Behind-the-Meter Photovoltaic Visibility from Net Smart Meters: Deployment Regimes and Operational Insights

Computer Science > Machine Learning arXiv:2610.12527 (cs) [Submitted on 8 Oct 2026] Title: Behind-the-Meter Photovoltaic Visibility from Net Smart Meters: Deployment Regimes and Operational Insights Authors: Jack Zheng, Hao Wang.

计算机科学 > 机器学习 arXiv:2610.12527 (cs) [提交于 2026 年 10 月 8 日] 标题：通过净智能电表实现表后光伏可见性：部署机制与运营洞察 作者：Jack Zheng, Hao Wang。

Abstract: Distribution operators need visibility of behind-the-meter photovoltaic (PV) generation, yet most residential customers are observed only through net smart meters. Deep-learning-based PV disaggregation can restore this visibility, but its operational value depends on whether the model serves known customers, new customers, future periods, or new regions.

摘要：配电运营商需要了解表后光伏（PV）发电情况，但大多数住宅用户仅通过净智能电表进行监测。基于深度学习的光伏分解技术可以恢复这种可见性，但其运营价值取决于模型是服务于已知用户、新用户、未来时段还是新区域。

We quantify this dependence using four deployment regimes that separate household transfer from temporal extrapolation, applied to Ausgrid smart-meter data with cross-dataset adaptation to the Pecan Street data. Using a fixed training protocol, we benchmark representative deep sequence architectures for PV disaggregation under regime-matched evaluation.

我们利用四种部署机制量化了这种依赖性，这些机制将家庭迁移与时间外推区分开来，并应用于 Ausgrid 智能电表数据，同时实现了向 Pecan Street 数据的跨数据集适配。通过固定的训练协议，我们在机制匹配的评估下对用于光伏分解的代表性深度序列架构进行了基准测试。

At a fixed 48-hour context length, daylight mean absolute error (MAE) of the best model rises from 0.104 kW on seen households in the future test period to 0.185 kW on unseen households over the same period, and under a fixed training contract the deployment regime explains far more of the error variation than model choice.

在固定的 48 小时上下文长度下，最佳模型在未来测试期间针对已知家庭的日间平均绝对误差 (MAE) 从 0.104 kW 上升到针对同一时期未知家庭的 0.185 kW；在固定的训练契约下，部署机制对误差变化的解释程度远高于模型选择。

Under target-domain normalization, fine-tuning on five target-region households reduces source-model transfer error by 28%—36%. Long input windows help known-customer extrapolation, while their new-customer benefits vary by model. These findings suggest regime-matched error budgets, limited target supervision for model adaptation in new regions, and use-case-dependent context selection for energy data analytics.

在目标域归一化下，对五个目标区域家庭进行微调可将源模型迁移误差降低 28%—36%。较长的输入窗口有助于已知用户的外推，但其对新用户的益处因模型而异。这些发现表明，能源数据分析应采用机制匹配的误差预算、针对新区域模型适配的有限目标监督，以及依赖于用例的上下文选择。