DiTEC-WDN
收藏资源简介:
DiTEC-WDN是由荷兰格罗宁根大学伯努利学院和阿姆斯特丹大学信息学院共同创建的大型水分布网络液压场景数据集,包含36,000个独特的模拟场景。该数据集通过自动化管道优化关键参数,进行大规模模拟,并记录了在标准条件下通过规则验证和事后分析得到的离散、合成的但水力现实的网络状态。DiTEC-WDN总共生成了2.28亿个基于图的 states,能够支持各种机器学习任务,包括图级、节点级和链接级回归以及时间序列预测。
DiTEC-WDN is a large-scale hydraulic scenario dataset for water distribution networks, co-created by the Bernoulli Institute of the University of Groningen in the Netherlands and the School of Informatics of the University of Amsterdam. It contains 36,000 unique simulation scenarios. This dataset conducts large-scale simulations by automatically optimizing key pipeline parameters, and records discrete, synthetic yet hydraulically realistic network states derived from rule-based validation and post-hoc analysis under standard conditions. DiTEC-WDN has generated a total of 228 million graph-based states, which can support a wide range of machine learning tasks including graph-level, node-level, link-level regression and time series forecasting.




