This paper presents an electromagnetic transient analysis of lightning-initiated overvoltage stresses of the step-up transformers installed inside a nacelle of onshore, multi-megawatt, new-generation wind turbines. The increase in the wind turbine (WT) nominal power output, necessitated introducing the step-up transformer into the nacelle. A transformer installed inside a nacelle is subjected to completely different overvoltage stresses from those present if it were installed at the base of the WT tower. This has serious repercussions on its overvoltage protection (i.e., selection and installation of surge arresters) and insulation coordination. Furthermore, the overvoltage protection of medium-voltage cables (inside the tower) is also problematic when considering their length, proximity to the tower wall, and their screen grounding practices, and needs to be tackled in conjunction with that of the step-up transformer. This paper presents detailed models for the various components of the latest-generation WTs, intended for fast-front transient analysis and assembled within the EMTP software package. We further present the comprehensive results of the lightning-transient numerical simulations, covering both upward and downward (first and subsequent) strikes, their analysis, and recommendations for the optimal selection of medium-voltage surge arresters for the step-up transformers installed inside a nacelle.

No abstract

This paper presents a comprehensive error analysis of the day-ahead photovoltaic (PV) production multi-step forecasting model that uses a chained support vector regression (SVR). A principal component analysis (PCA) is also implemented to investigate possible improvements of the SVR prediction accuracy. Special attention was given to the hyper-parameter tuning of the chained SVR and PCA+SVR models; specifically, the dispersion of the prediction errors when fine-tuning the model with an experimental halving random search algorithm implemented within scikit-learn, i.e. the HalvingRandomSearchCV (HRSCV). The obtained results were compared with the traditional randomized search technique, i.e. the RandomizedSearchCV (RSCV). The chained SVR model prediction errors were analysed for several different parameter distribution settings. After doing repetitive fine-tuning and predictions, it was observed that the HRSCV tends to choose sub-optimal hyper-parameters for certain scenarios, as will be elaborated in the paper. Moreover, when analysing prediction errors of the same model fine-tuned repetitively with the HRSCV and RSCV, it was found that HRSCV creates larger errors and more inconsistency (variability) in the prediction results. The introduction of the PCA to the chained SVR model, at the same time, reduces the influence of exogenous variables and, on average, increases its performance and decreases prediction errors regardless of the optimization technique used.

Curtailment losses for large-scale hybrid wind–solar photovoltaic (PV) plants with a single grid connection point are often calculated in 1 h time resolution, underestimating the actual curtailment losses due to the flattening of power peaks occurring in shorter time frames. This paper analyses the curtailment losses in hybrid wind–PV plants by utilising different time resolutions of wind and PV production while varying the grid cut-off power, wind/solar PV farm sizes, and shares of wind/PV capacity. Highly resolved 1 s measurements from the operational wind farm and pyranometer are used as an input to specialized wind and PV farm power production models that consider the smoothing effect. The results show that 15 min resolution is preferred over 1 h resolution for large-scale hybrid wind–PV plants if more accurate assessment of curtailment losses is required. Although 1 min resolution additionally increases the estimation accuracy over 15 min resolution, the improvement is not significant for wind and PV plants with capacity above approx. 10 MW/10 MWp. The resolutions shorter than 1 min do not additionally increase the estimation accuracy for large-scale wind and PV plants. More attention is required when estimating curtailment losses in wind/PV plants with capacity below approx. 10 MW/10 MWp, where higher underestimation can be expected if lower time resolutions are used.