Browsing by Author "Astudillo Calle, Jorge Eduardo"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Actualización y coordinación de las funciones de protección de las Centrales Saucay, Saymirín, El Descanso y Ocaña
    (2016) Astudillo Calle, Jorge Eduardo; Pinos Narváez, Wilson Stalin; Quizhpe Huiracocha, Kléver Leonardo
    This present study develops a proposal to upgrade protection functions of various Intelligent Electronic Devices (IED’s) with digital technology, which protect generators and transformers of Saucay, Saymirin, El Descanso and Ocaña power plants belonging to Elecaustro S.A. Also a proposal to activate new protection functions is made, especially on Saucay, Saymirin III, IV and El Descanso plants, where electromechanical relays were replaced by IED’s in the past. To know which protection function needs to be updated or activated, it is carried out a comparison between the recommended protection functions by the standards: IEEE Std C37.102-2006 for AC generator protection, IEEE Std C37.91-2008 for transformer protection and the protection functions that currently have the installed IED’s. Also a new proposal for coordination of overcurrent functions that protect generators, transformers and transmission lines is performed, considering that all plants belonging to Elecaustro S.A are embedded in the 69kV subtransmission system belonging to EERCS. Finally, with the help of the software DIgSILENT PowerFactory, various failures are simulated at different points of the grid to make a contrast between existing and proposed trip times.
  • No Thumbnail Available
    Item
    Pronóstico de la generación de potencia en un parque eólico a través de modelos de aprendizaje automático basado en datos históricos
    (Universidad de Cuenca, Facultad de Ingeniería, 2025-11-18) Astudillo Calle, Jorge Eduardo; Minchala Ávila, Luis Ismael
    This work presents an advanced hybrid model for wind power generation forecasting that inte-grates feature engineering, seasonality analysis, and signal decomposition techniques. Initially, an exhaustive study of annual, monthly, and daily seasonality is carried out using sine and co-sine functions, which effectively capture regular cyclic patterns. However, when analyzing the complete wind power generation signal over an extended horizon, it becomes evident that it is not fully stationary. To address this nonlinear and non-stationary behavior, the CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) technique is imple-mented, decomposing the signal into Intrinsic Mode Functions (IMFs) at different temporal sca-les. This decomposition allows the specialized LSTM model to focus on learning specific high-and low-frequency dynamics independently. Validation is performed using operational data from the Huascachaca wind farm; the tuning stage employs time-series cross-validation (TSCV ), and the final evaluation uses a rolling-origin scheme with metrics expressed in physical units. Overall, the model achieves RMSE values between 546 and 680 kW and an R2 between 0.74 and 0.80 for a 48-hour forecasting horizon, demonstrating the effectiveness of the proposed approach in real-world energy forecasting applications.