DHNs simulation data

This simulation dataset provides the topology structures of four different District Heating Networks (DHNs), along with randomly selected consumer clusters. It features high-resolution, with 1-minute time step, simulation results for 5.5 months, including water temperatures, mass flow rates, consumer substation heating demands, and heat source power productions. This data serves as the foundation for training, validating and testing diverse neural networks models to learn clusters underlying physics. Publicly available training and data pipeline code on Github [https://github.com/drod-96/smart_clusters_v1] facilitates data extraction for our machine learning models and post-treatment.

本仿真数据集提供了四种不同的区域供热网络（District Heating Networks, DHNs）的拓扑结构，以及随机选取的用户集群。该数据集具备高分辨率特性，采用1分钟时间步长，包含时长5.5个月的仿真结果，具体涵盖水温、质量流量、用户换热分站供热需求以及热源功率产出数据。本数据集可作为训练、验证与测试多种神经网络模型的基础，用于学习集群背后的内在物理规律。项目在GitHub平台（链接：https://github.com/drod-96/smart_clusters_v1）上公开了训练及数据流水线代码，可辅助机器学习模型的数据提取与后处理工作。

搜集汇总

背景与挑战

背景概述

该数据集包含四个区域供热网络的拓扑结构和消费者集群的高分辨率模拟数据(1分钟时间步长，5.5个月)，涵盖水温、流量、热需求等多维参数，专为神经网络模型训练设计，并配有公开的GitHub数据处理代码。