Dynamic neural network controller model of PEM fuel cell system

In this paper, the position of a pulse width modulation (PWM)-driven pneumatic actuator has been controlled using a dynamic neural network (DNN) and Proportional Integral Derivative (PID) controller. The harmony search algorithm (HSA) has been used to unravel the optimization problem. The DNN controller is optimally designed to control the position of the actuator. As to the performance of the PID controller, it can assist the DNN controller to give better results. Therefore, an optimal hybrid scheme with both DNN and PID controllers based on HSA is suggested. A pneumatic circuit containing a fast-switching valve is used to reduce the complexity of the PWM-driven servo pneumatic system along with its cost price.

Dynamic neural network control for non-linear systems: optimal neural network structure and stability analysis

DNNC has been successfully applied to nonisothermal CSTR with time varying parameters including activation energy and heat transfer coefficient. Thus, the objectives of the paper are: 1. To study the constraint handling properties of DNNC, 2. To extend SISO-DNNC to MIMO-DNNC, 3. To study the on-line adaptation properties of DNNC. >

An efficient Q-learning paradigm implemented on a fuzzy CMAC network is proposed. The fuzzy CMAC network topological architecture is described. First, the continuous states of the system are partitioned into a number of fuzzy boxes. Second, the proposed fuzzy CMAC evaluates the Q-values of agents in the fired fuzzy boxes and chooses control actions with maximum Q-values. Then a critic generates an external reinforcement signal according to the outcome or the effect of the control at every time-step, which is used later for further improving the estimation of these Q-values. To speed up the convergence of reinforcement learning, the traditional PID controller with several groups of different parameters is adopted so as to collect a number of taught-lessons. These taught-lessons together with the experienced lessons generated automatically, are sequentially replayed and learned, respectively, under the guidance of different reinforcement mechanisms. The hybrid adaptive and learning control system is applied to the control of a pH-neutralization process. Simulation investigations show that the fuzzy connectionist Q-learning control system has more adaptive, higher intelligence, and stronger generalization ability compared to neural network or fuzzy neural network control techniques using supervised learning.

Fuzzy Logic based Control Strategies for Proton Exchange Membrane Fuel Cell System

In the applications of the magnetostrictive actuators (MA), hysteresis of the MA is particularly significant and causes undesired effect in the control system. In order to reduce hysteresis effect and obtain precision position in actual application, a neural network supervisory control was proposed and three different neural networks were used. Hysteresis would be compensated and the precision control of the MA would be obtained. These neural networks were respectively radial basis function neural network (RBFNN), dynamic recurrent neural network (DRNN) and cerebellar model articulation controller (CMAC). Through comparing these neural networks in tracking performance, tracking speed and parameter number, the control performance is better when using DRNN and CMAC, especially CMAC with 4 parameters.

Development and utilization of environmentally friendly and more efficient energy conversion devices is necessary for sustainable development in the 21st century. A low temperature Proton Exchange Membrane Fuel Cell (PEMFC) system, fueled by natural gas is proposed and its performance is investigated by establishing a thermodynamic model of integral approach using energy and species mass conservation in addition to chemical and thermal equilibrium and electrochemical principles. Activation, ohmic and concentration potential losses are considered in cell voltage calculations. Estimate of the cost of electricity (COE) of the proposed system is also made using data fits of the available purchase costs. Computer programs using the Engineering Equation Solver (EES) supported with graphical user interface (GUI) are established for solution. Parametric studies are conducted to this system over ranges of: stack pressure, stack inlet temperature, reformer temperature, gas shift temperature and current density. Output power, electrical efficiency and the COE are calculated to investigate both performance-optimum and economic-optimum operating ranges of each input parameter.

<div class="section abstract"><div class="htmlview paragraph">The growing demand for air transport requires efficient and sustainable power systems to meet the pressing need for decarbonizing the sector. A hybrid unit, consisting of a proton exchange membrane fuel cell system and a lithium-ion battery, is a suitable option due to the advantages of reduced gravimetric and volumetric impacts, along with the flexibility of energy management strategies. This work addresses, using a model-based approach, the issue of integrating these electrochemical devices into the aircraft’s electrical architecture considering both design and energy management aspects. A literature derived DC-DC converter bi-dimensional power map is exploited to investigate scenarios differentiated by the fuel cell system power rating and number of stacks working in parallel such that the DC bus line voltage requirements can be respected. These maps relate the converter’s maximum deliverable power to the input and desired output voltage. The combined design and energy management problem is tackled via a multi-objective optimization supported by a design space exploration based on a full factorial parametric analysis. Therefore, a Pareto front is derived by considering the hydrogen consumption and overall hybrid unit mass as objective functions. As additional contributions, the auxiliaries’ power absorption, heat generated by the stack and resulting cooling load are also estimated, thus providing information for the preliminary sizing of the thermal management system. For the mission profile under consideration, the hydrogen consumption exhibited a variation ranging from 7 kg to 6.7 kg as the fuel cell system rated power increased. On the other hand, a smaller fuel cell system lowers the hybrid unit’s total mass. Particularly, when considering a configuration featuring a 70.5 kW fuel cell system, a single stack and two DC-DC converters, the total mass is estimated to be approximately 237 kg. Therefore, for the reduced impact on the aircraft’s maximal mass, the latter design choice is deemed as the most suitable option to be installed on hybrid aircraft with more electric onboard systems.</div></div>

Testing of proton exchange membrane fuel cell system is one of the most important issues due to the cost and security reasons in railway applications. To reduce the cost, it would be necessary to have a representative simulation model for this case. Having an accurate model of proton exchange membrane fuel cell system requires considering several auxiliary circuits. However, in industrial application, the accurate model of proton exchange membrane fuel cell system and corresponding information are not often provided by the designer. Moreover, the simplicity and representativity of model is required to be compatible with overall traction chain model. The model has to be suitable for system simulation and not design purpose. In this paper, a simplified model of PEMFC system is proposed using the experimental profile mission. The impact of all the environment system of proton exchange membrane fuel cell stack is estimated using the average of polarization curve on several operational points. The parameter identification process is applied, and the obtained parameters are then inserted on the overall system of fuel cell stack. The experimental data obtained through train is compared with the consecutive simulation results.

The air supply system has the great influence on the output power of Proton Exchange Membrane fuel cell system. In this paper a stair-like generalized predictive control algorithm based on multiple models switching is proposed to regulate the inlet flow rate of air for improving the output power of fuel cell system. A multiple models bank contains several fixed models and an adaptive model is used to describe the air supply system, and the best model can be selected by the optimal error functions and model switching rules. Then a stair-like generalized predictive controller is designed based on this model to control the oxygen excess ratio by adjusting the air inlet flow rate so as to improve the output power of system. The simulation results demonstrate that the proposed algorithm can raise the output power of fuel cell system, and has the better performance than traditional generalized predictive control with signal adaptive model.

Anode recirculation is essential to the pure-hydrogen proton exchange membrane fuel cell system. Keeping the pressure difference between the anode and the cathode is also important to the membrane health. In this paper, a dynamic platform was designed for the recirculation test of injection pump and real-time control of the anode pressure tracking. The test bench can work in a wide range of conditions for high- and low-pressure application. Based on the MATLAB/xPC Target environment, some S functions were written to drive the PC board for the hardware-in-loop application. Then an analytical full-order and a reduced-order model were built with good accuracy. By linearization of the nonlinear dynamic model, a linear quadratic Gaussian algorithm based on state feedback was used for set-point tracking. Moreover, an adaptive fuzzy neural network with an on-line neural network identifier was also designed to improve the control robustness. The foundation of the test bench and realization of the real-time control algorithms are meaningful to the future application in fuel cell systems.

Dynamic fuzzy neural networks modeling and adaptive backstepping tracking control of uncertain chaotic systems

An output feedback adaptive tracking scheme is proposed for a class of nonlinear systems with parameter uncertainty and input saturation. The unmeasured states are estimated by constructing filters, then dynamic surface control system is designed, which avoids explosion complexity of backstepping control. The effect of actuator saturation on system performance is compensated by neural network. It is shown that the proposed scheme can guarantee semi-global stability of the closed-loop system and the tracking error can be made arbitrarily small.KeywordsOutput feedback Dynamic surface control Input saturation Neural networks Adaptive control

Space power applications historically include fuel cells due to the high energy storage density of hydrogen and oxygen compared to batteries. Fuel cells are continuously under development to incorporate latest technology and focus on specific details of fuel cells systems relevant to harsh space transportation environments. The National Aeronautics and Space Administration is developing proton exchange membrane fuel cells systems for space power applications because of the potential for longer life, reduction in cost, and increase in safety compared to current alkaline fuel cell technology. Space fuel cell applications utilize oxygen instead of air, which introduces better performance but greater hazards. Circulation of reactants is beneficial for these systems to aid in removal of product water from the fuel cell stack and to humidify reactant fluid streams. Current space fuel cell prototype systems use a simple but effective pump for reactant recirculation known as a gas ejector. A gas ejector uses a high-pressure primary gas supply to produce suction to a secondary fluid at a lower pressure. A gas and water separator is then necessary to remove the fuel cell product water from the unutilized recirculated oxygen. The National Aeronautics and Space Administration is analyzing and testing several different means to separate the oxygen gas and water in both microgravity and increased gravity conditions. This paper addresses specific components and design concerns for proton exchange membrane fuel cell systems for space power applications.

Proton exchange membrane fuel cells are gaining increasing importance in vehicle applications. The exhaust gas composition regarding the water and oxygen content and the mass flow are important parameters in fuel cell research (e.g., for designing the test bed, quantifying the hydrogen loss in the exhaust, performing experiments with air pollutants, and monitoring degradation). The exhaust gas composition is also important for vehicle applications (e.g., ensuring safe hydrogen levels in the exhaust). Performing direct measurements of the exhaust mass flow and the relative humidity is challenging due to the high-humidity environment. This article presents a mathematical thermodynamic model used to calculate the exhaust gas mass flow and relative humidity, validated by balancing the gas species composition between cathode inlet and exhaust and by using data measured at the fuel cell system test bed. Four calculation model variations and their analyses are discussed. Furthermore, the exhaust gas composition throughout the fuel cell system operating range is presented. The results of air pollutant experiments provide comprehensive examples for the application of the calculation model. These results demonstrate the suitability of the model for its application in fuel cell system research.