| 基于MPFA内嵌物理神经网络的园区空调负荷短期预测 |
| 点此下载全文 |
| 引用本文：温茂林1,邓汉钧2,杨帅2,余敏琪2,孟珺遐3,许刚1.基于MPFA内嵌物理神经网络的园区空调负荷短期预测[J].电网与清洁能源,2025,41(4):1~11 |
| 摘要点击次数: 9 |
| 全文下载次数: 1 |
|
| 基金项目:国家电网有限公司总部科技项目（5400-202323233A-1-1-ZN） |
|
| 中文摘要:提出一种基于多参数特征关联（multi parameter feature association，MPFA）内嵌物理神经网络（physics-informed neural network，PINN）的园区空调负荷短期预测模型。通过MPFA方法对多维数据进行聚类标记，提取特征向量作为模型输入。基于空调热力学物理方程和深度神经网络构建PINN模型，在模型中嵌入物理约束，实现对园区空调负荷的预测。以商业园区空调负荷数据为例，验证了所提方法的有效性。 |
| 中文关键词:园区建筑空调 物理神经网络 多参数特征关联 负荷预测 |
| |
| Short Term Load Forecasting of Air Conditioning in the Park Based on MPFA Embedded Physical Neural Network |
|
|
| Abstract:A short-term prediction model for industrial park air conditioning load， based on multi-parameter feature association （MPFA） embedded within a physical neural network （PINN）， is proposed. Initially， multidimensional data is clustered and labeled using the MPFA method， and the feature vector is extracted as the model input. Subsequently， the PINN model is constructed on the basis of the thermodynamic physical equation of air conditioning and a deep neural network， incorporating physical constraints into the model to achieve the prediction of industrial park air conditioning load. Finally， the air conditioning load data from a commercial park is utilized as an example to validate the efficacy of the proposed method. |
| keywords:building air conditioning in the park physical neural network multi-parameter feature association load forecasting |
| 查看全文 查看/发表评论 下载PDF阅读器 |
