1. 1 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Markos García Arregui Theoretical study of a power generation unit based on the hybridization of a fuel cell and ultracapcitors Application to the design of an aircraft electrical emergency network Theoretical study of a power generation unit based on the hybridization of a fuel cell and ultracapcitors Application to the design of an aircraft electrical emergency network Sydney, 23/09/2008. Power Engineering Group Seminar

2. 2 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Introduction (1/4) This work is the heart of the work developed by the LAPLACE laboratory in the European Union CELINA project. (DASSAULT, AIRBUS, THALES AES…) The aim of this project is to design an an aircraft power generation unit based on a fuel cell stack. It is a transportation application. Therefore, the system mass, volume, cost and hydrogen consumption must be minimzed. C E LINA

3. 3 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion RAT The aim of the project is to design the system electrical architecture, control, gas supply (Air, pure oxygen, reformate…) and size all the devices involved in a fuel cell based aircraft emergency electrical network. Introduction (2/4) In this case the emergency electrical network of the aircraft is studied. The Ram Air Turbine (RAT) and its associated electrical network will be replaced by the fuel cell system. This case was chosen due to the aggressive dynamics of the mission power profile.

4. 4 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion 050100150200250300350 10 20 30 40 50 60 70 Time(s) Power (kW) Profile mission Mean value = 40 kW Peak value = 70 kW Storage device! Hybridization! Introduction (3/4) The sizing case is the lost of all electrical generation in the aircraft. In this case the FC must guarantee the RAT functions. Three analyzed cases: Pure H 2 / Air. Hydrogen bottles boarded Pure H 2 / Pure O 2. Hydrogen and oxygen bottles boarded H 2 reformate / Air. Auxiliaries dynamic limitations Air compressor = 1 second Kerosene reformer = 20 seconds

5. 5 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Specifications:  ± 270 Volts electrical network  Mission = 4 hours  One STACK  STACK connected to ground Objectives:  System sizing  Electrical architecture choice  Mass, volume and H 2 consumption minimization  SCV design Emergency Electrical Netwotk Design Introduction (4/4) Fuel Cell N, S?? Storage ? ? ? +270V -270V Pure H 2 / Air. Reformate H 2 / Air. Pure H 2 / Pur O 2. Compresseur Air Reformeur Compresseur Air

6. 6 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Adopted methodology System optimization Generic sizing models Efficiency and mass estimations: FC SCV SC Electrical architecture choice Electrical architectures study. Energy management strategy Technological constraints Specifications

7. 7 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Objectives Design variables  Hydrogen consumption minimization  Global boarded mass minimization  10 desing variables…….OPTIMISATION Variable conceptionTypeDomaine de définition Filtering frequencyContinous0.001 ≤ f ≤ 1 [Hz] FC minimal voltageContinous50 ≤ V FC min ≤ 400 [V] FC current densityContinous0.1 ≤ J FC max ≤ 0.8 [A/cm 2 ] UC maximal voltageContinous50 ≤ V UC max ≤ 400 [V] UC technologyDiscontinuousLibrary [1………6] GAIN energy managementContinous0.1≤ G ≤ 5 V REF energy managementContinous0.75 ≤ K Vref ≤ 1 FC SCV branch numberDiscontinuous2 ≤ n b FC ≤ 8 UC SCV branch numberDiscontinuous2 ≤ n b UC ≤ 8 Switching frequencyContinous8.000 ≤ f s ≤ 25.000 [Hz] Optimization problem formulation Hybridization rate FC UC Energy management SCV

8. 8 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion Optimization porcess Genetic Algorithm NSGA-II System Sizing System Simulation Iterative sizing method MatLab SimulinkC code FC V min FC J max f UC V max UC K ref V b n s f S N C UC R No Validated results? Weight Hydrogen consumption Yes Technological constraints fulfilled? Sizing criteria fulfilled? Not valid solution? CVS PAC SC

9. 9 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion 0 50 100 150 200 250 Mass (kg)) Mass distribution CVS PAC CVS SC PAC SC H2H2 4200440046004800500052005400 350 400 450 500 550 600 650 700 750 800 850 Consumption (moles) Mass (kg) Remarks:  FC and boarded hydrogen present similar weight in the three cases  The power electronics and the boarde UC makes the difference Case : pure H 2 / Air. Hydrogen bottles boarded Conclusion :  Best option: DC BUS.  Second option 2 SCV.  Worst choice DC Optimization results Hydrogen consumption Mass DC 2 SCV SC BUS 2 SCV SD BUS DC 2 SCV SC BUS Architectures comparison

10. 10 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion  Robustness optimization study - Power profile influence - Model  Experimental set-up - Energy management strategies validation - Power converter validation Future work

11. 11 HybridationPAC CVS ArchitecturesOptimisation SC Conclusion