Novel ANFIS Based Control for Solar Energy Heliostats

Abstract

Solar central receiver or "power tower" thermal energy systems demand high precision heliostat tracking. Large numbers of heliostats in the field need to precisely track their own targets. Such a scenario suggests a hierarchical control architecture in which a model based controller could occupy the supervisory level. A novel two-level control architecture consisting of the NREL SPA (Solar Position Algorithm) and ANFIS (Artificial Neural and Fuzzy Inference System) as supervisory control is presented. The ANFIS system offers reduced complexity and allows for the quick simulation of various heliostat geometries without solving the kinematics. The initial system design, simulation and tracking results using Matlab SimMechanics and Simulink are given here.

Deep Learning Techniques for Photovoltaic Solar Tracking Systems: A Systematic Literature Review

“…Table 3 shows a summary of the data type, data source and features used in the various publications.              [100]              [101]              [102]              [103]              [104]              [105]              [106]              [107]              [108]              [10]              [109]              [110]              [111]              [112]              [113]              [114]              [115]              [116]              [117]              [118]              [119]              [120]              [121]              [122]                           [125]              [126]              [127]              [128]              [129]              [130]              [131]              [132]             ...…”

“…Away et al [101] used azimuth and elevation angles and LDR data to train an ANN-based tetrahedron sensor to track the position of the sun. Further, while the works in [104,111,115,121,128] used LDR data, the authors in [118,120,122,130] used astronomical data to their proposed models.…”

“…Variations of an ANFIS, a DHL-based model, where used by several researchers. The works in[10,98,104,105,108,120,122,123] proposed ANFIS-based models for solar tracking. These ANFIS model architectures consisted of five layers, nodes ranging from 78 to 286, parameters ranging from 126 to 530, and fuzzy rules ranging from 27 to 125.…”

A method for calibrating a single facet small heliostat based on extracting invariant moments

“…Camerabased machine vision is used in large heliostat field tracking control and calibration. Closed loop control and calibration method using photodiodes and piezoelectric actuators have been introduced by Freeman [13]. The improved accuracy and robustness come at the expense of the extra cost of sensors and control system.…”