基于核能供热技术的热网调控与智慧化管理
Heating Network Regulation and Intelligent Management Based on Nuclear Heating Technology
王锋 1孙英策 1刘慧文 2仇天星1
作者信息
- 1. 国家电投集团东北电力有限公司大连大发能源分公司,辽宁 大连 116000
- 2. 内蒙古工业大学电力学院,内蒙古 呼和浩特 010051
- 折叠
摘要
核电供热技术凭借其显著的低碳和高效特性,已成为实现"双碳"目标的重要手段之一.然而,当前核能供热负荷预测不准确、调控手段单一,无法满足该技术发展需求.为此,提出了一种基于卷积循环注意力机制模型(CGAM)的精细化调控和热负荷预测方法用于核电供热系统.该方法首先采用多层次数据存储技术和数据同步与优化策略,确保生产数据的实时性与准确性;其次引入深度网络对核电换热站短期热负荷进行精确预测并实施策略调整,确保预测精度;最后依据预测结果设置多层次精准调控手段,完成供热系统的智能调节.将该方法应用于大连红沿河核电供热项目中,实验结果显示,对实际工况的热负荷实现了准确预测,对供热系统起到了精准调控的作用,证明了算法的有效性.
Abstract
With its remark able low-carbon and high-efficiency characteristics,nuclear power heating technology has become one of the important means to achieve the goal of"dual carbon".However,the current nuclear heating load prediction is inaccurate and the control means are single,which cannot meet the needs of this technology development.Therefore,a refined regulation and heat load prediction method based on convolutional cycle attention mechanism model(CGAM)was proposed for nuclear power heating system.Firstly,multi-level data storage technology and data synchronization and optimization strategy were adopted to ensure the real-time and accuracy of production data.Secondly,a deep network was introduced to accurately predict the short-term heat load of the nuclear power heat exchange station and implement strategy adjustment to ensure the prediction accuracy.Finally,multi-level precise control means were set according to the prediction results to complete the intelligent adjustment of the heating system.The method was applied to Dalian Hongyanhe Nuclear Power Heating Project.The experimental results show that the heat load of the actual working conditions is accurately predicted,and the role of precise regulation of the heating system is achieved,which shows the effectiveness of the algorithm.
关键词
核电供热/精细化调控/热负荷预测/智慧化管理Key words
nuclear power heating/fine regulation/heat load forecasting/intelligent management引用本文复制引用
基金项目
国电投科技项目(DNYZC2022080169813)
国家自然科学基金(62241309)
内蒙古自治区科技计划(2020GG0283)
出版年
2024
NETL
NSTL
维普
万方数据