Nonlinear Recurrent Neural Network Predictive Control for Energy Distribution of a Fuel Cell Powered Robot

Abstract

This paper presents a neural network predictive control strategy to optimize power distribution for a fuel cell/ultracapacitor hybrid power system of a robot. We model the nonlinear power system by employing time variant auto-regressive moving average with exogenous (ARMAX), and using recurrent neural network to represent the complicated coefficients of the ARMAX model. Because the dynamic of the system is viewed as operating- state- dependent time varying local linear behavior in this frame, a linear constrained model predictive control algorithm is developed to optimize the power splitting between the fuel cell and ultracapacitor. The proposed algorithm significantly simplifies implementation of the controller and can handle multiple constraints, such as limiting substantial fluctuation of fuel cell current. Experiment and simulation results demonstrate that the control strategy can optimally split power between the fuel cell and ultracapacitor, limit the change rate of the fuel cell current, and so as to extend the lifetime of the fuel cell.

Bioenergy Based Power Sources for Mobile Autonomous Robots

“…The crucial role in the efficiency and operation stability of specified FC types play the supply of substrate to the catalyst and removal of reaction products [106,107]. As it was shown in [31], optimization of the cell design and its hydraulic control potentially can offer some power increase and lead to a reduction of its weight and size characteristics;…”

“…Microbial and enzyme BFC do not completely decompose large organic compounds to carbon dioxide and water [15,108]. It was demonstrated in [31] that the cascade system with successive use of substrate by several FCs can be a good option to increase the power of energy sources in general.…”

“…The distribution of energy between various power sources requires adaptive control and prediction of power consumption. Such energy distribution system based on a neural network model was proposed in [31] for robots with proton exchange membrane (PEM) fuel cell and a bank of ultracapacitors as power sources. 3.…”

A Feasibility Study for the Automated Monitoring and Control of Mine Water Discharges

“…ANN's have demonstrated the ability to derive highly nonlinear relationships, and they can be continually updated as more data is collected. Furthermore, this process modeling approach has been successfully implemented by a number of researchers to develop accurate model predictions in a number of application areas, including fuel cells, batteries, heat exchangers, chemical reactions, and surface water quality parameters (Yu and Gomm, 2003;Singh et al, 2009;Vasickaninova et al, 2011;Shen et al, 2013;Chen et al, 2014;Elbisy et al, 2014).…”