Real-time estimates of Swiss electricity savings using streamed smart meter data

• 全文链接
• 请求原文

作   者：

Mari A.;  Remlinger C.;  Castello R.;  Obozinski G.;  Quarteroni S.;  Heymann F.;  Galus M.; 

作者机构：

Swiss Federal Office of Energy;  Swiss Data Science Center EPFL & ETH Zürich; 

关键词：

Generalised additive models;  Energy policy;  Smart meters;  Energy saving; 

期刊名称：

Applied energy

i s s n：

0306-2619

年卷期：

2025 年 377 卷 Jan.1 Pt.C 期

页   码：

1.1-1.7

页   码：

摘   要：

© 2024 The AuthorsThe gas crisis of 2022 put pressure on electricity prices in Europe, prompting the Swiss government to launch a national energy-saving campaign. To effectively quantify potential savings and guide timely decision-making, this campaign called for rigorous near-real-time modeling of changes in electricity consumption habits. The proposed approach estimates national electricity consumption at an hourly resolution across three consumer categories using thousands of streamed smart-meter load curves. These curves are aggregated to produce a national consumption estimate using scaling factors that account for differences among Swiss distributors. These factors are derived by regressing historical annual consumption against public socio-economic variables. The obtained national load curve is adjusted for the influence of weather conditions, the calendar and global trends, in order to compare different periods with a reference scenario. Such external effects are modeled with splines using Generalized Additive Models, trained on a 5-year dataset, to precisely measure each contribution on the national consumption and evaluate the consumers response to the saving plan. The results indicate a reduction of approximately 4.8% of the adjusted electricity consumption during winter 20222023, equivalent to an average monthly savings of 246 GWh, distributed across residential, service, and industrial sectors.