Triviño Castañeda, Valentina, Escobar Mejía, Andrés, and Giovanny Valdés, Henry
Abstract
The Comisión de Energía y Gas (CREG) published in 2016 the guidelines to calculate the reliability charge for solar plants. The methodology indicates that in order to calculate the base load power for the reliability charge, it is required ten years of information with solar radiation and ambient temperature. In Colombia there is not such information, thus it is required to use strategies to estimate the required information from available data. In this paper Artificial Neutral Networks are used to estimate solar radiation considering ambient temperature, relative humidity and month of the year. [ABSTRACT FROM AUTHOR]
Montoya Giraldo, Oscar Danilo, Ramírez Vanegas, Carlos Alberto, and Grisales Noreña, Luis Fernando
Abstract
In this paper describes the design and simulation of a two-stage hybrid controller for the inverted reaction wheel pendulum (RWP) is presented. In the first stage, the general pendulum arm modeling is performed through a nonlinear model to determine the stored energy in the plant, and using a strategy known as swing up energy regulation, the data required for the training of an artificial neural network are obtained. In the second stage, via a soft switching system, neuronal control is exchanged by a controller based on the extended linearization of the state variables, whereby the permanence of the pendulum in the region of operation is guaranteed. The control strategy proposed shows excellent performance against external disturbance phenomena and ensures the overall operation of the physical system. [ABSTRACT FROM AUTHOR]