“Modelo de optimización para el planeamiento multietapa coordinado del sistema generación –transmisión con pronóstico de demanda mediante redes neuronales”

Due to the growing energy demand and the importance of adequate and efficient planning in electrical systems, the implementation of a multistage planning model for the expansion of generation-transmission is proposed. This approach represents an alternative to traditional independent planning models, which generate additional costs and face limitations regarding to current energy demand. In this research work, a multistage DC planning model is implemented in the generation-transmission expansion for the 230 kV ring of the National Interconnected System (NIS) of Ecuador with the objective of minimizing the costs of operation and construction of new infrastructure. The applied model, mixed integer nonlinear programming (MINLP), analyzes the expansion of electrical generation-transmission for the period 2023-2033. As a starting point and strength of this research, the demand forecasting is carried out through a neural network trained with a historical database using different neural network models (LSTM, LSTM MANUAL, SARIMA model, GRU), each with specific characteristics. The best demand prediction in the four models is used as the basis for the multistage DC model. Finally, the approach to demand forecasting with neural networks will allow the identification of expansion zones in the model, which in turn will significantly reduce long-term investment and operating costs.

• 1. autor: Chasi Cando, Alex Fernando (author)
• Kolejni autorzy: Puetate Mancheno, Vicky Johanna (author)
• Format: bachelorThesis
• Język: spa
• Wydane: 2024
• Hasła przedmiotowe: SISTEMA NACIONAL INTERCONECTADO; ANILLO DE 230 KV; REDES NEURONALES; GENERACIÓN; ELECTRICIDAD
• Dostęp online: http://repositorio.utc.edu.ec/handle/27000/11978