Wind speed time series reconstruction using a hybrid neural genetic approach

2017 International Conference on New Energy and Future Energy System

Rodriguez, H.^1 ; Flores, J.J.^2 ; Puig, V.^3 ; Morales, L.^4 ; Guerra, A.^1 ; Calderon, F.^2

Instituto Tecnológico de Culiacán, División de Estudios de Posgrado e Investigación, Juan de Dios Bátiz 310 pte., Col Guadalupe, Culiacán

C.P. 80220, Mexico^1

División de Estudios de Postgrado, Facultad de Ingenieria Eléctrica, Universidad Michoacana de San Nicolás de Hidalgo, Michoacán, Morelia, Mexico^2

Institut de Robótica i Informática Industrial (CSIC-UPC), Carrer LLorens Artigas 4-6, Barcelona

08028, Spain^3

Facultad de Ingenieria Civil, CONACYT-Universidad Michoacana de San Nicolás de Hidalgo, Michoacán, Morelia, Mexico^4

关键词: Compact genetic algorithm;    Energy productions;    Hybrid methodologies;    Irreversible damage;    Maintenance planning;    Operational management;    Solar and wind energies;    Wind speed time series;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/93/1/012020/pdf
DOI  :  10.1088/1755-1315/93/1/012020

来源: IOP

【 摘 要 】

Currently, electric energy is used in practically all modern human activities. Most of the energy produced came from fossil fuels, making irreversible damage to the environment. Lately, there has been an effort by nations to produce energy using clean methods, such as solar and wind energy, among others. Wind energy is one of the cleanest alternatives. However, the wind speed is not constant, making the planning and operation at electric power systems a difficult activity. Knowing in advance the amount of raw material (wind speed) used for energy production allows us to estimate the energy to be generated by the power plant, helping the maintenance planning, the operational management, optimal operational cost. For these reasons, the forecast of wind speed becomes a necessary task. The forecast process involves the use of past observations from the variable to forecast (wind speed). To measure wind speed, weather stations use devices called anemometers, but due to poor maintenance, connection error, or natural wear, they may present false or missing data. In this work, a hybrid methodology is proposed, and it uses a compact genetic algorithm with an artificial neural network to reconstruct wind speed time series. The proposed methodology reconstructs the time series using a ANN defined by a Compact Genetic Algorithm.