基于机器学习的电网调度指令智能交互技术研究  

Research on intelligent interactive technology of power grid dispatching instructions based on machine learning

作　　者：郭琦[1] 袁少卿 吕小凡 蔡硕琦 GUO Qi;YUAN Shaoqing;LV Xiaofan;CAI Shuoqi(Inner Mongolia Electric Power Group Co.,Ltd.,Hohhot 010090,China)

机构地区：[1]内蒙古电力(集团)有限责任公司,内蒙古呼和浩特010090

出　　处：《电子设计工程》2022年第9期153-157,共5页Electronic Design Engineering

基　　金：内蒙古电力公司技改项目(DUKZZZ-YBHT-2020-JSC0401-0055)。

摘　　要：随着新一代人工智能和通信技术的快速发展,电网调度指令技术面临着新的技术革新,传统的调度指令操作已无法满足目前提质增效的要求。文中深入研究了基于机器学习的电网调度指令智能交互技术,将电网调度指令分解为执行和评估两个步骤,并建立系统化的人机交互模型。模型设计涵盖了人机交互的4个阶段和一个循环过程,通过交互式迭代提升了调度系统的可交互性。采用机器学习算法完成调度指令的数据清洗,利用端对端的传输结构实现电网调度指令的传输,同时建立传输通道模型。经过系统测试验证了所提方法的有效性,电网调度指令的平均识别率达到91.45%,能够满足电网调度的基本要求。With the rapid development of the new generation of artificial intelligence and communication technology,the power grid dispatching instruction technology is facing new technological innovation.The traditional dispatching instruction operation has been unable to meet the current requirements of improving quality and efficiency.In this paper,the intelligent interaction technology of power grid dispatching instruction based on machine learning is deeply studied.The power grid dispatching instruction is divided into two steps: execution and evaluation,and a systematic human-computer interaction model is established.The design of the model covers four stages of human-computer interaction and a cycle process.The interactivity of the dispatching system is improved through interactive iteration.The machine learning algorithm is used to complete the data cleaning of dispatching instructions.The endto-end transmission structure is used to realize the transmission of dispatching instructions. At the same time,the transmission channel model is established.The effectiveness of the proposed method is verified by system test,and the average recognition rate of power grid dispatching instructions reaches 91.45%,which can meet the basic requirements of power grid dispatching.

关 键 词：机器学习 调度指令 人机交互 语音合成 特征提取 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]