the process of operation through PCA(Principal Component Analysis) algorithm. Subsequently, the key risk factors are taken as input parameters to realize the training and prediction of operation risk SA via CAPSO-SNN. Finally, the application process of the proposed method is demonstrated based on a sample of actual historical operation accidents of a power grid company. The simulative results show that the proposed method is not only suitable for analyzing the risk composition of power operation, but also can effectively perceive the risk status information and track the risk development trend during the process of operation, which contributes to implementing the whole process delicacy situation orientation management and control of power operation risk. |
| Key words: data mining situation awareness cloud adaptive particle swarm optimization spiking neural networks situation orientation power operation |
