摘要
在配电网规划中,准确的电力负荷预测对电力系统的安全运行和供能质量至关重要。针对传统负荷预测方法对短时负荷预测结果较差的问题,提出基于BP神经网络的电力负荷预测方法。选定相似日数据进行矫正,基于神经网络模型快速得到负载功率。依托矫正输出的小数据集,减少了神经网络学习所需的数据,有效降低了神经网络的结构复杂度和时间,实现模型快速响应。当预测日期发生变化,神经网络能够重新训练并捕捉预报日的负荷与温度的关系,迅速响应输出相应的温度变化矫正。选取不规则变化温度数据与特殊日期作为预测对象验证模型有效性,预测最大绝对误差为14%,平均绝对误差1.63%。与传统方法相比,该方法能够在短时内进行预测,并大幅提升了预测效率,减少了数据需求,结果更加精准。
In distribution network planning,accurate power load prediction is crucial for the safe operation of power systems and the quality of energy supply.Addressing the issue of poor short-term load forecasting results in traditional load forecasting methods,a power load forecasting method based on the BP neural network was proposed.Selecting similar daily data for correction and quickly obtaining load power based on neural network models.By relying on a small corrected dataset,the data requirements for neural network training were reduced,the structural complexity and time of neural networks were effectively reduced,thereby achieving fast response of the model.When the forecasting date changes,the neural network can retrain and capture the relationship between load and temperature for the specific day,quickly adjusting the output based on temperature corrections.The irregular temperature variation data and special dates were selected as forecasting objects to verify the effectiveness of the model,the maximum absolute error of forecasting is 14%,and the average absolute error is 1.63%.Compared with traditional methods,this method can make forecasting in a short period of time,greatly improve forecasting efficiency,reduce data requirements,and produce more accurate results.
作者
杨欣
徐飞
贺国伟
周帆
丛昊
邓言振
Yang Xin;Xu Fei;He Guowei;Zhou Fan;Cong Hao;Deng Yanzhen(Economic and Technological Research Institute,State Grid Anhui Electric Power Co.,Ltd.,Hefei 230000,China;State Grid Anhui Electric Power Co.,Ltd.,Hefei 230000,China;Zhongneng Bowang(Beijing)Technology Co.,Ltd.,Beijing 102400,China)
出处
《能源与环保》
2025年第6期174-178,共5页
CHINA ENERGY AND ENVIRONMENTAL PROTECTION
基金
国网安徽科技项目(B1120922000C)。
关键词
神经网络
电力系统负荷预测
大数据分析
快速响应
neural networks
power system load forecasting
big data analysis
rapid response
作者简介
杨欣(1985-),男,安徽枞阳人,高级工程师,现从事配电网规划及其智能辅助决策技术工作。
