OPTIMUM ENERGY MANAGEMENT OF PEM FUEL CELL SYSTEMS BASED ON MODEL PREDICTIVE CONTROL

Transcription

1 E3S Web of Conferences (017) OPTIMUM ENERGY MANAGEMENT OF PEM FUEL CELL SYSTEMS BASED ON MODEL PREDICTIVE CONTROL Chrysovalantou Ziogou (1) Michael C. Georgiadis () Syros Voutetakis (1) Simira Paadooulou (3) (1) Chemical Process and Energy Resources Institute (CPERI) Centre for Research and Technology Hellas (CERTH) Thessaloniki Greece () Deartment of Chemical Engineering Aristotle University of Thessaloniki Thessaloniki Greece (3) Deartment of Automation Engineering Alexander Technological Educational Institute of Thessaloniki Thessaloniki Greece ABSTRACT This work resents an otimum energy management framework which is develoed for integrated Polymer Electrolyte Membrane (PEM) fuel cell systems. The objective is to address in a centralized manner the control issues that arise during the oeration of the fuel cell (FC) system and to monitor and evaluate the system s erformance at real time. More secifically the oeration objectives are to deliver the demanded ower while oerating at a safe region avoiding starvation and concurrently minimize the fuel consumtion at stable temerature conditions. To achieve these objectives a novel Model Predictive Control (MPC) strategy is develoed and demonstrated. A semiemirical exerimentally validated model is used which is able to cature the dynamic behaviour of the PEMFC. Furthermore the MPC strategy was integrated in an industrial-grade automation system to demonstrate its alicability in realistic environment. The roosed framework relies on a novel nonlinear MPC (NMPC) formulation that uses a dynamic otimization method that recasts the multivariable control roblem into a nonlinear rogramming roblem using a warm-start initialization method and a search sace reduction technique which is based on a iecewise affine aroximation of the variable s feasible sace. The behaviour of the MPC framework is exerimentally verified through the online deloyment to a small-scale fully automated PEMFC unit. During the exerimental scenarios the PEMFC system demonstrated excellent resonse in terms of comutational effort and accuracy with resect to the control objectives. 1. INTRODUCTION AND MOTIVATION The shift towards a low carbon efficient and secure economy requires targeted deloyment of innovative technologies and increased exloitation of renewable energy sources. Furthermore it is recognized that a technological shift and the develoment of new clean technologies are vital for a successful transition to a decarbonized and sustainable future economy. Although a number of diverse technologies exist that aim at the same target the synergy between the increased use of renewable energy sources renewable hydrogen and electricity from fuel cells reresent one of the romising ways to realize sustainable energy. Fuel cell and hydrogen technologies have the otential to contribute to the ambitious energy and climate objectives of the Euroean Union for 00 which are the reduction of the greenhouse gas emissions by 0% the increase of the share of renewable energy to 0% and the imrovement of the energy efficiency by 0% [1]. 1.1 Imortance of Control Strategies to FC systems In order to take advantage of FC s great otential a number of technological challenges must be addressed that require significant efforts of research and develoment. Continuous research efforts are imerative for the realization of FC system-level goals for erformance redictability stability and other roerties through aroriate analysis design and imlementation. In this context the otimum energy management through aroriate control methods can function as a catalyst that transforms technological innovation to systems engineering novelties. The imact of control technology is evident in a wide range of alication areas including fuel cells as it is the necessary facilitator for achieving desired objectives and fulfilling alication-secific goals []. Fuel cell systems exhibit fast dynamics nonlinearities and uncertainties that constitute challenges requiring aroriate control in order to be confronted effectively. The use of efficient control strategies would not only increase the erformance of these systems but would increase the number of oerational hours as their lifetime is reserved by oerating at otimal levels and also reduce the cost er roduced kilowatt-hour. Overall control can be considered as an enabling technology for the otimum energy management of fuel cell systems. Based on this motivation initially the exerimental system is described along with the basic modelling aroach (Section ). Subsequently the advanced control (Section 3) and structure of the online energy management framework is discussed (Section ) which is deloyed to the PEMFC into consideration. The Authors ublished by EDP Sciences. This is an oen access article distributed under the terms of the Creative Commons Attribution License.0 (htt://creativecommons.org/licenses/by/.0/).

2 E3S Web of Conferences (017). PEM FUEL CELL EXPERIMENTAL SETUP AND MODELING In this work a small-scale fully automated fuel cell unit was used which was designed and constructed at CPERI/CERTH. Fig. 1 resents the simlified rocess and instrumentation (P&ID) diagram of the exerimental PEMFC along with the measured variables and the control actions that will be determined by the the advanced control scheme. The variables that constitute the Inut Outut (I/O) field of the unit are acquired by a Suervisory control and Data Acquisition (SCADA) system. The PEM fuel cell is by Electrochem has a Nafion 1135 (89m thickness) membrane electrode assembly (MEA) with an active area of 5cm while the ower is drawn from the fuel cell through a DC load (KIKUSUI ). The air and the hydrogen are sulied from ressurized cylinders by two mass flow controllers (MFCs). Hydrators and heated lines are used to rovide the roer humidity of the inlet gases. The temerature is maintained by an air cooling subsystem and an electrical resistance. Figure 1. Simlified P&ID of CERTH/CPERI s PEMFC unit The objective is to manage the roduced energy by roer maniulation of the four distinct yet interacting subsystems which are related to the ower roduction the gas suly the temerature and the water management. Each subsystem has its own control objective and a number of sensors and actuators are used to monitor the system s behavior. The inlet gasses the FC temerature and the ower are controlled by the roosed model-based framework while the hydrator s temerature and the heated lines are controlled by conventional PIs due to their slow dynamic behaviour. All other system comonents (ums heaters valves etc) are controlled by the SCADA system..1 Nonlinear Dynamic Model The fuel cell unit (Fig. 1) was modelled by an exerimentally validated semi-emirical nonlinear dynamic model [3] which considers the mass dynamics in the gas flow channels the gas diffusion layers (GDL) and the membrane. Also the resective electrochemical equations are used for the calculation of the voltage and an energy balance differential equation is integrated to cature the temerature evolution. An illustrative reresentation of the involved variables and main comonents of the PEMFC model is show at Fig.. I T m anch in anch in anch in N v angdl v anch H anch N v cagdl v cach o cach m m van H an m m m vca O ca N ca T m cach in cach in cach in N vmem vangdl v cagdl T fc N vmem T fc N m van h GDL Nv ca m o GDL T fc Tfc H angdl v cagdl o cagdl Figure. PEMFC mathematical model structure The develoed model coules both the theoretical validity (mechanistic equations) and the inherent simlicity of the emirical alication (semi-emirical equations). The described dynamic model is not system deendent as it can be adjusted to describe any other PEM fuel cell system by erforming a number of sequentially executed actions. These actions are determination of - the hysical characteristics of the PEM fuel cell - the unit secific arameters - the emirical arameters utilized by the electrical subsystem of the model. The value of the emirical arameters are determined by a systematic arameter estimation rocedure as long as exerimental data or data from the manufacturer are rovided. 3. ADVANCED MODEL-BASED PREDICTIVE CONTROL STRATEGIES The objective of advanced control strategies is to achieve a set of redefined conditions for the rocess and maintain the oeration at the desired or otimal values []. Model redictive control (MPC) also known as receding horizon control (RHC) is becoming a referred control strategy for a large number of industrial rocesses [5]. The main reasons for the increased oularity include the ability to handle state and inut constraints dynamic nonlinearities of the rocess into consideration and simultaneous satisfaction of economical and oerating constraints under an integrated otimization methodology. Moreover recent advances in otimization enable the move towards direct online otimizing control. These characteristics in combination with the demanding oeration of a PEMFC signify that the alication of NMPC to such V fc

3 E3S Web of Conferences (017) systems is a suitable aroach. In the current work an MPC formulation is selected to effectively handle the interrelated control objectives of a PEMFC. In general MPC refers to a methodology which makes exlicit use of a rocess model to otimize the future redicted behavior of a rocess. The main objective is to obtain control actions that minimize a cost function related to selected objectives or erformance indices of the system. At each samling time an otimal control roblem is solved using measurements acquired from the system and it yields the aroriate control inuts for the system. MPC comutes online a finite-time constrained otimization roblem over a rediction horizon ( T ) using the current state of the rocess as the initial state. The otimization yields an otimal control sequence ( uk.. uk Nc) over a control horizon ( T c ) which is artitioned into N c intervals and only the first control action ( u k ) for the current time is alied to the system. We consider the following formulation of the MPC roblem [6]:. INTEGRATED ONLINE ENERGY MANAGEMENT FRAMEWORK Overall the online energy management framework used in this work has a number of interacting entities that form a comuter-aided latform for the monitoring and control of the PEMFC unit [9]. The information flow from the I/O field to the end user is shown at Fig. 3. Furthermore Fig. 3 illustrates the interconnection of the control system s entities and the I/O field including the concetual flow of information from the signal acquisition to the final resentation at the user level through the develoment of a grahical interface. N Nc1 T T ( ˆ ˆ k j s k j) QR( k j s k j) klr kl j1 l0 min J y y y y u 1u u (1) s.t.: (1a) e ( y y ) (1b) k meas red k y y e (1c) ˆk j red k j k where uyx are the maniulated the controlled and the state variables ymeas yred y s are the redicted the measured variables and the desired set-oints and Q R are the outut tracking and the inut move weights. The minimization of functional J (Eq(1)) is subject to constraints of u and y. In this work two MPC-based strategies are combined in a unified control framework in order to take advantage of their synergistic benefits [7]. The first methodology is an online Nonlinear Model Predictive control (NMPC) strategy which is very aealing due to its ability to handle dynamic nonlinearities of the rocess under consideration [6]. The second methodology is an exlicit or multi-arametric Model Predictive Control (mmpc) strategy. The mmpc can rovide the otimal solution in real-time as the solution is comuted offline and can be imlemented online by simle look-u functions [8]. Both aroaches handle state and inut constraints and satisfy oerating objectives. These characteristics in combination with the tight oeration of a PEMFC signify that an advanced control framework is a suitable aroach. Figure 3. Information flow of the Energy Management Framework of CERTH/CPERI s PEMFC unit The MPC framework is deloyed online to the fuel cell unit using a custom made OPC-based interface that was develoed for the communication between the otimizer and the SCADA system [8]. The develoed framework has the following desired features: Fast calculation of the otimal control actions while taking into account the hysical and oerating constraints. Flexibility to adat to changing fuel cell resonse under the influence of disturbances or during start-u and shutdown. Incororation of a multitude of erformance criteria under strict comutational time demands. Easily deloyable and maintainable control and otimization solutions. As far as the MPC related arameters are concerned the rediction horizon is selected to be sufficiently long with resect to the dynamics of the fuel cell (T =5sec) divided into N intervals. The control horizon (T c ) was set to be equal to the samling time of the SCADA system (500ms) and it is divided into N c. Finally the selected otimization method is the direct transcrition using a reduced gradient NLP solver. The nonlinear fuel cell model is discretized based on OCFE. More secifically there are 10 finite elements (NE) with collocation oints (N co ) each. 3

4 E3S Web of Conferences (017) 5. PEM FUEL CELL EXPERIMENTAL RESULTS Overall the objectives for the control system are to effectively address the issue of ower generation in an otimum manner. In this context the otimality is defined by the following three terms: Oeration at a safe region regardless of the load fluctuations. Minimization of the fuel consumtion and air suly. Maintenance of stable temerature conditions ensuring roer gas humidification. The safe oeration is maintained by controlling the reactants at a certain excess ratio level in order to avoid starvation caused by sub-stoichiometric reaction conditions at the cathode and the anode. The safe oerating region for the cathode and the anode is defined by two unmeasured variables the oxygen and hydrogen excess ratios ( ) exressed as the ratios of the inut flow of each gas to the consumed quantities er unit time due to the reaction [10]: where () (3) are the oxygen and hydrogen inut flows at the channels while are the resective reacted quantities. In order to reach the required excess ratio set-oint the air and hydrogen flows ( ) are used as maniulated variables. The safety of the oeration is ensured by maintaining the excess ratios above one ( SP >1 SP >1). The ower generation and the starvation avoidance objective can be achieved by control actions that aim at an accurate set-oint tracking of P SP OSP SP. On the other hand the temerature control (T fcsp ) involves two mutually exclusive subsystems one for the heat-u and another for the cooling 5.1 Online PEMFC Oeration An indicative result form the online oeration of the PEMFC is resented. The resonse of the MPC strategy to various ower demands was studied at secific oerating conditions of temerature and ressure (T=338K Pt=1bar) is shown in this section. The ower modifications were within a range of 1W to 5W which covers the oerational range of the system. The ower resonse to the corresonding control actions of the maniulated variables (Fig. 5) as the ower demand changes while the oxygen and hydrogen excess ratio remains at a constant set-oint ( s = 3 s =.5). Power (W) 5 3 Power Powe SP Figure. Power demand changes and delivered ower by the NMPC controller Flows (cc/min) Current (A) Air Flow H Flow 0 Figure 5. Control actions for the maniulated variables of (a) Current and (b) air and hydrogen mass flows derived by the NMPC controller The NMPC controller is able to control the fuel cell ower to any admissible set-oint while maintaining the required excess ratio level. The mean squared error (MSE) for each objective and some erformance metrics are resented at Tab 1. Table 1. Performance of the Online NMPC framework MSE from Set-oint Power: 0mW Temerature: 0.6 C O : 1.5*10-3 :.1*10-3 Performance Metrics Max. ot. time: 19ms Average ot. time (SP change): 30ms Average ot. time (steady state): 75ms The NMPC controller that was deloyed to the PEMFC system can efficiently address the issues of ower generation in a safe and controlled manner. Overall it has been clearly illustrated that the NMPC framework is able to deal with uncertainties and achieve trajectory tracking in a satisfactory manner. Also the NMPC controller has comutational requirements that satisfy the time samling interval without convergence failures while satisfying realistic constraints imosed by the nature of the PEM fuel cell system. The resented initial results demonstrate that the roosed control method can be alied online at

5 E3S Web of Conferences (017) realistic time frame and enable the achievement of multile often conflicting objectives in a dynamically changing environment. Furthermore due to the nonlinear nature of the model the NMPC aroach is caable to adjust to the fuel cell degradation that will shift the oeration of the system out of the initial area of oeration in time. 6. CONCLUSIONS In this work an integrated energy management aroach that relies on a MPC strategy is resented using an advanced MPC strategy to address the oeration objectives of a PEMFC unit. A synergy that reveals the otential of exloiting at best the intrinsic features of two model-based control methods (exlicit and nonlinear MPC) and the handling of an interesting control roblem that involves the otimum oeration of an exerimental PEM fuel cell unit is demonstrated by the roosed framework. A challenging multivariable nonlinear control roblem with measured and unmeasured variables that involves concurrently four oeration objectives for a PEM fuel cell system is addressed. The roosed framework is deloyed to the exerimental fuel cell and it is validated online. The erformance of the MPC strategy shows romising behavior in terms of comutational demands and at the same time it guarantees that the PEMFC system oerates at an otimum manner. 6. Allgöwer F. Findeisen R. Nagy Z. K. (00). Nonlinear Model Predictive Control: From Theory to Alication J. Chin. Inst. Chem. Engrs. 35 (3) Ziogou C. Pistikooulos E. N. Georgiadis M. C. Voutetakis S. Paadooulou S. (013). Emowering the erformance of advanced NMPC by multi-arametric rogramming An alication to a PEM fuel cell system Industrial Engineering and Chemistry Research 5 (13) Bemorad A. Morari M. Dua V. Pistikooulos E. N. (00). The exlicit linear quadratic regulator for constrained systems. Automatica Ziogou C. Paadooulou S. Georgiadis M. C. Voutetakis S. (013). On-line nonlinear model redictive control of a PEM fuel cell system Journal of Process Control 3() Pukrushan J. T. Peng H. and Stefanooulou A.G. (00). Control-oriented modeling and analysis for automotive fuel cell systems. Journal of Dynamic Systems Measurement and Control- Transactions of the ASME 16(1) REFERENCES 1. Euroean Commission (011). A roadma for moving to a cometitive low carbon economy in 050 COM(011) 11 Brussels.. CSS (011). The Imact of Control Technology. T. Samad and A.M. Annaswamy (eds.) IEEE Control Systems Society (Accessed March 016). 3. Ziogou C. Voutetakis S. Paadooulou S. Georgiadis M.C. (011). Modeling simulation and exerimental validation of a PEM fuel cell system Comuters and Chemical Engineering 35 (9) Qin S. and Badgwell T. (003). A survey of industrial model redictive control technology. Control Engineering Practice 11(7) Bauer M. and Craig I.K. (008). Economic assessment of advanced rocess control A survey and framework. Journal of Process Control 18 (1)