三、摘要中英对照（精读核心）
🎧 音频在文末，可先听一遍原文再读；每个英文句都配了逐句翻译。
摘要 Abstract
EN · 原文
Reliable photovoltaic (PV) forecasts are needed for low-carbon energy systems, but newly deployed sites often have short, imperfect records.
CN · 翻译
低碳能源系统需要可靠的光伏功率预测，但新建电站往往只有
短期且不完整的数据记录。
This makes standard day-ahead forecasting difficult: persistence and physical baselines can be sensitive to calibration and timestamp alignment, while single machine-learning models may capture only one structure in the data and overstate skill under non-temporal validation.
这使标准的日前预测变得困难：持久性基线和物理基线对标定与时间戳对齐非常敏感，而单一的机器学习模型往往只能捕捉数据中的某一种结构，并在非时序（随机划分）验证下高估自身水平。
We study this problem at a United Kingdom charging-station site, where PV forecast errors affect charging availability, storage scheduling, and downstream control.
我们在英国的一个充电站站点研究这一问题——在那里，光伏预测误差直接影响充电可用性、储能调度和下游控制。
Using measured inverter output and publicly available meteorological inputs, we develop a deployment-oriented environmental-AI pipeline for day-ahead hourly PV forecasting.
利用实测的逆变器输出和公开可获取的气象数据，我们开发了一条面向实际部署的环境人工智能流水线，用于日前逐小时光伏功率预测。
The pipeline corrects timestamp conventions, constructs leakage-safe solar-geometry and clearness-index features, adds short-term atmospheric context, and combines complementary predictors through validation-learned stacking.
该流水线：① 校正时间戳约定；② 构造无泄漏的太阳几何与晴朗指数特征；③ 加入短期大气背景信息；④ 通过「在验证集上学习」的堆叠方法融合互补的预测模型。
Against smart persistence, a clear-sky baseline that adjusts recent PV output using expected clear-sky irradiance, the best ensemble reduces daylight normalised RMSE by about 32% under random day-blocked evaluation and 9% under the stricter rolling-origin protocol.
与「智能持久性」基线（利用预期晴空辐照度修正近期光伏输出的晴空基准）相比，最优集成模型在随机按天分块评估下将白天归一化 RMSE 降低了约 32%，在更严格的滚动起点协议下降低了 9%。
It also reduces daylight RMSE relative to the strongest individual machine-learning baseline by 6.6% and 6.4%, respectively.
与最强的单一机器学习基线相比，白天 RMSE 也分别降低了 6.6% 和 6.4%。
The results show that physics-aware stacking can support PV forecasts from limited site data, but its value depends on model class, evaluation protocol, and deployment context.
结果表明，融入物理信息的堆叠方法能在数据有限的电站条件下支撑光伏预测，但它的价值取决于模型类别、评估协议和部署场景。
关键词 Keywords：Environmental Artificial Intelligence 环境人工智能 ｜ Photovoltaic Forecasting 光伏功率预测 ｜ Decision Support 决策支持
四、引言精选（为什么这个问题重要）
① 为什么机器学习做环境预测 ≠ 普通预测
Machine learning is now central to forecasting environmental quantities that affect energy, climate, and infrastructure decisions, including solar generation, wind power, electricity demand, river flow, and air quality. In these settings, the practical value of a forecast depends on whether it can support timely and reliable decisions, not only on whether it improves an isolated error metric.
如今，机器学习已成为预测影响能源、气候和基础设施决策的环境量的核心手段，包括太阳能发电、风力发电、电力需求、河流流量和空气质量等。在这些场景中，预测的实际价值取决于它能否支撑及时可靠的决策，而不仅仅是改善某个孤立的误差指标。
These tasks are not standard tabular prediction problems. The target is usually a structured time series; the drivers are multivariate, weather-dependent, and partly predictable; and forecast skill often depends as much on how data are aligned and represented as on the model itself.
这类任务并不是标准的表格数据预测问题：预测目标通常是结构化的时间序列；影响因素是多变量的、依赖天气的、且只能部分预测；预测水平往往既取决于模型本身，也取决于数据如何对齐和表示。
② 为什么光伏预测是环境 AI 的理想「案例研究」
Day-ahead forecasting, predicting the next 24 hours at hourly resolution, is a particularly important decision-support setting. It is close enough for weather information to remain informative, but far enough ahead for cloud evolution and atmospheric uncertainty to make the task difficult.
日前预测——以小时分辨率预测未来 24 小时——是一个特别重要的决策支持场景。它的时间尺度「近到天气信息仍然有用，远到云层演变和大气不确定性足以让任务变得困难」。
Photovoltaic (PV) power forecasting is a useful case study for environmental artificial intelligence (AI) because generation is bounded, non-linear, strongly seasonal, and driven by physical variables that are often available only through numerical weather products or reanalysis rather than direct site measurements.
光伏功率预测之所以是环境 AI 的理想案例，是因为光伏发电是有界的、非线性的、强季节性的，且由物理变量驱动——而这些变量往往只能通过数值天气预报产品或再分析数据获得，而非现场直接测量。
③ 这篇论文的出发点（三个研究问题）
This paper starts from a simple observation: in day-ahead environmental forecasting, the main limitation is not always the absence of a deeper model. Forecasts can fail because measurements and weather fields are misaligned, because calendar encodings do not capture solar physics, or because a single learner cannot exploit multiple complementary structures in the data.
本文从一个简单的观察出发：在日前环境预测中，主要瓶颈并不总是「模型不够深」。预测失败可能是因为测量数据与天气数据错位、因为日历编码没能捕捉太阳物理规律、或者因为单个模型无法利用数据中多种互补的结构。
💡 这是全文最有味道的一句话：很多论文都在「堆更复杂的模型」，而这篇论文冷静地指出——工程细节（数据对齐、特征构造）往往比模型更关键。这种思考方式就是科研品味。
五、论文贡献（5 个要点）
1. A timestamp-aligned forecasting pipeline. The authors identify and correct a one-hour mismatch between site measurements and gridded weather inputs. This improves physical estimates, raising the daylight coefficient of determination (R²) of a clear-sky proxy from 0.42 to 0.60, while showing that the learned ensemble is largely insensitive to the correction.
1. 时间戳对齐的预测流水线。他们识别并修正了站点测量与网格天气数据之间「差一小时」的错位，使晴空代理模型的白天决定系数 R² 从 0.42 提升到 0.60；同时发现学习型集成对该修正大体不敏感。
2. A leakage-safe, physics-informed problem representation. Irradiance-, geometry-, and weather-aware covariates are derived from public reanalysis data without using held-out targets. Physics is most effective through the prediction target, using the clearness index, rather than by simply adding raw solar-geometry features.
2. 无泄漏、融入物理信息的问题表示。从公开再分析数据构造辐照度、几何、天气相关的协变量，且不使用留出期的目标数据（避免泄漏）。物理信息通过「预测目标用晴朗指数」注入最有效，而不是简单堆原始太阳几何特征。
3. A diverse set of day-ahead forecasting models. Tabular, clear-sky-normalised, hour-specialised, and spatiotemporal convolutional models are trained under a common setup, enabling comparison of both accuracy and the structures captured by each approach.
3. 一组多样化的预测模型。在统一框架下训练了表格模型、晴空归一化模型、逐小时专用模型和时空卷积模型，既比较精度，也比较每种方法捕获的「结构」。
4. An interpretable validation-learned ensemble. Base forecasts are combined using non-negative weights learned only on validation data. The convex ensemble improves over the best individual model and assigns weight to complementary predictors.
4. 可解释的验证集学习集成。用仅在验证数据上学得的非负权重组合各基础预测。这个凸集成优于最佳单一模型，并把权重分配给互补的预测器。
5. A deployment-oriented evaluation and ablation study. Forecasts are assessed with day-blocked cross-validation and rolling-origin backtesting. Ablations show that each pipeline component has model- and protocol-dependent value.
5. 面向部署的评估与消融研究。用按天分块的交叉验证和滚动起点回测评估预测。消融实验表明：每个流水线组件的价值都取决于模型与评估协议。
六、结论中英对照
This paper presented a physics-aware AI pipeline for day-ahead PV forecasting as a decision-support tool for low-carbon energy systems under a realistic limited-data setting. ... Experimental results show that physics-aware pipeline steps are not uniformly beneficial; their value depends on the model class and the evaluation protocol, and this dependence can be measured.
本文在真实的有限数据条件下，提出了一条融入物理信息的 AI 流水线，作为低碳能源系统的日前光伏预测决策支持工具。……实验结果表明：融入物理信息的流水线步骤并非一律有益；其价值取决于模型类别与评估协议，而这种依赖性是可以被测量出来的。
A genuine one-hour data misalignment was decisive for parameter-free physical models and physically interpretable quantities, yet was absorbed by the flexible learners. Clear-sky structure was most useful as a normalising target, while supplying the same structure as raw seasonal covariates degraded rolling-origin skill through poor extrapolation.
真实存在的一小时数据错位对无参数物理模型和物理可解释量是决定性的，却被灵活的机器学习模型「吸收消化」了。晴空结构作为归一化目标时最有用；而把同样的结构作为原始季节协变量输入，则会因外推能力差而降低滚动评估的表现。
Short-term weather context was the single ingredient that most improved deployment-relevant performance. Robust environmental AI should therefore be pipeline-centred and protocol-aware: it should diagnose alignment for the components that need it, inject physics through stationary normalising representations, preserve short-horizon weather context, and report the stricter temporal protocol as the primary estimate of operational skill.
短期天气背景是提升部署性能贡献最大的单一要素。因此，稳健的环境 AI 应当「以流水线为中心、对评估协议有感知」：为需要的组件诊断对齐问题；通过平稳的归一化表示注入物理信息；保留短时天气背景；并把更严格的时序评估协议作为运行水平的主要估计。
The achieved performance is notable because it is obtained from a short and imperfect operational record, suggesting that useful deployment-oriented PV forecasts can be developed before extensive multi-year archives exist.
这样的成绩之所以值得注意，是因为它是在短期且不完整的运行记录上取得的——这提示我们：在积累多年数据档案之前，就能开发出有用的、面向部署的光伏预测。