Stability and sustainability are the operational bottom lines of modern power grids. However, the inherent volatility and non-stationarity of wind energy, particularly in complex terrains, severely threaten power grid stability. To address this challenge, we propose an end-to-end architecture named the Adaptive Multi-scale Routing Wind Power forecasting (AMR-Wind) framework. The framework is principally composed of three sequential modules: an Adaptive Frequency Disentanglement Module (AFDM), an inverted Transformer (iTransformer), and a Scale-Routing Gated Recurrent Unit (SRGRU). The AFDM utilizes a differentiable filter bank to dynamically disentangle complex spectral signatures and mitigate mode mixing. The iTransformer is employed to effectively capture the complex multivariate dependencies between these disentangled modes and exogenous meteorological features. The SRGRU utilizes hierarchical temporal routing to synchronize localized high-frequency ramp events with macroscopic evolutionary trends. Comprehensive evaluations across four diverse wind farms demonstrate that AMR-Wind reduces the RMSE by an average of 8.4% and improves the $R^2$ by at least 1.0% compared to state-of-the-art baselines. Ablation studies further confirm the modules’ strong synergistic effects, yielding a 7.6% reduction in forecasting errors. This framework reduces the error in wind energy prediction, providing a reliable tool for the stability and sustainability of the power grid.

Industrial equipment-maintenance knowledge is scattered across technical manuals, historical work orders, and field experience. Rapidly retrieving these heterogeneous sources and generating accurate guidance are essential to reducing downtime losses and ensuring operational safety. General-purpose large language models perform well in open-domain question answering, but their pretraining corpora lack industrial maintenance terminology and diagnostic reasoning patterns, causing severe mistakes in direct applications. Semantic-similarity-based retrieval-augmented generation can introduce domain knowledge to reduce hallucinations, yet it cannot preserve causal direction and procedural-order constraints. This limitation can logically misalign retrieved evidence and misguide maintenance decisions. Field queries also vary in wording and information completeness and often contain equipment abbreviations, colloquial transcriptions, and omitted key details, making fixed retrieval and response strategies unreliable. Our framework combines progressive fine-tuning with dual-path evidence fusion. A perplexity-driven sample-difficulty partition organizes two-stage fine-tuning so that the model learns basic domain knowledge before complex diagnostic reasoning, mitigating domain hallucinations. Parallel knowledge-graph and vector retrieval uses graph topology to preserve recorded relation directions and source procedures to supply operating-sequence information. A three-level router driven by retrieval necessity and entity-matching status selects direct answering, vector retrieval, or joint graph–vector retrieval for queries with different levels of information completeness. Independent tests, noisy-query experiments, and blinded expert evaluation show improved answer accuracy, logical consistency, and noise robustness. Relative to the base model, two-stage fine-tuning improves BLEU and ROUGE by approximately 14% and 7%, respectively. The results indicate that progressive domain learning and structured external evidence can provide reliable decision support for industrial equipment maintenance.

Automobile sales analysis provides an important basis for production planning, procurement, supply-chain coordination, inventory management, and dealership resource allocation. However, dealership sales are jointly influenced by operational practices, market conditions, and regional policies. Conventional machine-learning models may have difficulty capturing spatial heterogeneity, maintaining predictive performance under regional sample imbalance, and providing interpretable evidence for managerial analysis. To address these issues, this study constructs a multi-source automotive dealership dataset based on automaker marketing records and proposes CDMEP, a multi-paradigm framework integrating spatial diagnosis, ensemble transfer learning, predictive interpretation, and assumption-conditional structural diagnosis. The framework comprises a Spatial Heterogeneity Decoupling Diagnostic Model (SHDM) for examining spatial clustering and regional differences; a Regional Adaptive Stacked Learner (RASL) combining balanced cross-regional transfer, TPE-tuned base learners, and constrained ensemble weighting; and a Causally Scrutable Optimization Policy (CSOP) incorporating SHAP-, GAM-, and PDP-based predictive interpretation, temporally ordered LiNGAM diagnosis, edge-level stability and sensitivity analysis. The results indicate the following: (1) dealership sales exhibit spatial clustering and regional heterogeneity in variables related to lead generation, customer conversion, and dealership service capability; (2) RASL achieves an R² of 0.875 on the independent test set, corresponding to an absolute improvement of 0.051 over the second-best benchmark in the present dataset; and (3) under the specified preprocessing procedures and LiNGAM assumptions, lead volume, customer traffic, and conversion rate repeatedly appear in candidate directional structures associated with next-month sales, although their estimated coefficients and diagnostic stability vary across regions. These findings provide preliminary evidence that CDMEP may serve as a useful analytical framework for regionally differentiated sales prediction and assumption-aware dealership analysis.