Nonparametric Correlative-Probabilistic Microgrid Power Energy Management Based Sine-Cosine Algorithm
In modern power systems, microgrids play a pivotal role with several economical, technical, and environmental benefits. However, there are still challenges that need to be properly addressed, including: i) accurate modeling of the uncertain parameters behavior, ii) considering the correlation betwee...
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
IEEE
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/c412a3e0db4c46839a02f98f668bd023 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | In modern power systems, microgrids play a pivotal role with several economical, technical, and environmental benefits. However, there are still challenges that need to be properly addressed, including: i) accurate modeling of the uncertain parameters behavior, ii) considering the correlation between the random variables, and iii) find the optimal solutions with low computational burden. To address these issues, this paper proposes a nonparametric-correlative stochastic framework for microgrids (MGs) energy management. The proposed method imposes no assumption on the probability density function of renewable generations and electrical load consumption. To this end, an improved kernel density estimator (IKDE) is presented to estimate the probability density function (PDF) of uncertain parameters, e.g., renewable generations and load. To account for the correlation between the uncertain parameters, Cholesky decomposition is utilized. Furthermore, a multi-objective MG energy management problem considering reliability has been reformulated, and to solve the problem, a sine-cosine optimization algorithm (SCOA) is developed. Numerical results demonstrate the effectiveness and superiority of the proposed stochastic framework through comparison with several optimization algorithms by reducing the total cost of MG more than 11% in comparison with several metaheuristic algorithms and stochastic frameworks with less than 1% error. |
|---|