Elsevier

Energy Conversion and Management

Volume 185, 1 April 2019, Pages 455-464
Energy Conversion and Management

A novel intelligent-based method to control the output voltage of Proton Exchange Membrane Fuel Cell

https://doi.org/10.1016/j.enconman.2019.01.086Get rights and content

Highlights

  • Overwhelming the inappropriate efficiency of primary parameters is studied.

  • Co-Evolutionary Ribonucleic Acid Genetic Algorithm is suggested.

  • Some genetic operators are proposed to conserve the diversity of individuals.

  • Two subpopulations are considered to compromise between searching and exploitation.

  • By using proposed method; reliability and dependability of the controller increases.

Abstract

The Proton Exchange Membrane Fuel Cell is a low-temperature electrochemical device that offers promising advantages such as higher efficiency compared to conventional power sources, possibly a green choice with avoiding air polluting problems. The mentioned advantages will be obtained once the Fuel Cell is accurately and efficiently controlled. Experimentally, there are two significant problems in efficiently controlling of the Fuel Cell voltage including high-speed voltage oscillations and low-speed dynamic response. As well, the co-evolution ribonucleic acid genetic algorithm is presented as a novel algorithm to obtain the optimal control parameters. This algorithm is motivated from the biological Ribonucleic Acid and encodes the chromosomes by Ribonucleic Acid nucleotide basics and accepts some Ribonucleic Acid operations. Present work adopted some genetic operators to preserve the diversity of individuals, and individuals are separated into two sets. Different evolutionary methods are considered for these two sub-populations for compromising between exploration and explanation. Primarily, a comprehensive model of Proton Exchange Membrane Fuel Cell is presented. For presenting a simple application and reliable industrial control system, this paper showed a lead-lag controller. Firstly, this control system is adjusted to a certain operational point. Despite to reveal appropriate information considering the present condition of the plant, its efficiency will be decreased by varying the situations. So, the next step is employing the proposed co-evolutionary ribonucleic acid genetic algorithm for obtaining optimum values for controller parameters versus the varying conditions as well as fault occurrences. Finally, the obtained results are presented for efficiency validation of suggested control system, and these obtained results are analyzed.

Introduction

The global concern about environmental issues as well as increasing the energy crisis has encouraged the governments to provide their energy demands through renewable energy sources. From renewable energy resources, the Fuel cells can be considered as a practical energy resource for power generation [1]. The fuel cells are the electrochemical devices that generate electricity from a series of chemical reactions. Regarding the characteristics of the fuel cells, the performance of the fuel cell can be stated as one of the most important advantages. Evidently, the performance of fuel cell is more than the traditional distributed generations, which is because of its independency to middle mechanical stages in power generation process [2]. From FC systems, the PEM fuel cell belongs to first kinds of FC systems. Proton Exchange Membrane Fuel Cell is considered as the low-temperature devices which operate at the range of 50–100 °C [3]. PEM fuel cells can generate the electricity straightly by the reactions between the hydrogen and oxygen [4]. Moreover, thermal power is a secondary production in this procedure. The PEM fuel cells also have some other merits containing:

  • (a)

    Its startup is simple and has high speed that is proper for automotive utilities.

  • (b)

    It operates properly in low-temperature situations that is appropriate for portable energy utilities.

However the stated advantages make the PEM fuel cell a resource with high reliability; still, there are some problems for this system that must be stated comprehensively. For instance, its voltage variations have high speed, and its dynamic response is low-speed versus constructional variations. On the other hand, almost all of the electronic machines need a fixed voltage signal with fast transient conditions [5].

The voltage of the PEM fuel cell must be controlled effectively to reach reliability. In this field, a precursor study is performed in [6], where a feedback controller is presented to get a fixed voltage/current in an FC accurately. As well, [7] proposed a Proportional Integral Differential (PID) controller that is designed by the model-reference method. This suggested method kept the voltage in a fixed value versus various noises and disturbances. Ref. [8] presented a neural network (NN) based controller to regulate the output voltage. Also, [9] proposed an NN-based MPC (denoting the model predictive control) system. This proposed controller is learned by presented NN to control the rate of hydrogen for achieving to an appropriate efficiency. Furthermore, some other approaches like fuzzy method are used to regulate the output voltage of FCs with reduced variations [10]. These methods are intelligent ones that are proper for development of a broad operational range with the lowest human intervention. Nevertheless, the performance of these methods has a great dependency on the used database. It translates, with deviation of the operational-point, performance of the control system will be reduced in terms of reliability and dependability. Also, [11] proposed an adaptive control method for a PEM fuel cell based on a Z-source inverter. Even though the model-based methods are mainly considered as efficient methods, they require a precise model from studied systems that is difficult to obtain. Narissara et al. [12] presented the model predictive controlling system for the voltage controlling of the PEMFC. The transient behavior of the cell is studied, and the impact of essential input variables on the voltage of the cell is examined. Fan et al. [13] proposed a Fuzzy logic controller for the voltage and current controlling of the PEMFC. Their proposed method showed an excellent steady-state response.

Present work aimed to design a high performance control system for PEM fuel cells application. The most important target is to preserve the voltage of fuel cell in a fixed value against variation of loads and probable disturbances. Here, a favorite model for PEM fuel cell is obtained considering the whole operational procedures. This extracted model also puts up the electrochemical, thermodynamics, as well as fluid concepts in the efficiency of the fuel cell. Respect to disadvantages in reviewed papers, a renowned lead-lag control system is proposed in this paper to control the voltage of PEM fuel cells. This kind of control system due to simple application and less implementation and maintenance expenses is mainly used in industrial utilities. Its factors can be adjusted in starting phase, according to the precise mathematical base in integration with trial-and-error approaches. Moreover, the suggested control system implements the coRNA-GA to give the optimal factors in on-line manner. Clearly, this manner has high performance versus variations of load as well as constructional variations of studied systems, in which the primary controller parameters aren’t able to maintain the voltage in a fixed value and favorite dynamic response. In the on-line manner, the proposed optimization method searches minima deviation in the voltage of fuel cell. Therefore, the coRNA-GA algorithm can adjust the parameters of controller in a real-time manner. So, the basic novelties of this paper can be provided as:

  • This paper overcame the improper efficiency of primary parameters in studied problem.

  • Here, coRNA-GA optimization method is proposed to adjust the control gains in a real-time manner.

  • Here, some genetic operators to preserve the diversity of individuals, and individuals are separated into two sets.

  • In the two considered sub-populations, various evolutionary techniques have been utilized to compromise between searching and exploitation.

  • By applying the suggested method, the control system will be highly reliable and dependable in wide range of oscillations.

Some numerical evaluations are considered here to validate the efficiency of the suggested method, and obtained results are analyzed.

The rest of the present work is structured in the following form. A comprehensive presentation of the PEM fuel cell model is provided in part 2. Then, the designing process of the proposed control system is given in part 3. The proposed coRNA-GA for optimally adjusting of control parameters is comprehensively described in part 4. Then, part 5 implemented the proposed optimization method to design the final control system. Finally, the conclusion of the whole paper is presented in part 6.

Section snippets

Proton Exchange Membrane Fuel Cell modeling

In this part of the work, a comprehensive definition of the PEMFC fundamental is given in the first sub-section. Furthermore, a thorough mathematical modeling of the PEMFC is presented in the second sub-section.

Controller designing

As mentioned, many factors affect the PEM fuel cell voltage. In supplement state, a variation in loading situations can generate a deviation in the output voltage of fuel cell. Whereas, preserving the output voltage in a fixed value is an elementary condition in the effective procurement of power. Therefore, the PEM fuel cell needs an efficient voltage controller to reach a favorite operating. In this way, this paper proposed the renowned lead-lag control method. This kind of controller is a

Proposed co-evolution ribonucleic acid genetic algorithm

Present work suggested the coRNA-GA approach that is represented as follows.

Results and discussion

The primary presented control system in Section 3 is improved here to an optimal on-line controller equipped with coRNA-GA algorithm. Aiming this regard, the Integral of Time Multiply Absolute Error is introduced here as an efficient metric for assessment of the PEM fuel cell voltage deviation. It translates, once the operational point has considerably deviated from the original situations, the suggested controller tries to find the optimal variables for minimization of the errors. So, the cost

Conclusion

Performance of PEM fuel cell is improved in present work with ensuring a fixed value. Also, this paper enhanced the dynamic response of PEM fuel cell voltage. Firstly, a traditionally adjusted lead-lag control system was presented for PEM fuel cell stack. Then, a step variation was applied, and designed traditional controller could efficiently preserve the voltage in a constant value. But, this control system had some weakness in facing with higher values of changes. So, an on-line optimization

Conflict of interest

The authors declared that there is no conflict of interest.

References (26)

Cited by (26)

  • Optimal dynamic operation and modeling of parallel connected multi-stacks fuel cells with improved slime mould algorithm

    2021, Renewable Energy
    Citation Excerpt :

    Microgrids (MGs) are developed through a combination of several RESs and energy storage systems (ESSs) as hybrid systems that have gained attention of researchers in their recent studies [1–3]. Recently, MG has a potential to substitute traditional power systems since many problems encounter the power networks [4]. Fuel Cells (FCs) are considered as a chief RES that uses energy of chemical reactions, and shows much promising improvement to the resiliency of energy sector.

  • Nonlinear control of a PEM fuel cell integrated system with water electrolyzer

    2021, Chemical Engineering Research and Design
    Citation Excerpt :

    The formulated MPC was implemented to a linearized PEM fuel cell model. A novel intelligent-based method (co-evolution ribonucleic acid genetic algorithm) is used to control the output voltage of PEM fuel cell (Nejad et al., 2019). The work included only the dynamics of fuel cell voltage.

  • Working zone for a least-squares support vector machine for modeling polymer electrolyte fuel cell voltage

    2021, Applied Energy
    Citation Excerpt :

    In the real application of PEFCs, voltage is easy to measure and has been widely used to gauge the performance of fuel cells. Thus, in order to monitor the performance of operating fuel cells and then achieve good performance, the establishment of an efficient, fast and reliable model that can predict PEFC system output in real-time is necessary [10]. Mathematical models derived from physical mechanisms [11] and black box models [12] utilizing collected data are two of the predominant methods of fuel cell modeling.

View all citing articles on Scopus
View full text