Abstract:A fuel cell system must output a steady voltage as a power source in practical use. A neural network (NN) based model predictive control (MPC) approach is developed in this work to regulate the fuel cell output voltage with safety constraints. The developed NN MPC controller stabilizes the polymer electrolyte fuel cell system's output voltage by controlling the hydrogen and air flow rates at the same time. The safety constraints regarding the hydrogen pressure limit and input change rate limit are considered. The neural network model is built to describe the system voltage and hydrogen pressure behavior. Simulation results show that the NN MPC can control the voltage at the desired value while satisfying the safety constraints under workload disturbance. The NN MPC shows a comparable performance of the MPC based on the detailed underlying system physical model.

What problem does this paper attempt to address?

This paper aims to solve the problem of unstable output voltage in polymer electrolyte fuel cell systems (PEFC) in practical applications. Specifically, the paper develops a neural - network - based model predictive control (NN MPC) method for regulating the output voltage of fuel cells under safety - constraint conditions. This method stabilizes the output voltage of the system by simultaneously controlling the flow rates of hydrogen and air, and takes into account safety - constraint conditions such as hydrogen pressure limits and input rate - of - change limits. The goal of the paper is to show that NN MPC can meet these safety constraints while controlling the voltage, and its performance is comparable to that of MPC based on a detailed system physical model. ### Main Research Contents: 1. **Problem Background**: As an environmentally - friendly and high - energy - efficiency power source, PEFC is becoming more and more popular in fixed, portable and transportation applications. However, its commercialization still faces some technical challenges, especially the problem of reliable operation. Control algorithms play an important role in ensuring the reliability of PEFC systems, but the complexity and safety limitations of PEFC systems require more advanced control methods. 2. **Method Introduction**: The paper proposes a neural - network - based model predictive control (NN MPC) method for controlling the output voltage of PEFC systems. This method describes the behavior of system voltage and hydrogen pressure by establishing a neural network model, and realizes stable control of voltage by optimizing control inputs (hydrogen and air flow rates). 3. **Model Construction**: The inputs of the neural network model include hydrogen flow rate \( Q_{\text{H2}} \), air flow rate \( Q_{\text{air}} \) and current \( I \), and the outputs are the output voltage \( V_{\text{FC}} \) of the fuel cell and hydrogen pressure \( P_{\text{H2}} \). After linearization processing, the model is used to construct the state - space model of MPC. 4. **Control Algorithm**: The NN MPC algorithm obtains the optimal control input by solving a quadratic programming (QP) problem. The goal of the QP problem is to minimize the error between the output voltage and the reference value, while considering the safety constraints of input rate - of - change and hydrogen pressure. 5. **Simulation Results**: The simulation results show that NN MPC can stabilize the output voltage at the target value under workload perturbations, and meet the safety limits of hydrogen pressure in most cases. Compared with MPC based on a physical model, NN MPC has a slight constraint violation in some step - current - increase scenarios, but the overall performance is comparable. ### Conclusion: The paper successfully develops a neural - network - based model predictive control method for controlling the output voltage of PEFC systems. This method can effectively stabilize the output voltage on the premise of meeting safety constraints, and its performance is comparable to that of MPC based on a detailed physical model. Although there are slight constraint violations in some cases, NN MPC shows its potential in complex - system control.

Neural network based model predictive control of voltage for a polymer electrolyte fuel cell system with constraints

A Thermal Management Strategy for Proton Exchange Membrane Fuel Cell via Nonlinear Model Predictive Control

Multiple Model Predictive Control for Oxygen Starvation Prevention of Fuel Cell System

Voltage Regulation in Polymer Electrolyte Fuel Cell Systems Using Gaussian Process Model Predictive Control

Integrated Motion and Powertrain Predictive Control of Intelligent Fuel Cell/Battery Hybrid Vehicles

A Combined Voltage Control Strategy for Fuel Cell

Neural-Network-Based Predictive Control for PMSM Drives

Robust composite adaptive neural network control for air management system of PEM fuel cell based on high-gain observer

Design Of Model Predictive Controller Based On Neural Systems

Model predictive control with thermal constraints for fuel cell hybrid electric vehicle based on speed prediction

Model-based Fuel Pressure Regulation Algorithm for a Hydrogen-Injected PEM Fuel Cell Engine

An energy management strategy for fuel cell hybrid electric vehicle based on a real-time model predictive control and pontryagin's maximum principle

Control of fuel cell air supply system based on LPV model

Design and implementation of model predictive control for an open-cathode fuel cell thermal management system

Local Control of Fuel Cell Systems Within Hybrid Renewable Energy Generation Using Model Predictive Control

Constrained computationally efficient nonlinear predictive control of Solid Oxide Fuel Cell: Tuning, feasibility and performance

System Modeling and Temperature Control for a Fuel Cell System Based on Local Model Networks

Model based control strategies to control voltage of Proton Exchange Membrane Fuel Cell

Predictive Energy Management for Fuel Cell Hybrid Vehicles Considering Efficiency and Safety

Predictive energy management of fuel cell plug-in hybrid electric vehicles: A co-state boundaries-oriented PMP optimization approach

An ultrahigh synchronous drive step-up converter for PEMFC and its explicit model predictive control: A neural network fitting strategy