摘要
电源状态监测数据易存在冗余与过拟合的问题,影响最终监测结果的准确性。因此,设计了基于改进卷积神经网络(Convolutional Neural Network,CNN)的变电站数据网屏电源状态实时监测方法。采集变电站数据网屏电源状态数据,将电源状态监测数据转化成统一的个数,并在监测终端处理数据,使电源状态数据更加真实。基于改进CNN监测变电站数据网屏电源充放电阈值,通过数据采集获取电源标定电容、充放电电流、电压、温度等数据,并建立电源状态评估矩阵,确保电源状态监测的准确性。采用对比实验,验证了该方法的监测结果更准确,能够应用于实际生活。
The power supply status monitoring data is prone to redundancy and overfitting,which affects the accuracy of the final monitoring results.Therefore,the real-time monitoring method of power status of substation data network screen based on improved Convolutional Neural Network(CNN)is designed.Collect the power status data of the substation data network screen,transform the power status monitoring data into a unified number,and process the data at the monitoring terminal to make the power status data more real.Based on the improvement of the power charge and discharge threshold of CNN monitoring substation,the power calibration capacitance,charge and discharge current,voltage,temperature and other data are obtained through data collection,and the power status evaluation matrix is established to ensure the accuracy of power status monitoring.The comparative experiment proves that the monitoring results of this method are more accurate and can be applied in real life.
作者
周洁
陈坊
ZHOU Jie;CHEN Fang(State Grid Zhejiang Electric Power Limited Company Zhuji Power Supply Company,Shaoxing Zhejiang 311800,China)
出处
《信息与电脑》
2023年第22期145-147,共3页
Information & Computer
基金
国网浙江省电力有限公司科技项目“国网浙江绍兴基于智能控制的变电站调度数据网第三路电源技术的研究与应用测试加工服务项目”(项目编号:5211S123000D)。
关键词
卷积神经网络
变电站
数据网屏
电源状态
实时监测方法
Convolutional Neural Network
substation
data network screen
power supply status
real-time monitoring method
作者简介
周洁(1989-),女,山西太原人,硕士研究生,工程师。研究方向:调度自动化。
