An Advanced Fueling Algorithm The MC Method. Ryan Harty and Steve Mathison Honda R&D Americas, Inc

Transcription

1 An Advanced Fueling Algorithm The MC Method Ryan Harty and Steve Mathison Honda R&D Americas, Inc

2 Outline 2 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

3 Outline 3 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

4 Need for an advanced fueling algorithm 4 A fueling standard should work well for all tank systems, in all conditions, for any station configuration, in all ambient conditions, all the time. Guarantee safe fueling in all conditions Guarantee fast fueling in a wide range of conditions Real customers using real stations need great station performance

5 Precooling Temperature No Cooling Need for an advanced fueling algorithm 5 Some Forklift Stations Many Current Stations and Bus Stations (No Precooling) Station Operating Envelope Current Fueling Standard Advanced Algorithm 0C -20C Most of the Precooled Stations in US Today -40C Some of the Future Stations in the US Tomorrow 25MPa 35MPa 50MPa 70MPa Fill Pressure

6 Precooling Temperature No Cooling Need for an advanced fueling algorithm 6 Some Forklift Stations SAE J2601 Type D35 CAFCP Rev6.1 Many Current Stations and Bus Stations (No Precooling) Station Operating Envelope Current Fueling Standard Advanced Algorithm 0C SAE J2601 Type C35-20C -40C No Standard No Standard SAE J2601 Type B35 SAE J2601 Type A35 SAE J2601 Type B70 SAE J2601 Type A70 Most of the Precooled Stations in US Today Some of the Future Stations in the US Tomorrow 25MPa 35MPa 50MPa 70MPa Fill Pressure Do the existing fill protocols always give good fueling performance to the maximum capability of the tank system?

7 Precooling Temperature No Cooling Need for an advanced fueling algorithm 7 0C -20C -40C SAE J2601 Type D35 CAFCP Rev6.1 Advanced Algorithm SAE J2601 Type C35 SAE J2601 Type B35 SAE J2601 Type A35 SAE J2601 Type B70 SAE J2601 Type A70 Station Operating Envelope Current Fueling Standard Advanced Algorithm Target: Safe, High Quality, Fast Fills at All Stations 25MPa 35MPa 50MPa 70MPa Fill Pressure Maximize the possible performance at all conditions.

8 Outline 8 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

9 An advanced fueling algorithm 9 -should be based on the actual conditions of a hydrogen station enthalpy of delivered hydrogen (any condition) -should be applicable for all initial fill temperature, pressure internal energy in the tank (any condition) -should be based on the actual capabilities of a tank system actual volume, NWP real thermodynamic characteristics hot soak, cold soak based on test results and conditions expected by the OEM and the Station -should make use of available technology to communicate these parameters to the station. ID-Fill? IRDA? RFID? etc

10 An advanced fueling algorithm 10 Photo courtesy of Air Products and Chemicals Inc Knowable by the Vehicle: -NWP -Tank Volume -Max Hot Soak Temp -Max Cold Soak Temp -Tank Thermodynamics -Other? Knowable by the Station: -Ambient Temperature -Initial Tank Pressure -Enthalpy of Fill -Precooler Output -Starting Pressure -Ending Pressure

11 An advanced fueling algorithm 11 Data Toss (Static Communication RFID, Barcode, HVAS, ID-Fill, etc. Not integrated to ECU) Photo courtesy of Air Products and Chemicals Inc Knowable by the Vehicle: -NWP -Tank Volume -Max Hot Soak Temp -Max Cold Soak Temp -Tank Thermodynamics -Other? Knowable by the Station: -Ambient Temperature -Initial Tank Pressure -Enthalpy of Fill -Precooler Output -Starting Pressure -Ending Pressure

12 Outline 12 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

13 MC Method Development 13 Only source of energy? A control volume analysis yields several insights. u 2 m h m h i i i ( T, P) i ( T, P) m u 1 1 ( T, P) m cv ( m u 2 u mih m 2 cv ( T, P) 2 ( T, P) i ( T, P) Q m u 1 Q 1 ( T, P) ) Q h i =enthalpy at the inlet (at T,P,dm) m 1 =initial mass of hydrogen in the control volume u 1 =initial internal energy in the control volume m 2 =final mass of hydrogen in the control volume u 2 =final internal energy in the control volume Q=heat transferred across the control volume boundary Since internal energy, u, is a state property, if we know the density, and the pressure, for any given amount of heat transfer, Q, we can directly calculate the temperature. Model based on Inlet Enthalpy and Internal Energy

14 MC Method Development 14 Only source of energy h i =enthalpy at the inlet (at T,P,dm) m 1 =initial mass of hydrogen in the control volume u 1 =initial internal energy in the control volume m 2 =final mass of hydrogen in the control volume u 2 =final internal energy in the control volume Q=heat transferred across the control volume boundary Density=target density at end-of-fill T adiabatic =adiabatic temperature if no heat was transferred 0 Density, T adiabatic u A control volume analysis yields several insights. u 2 m h i m h i i ( T, P) i ( T, P) m u 1 1 ( T, P) m cv ( m u 2 u mih m 2 cv ( T, P) 2 ( T, P) i ( T, P) Q m u 1 Q 1 ( T, P) If we study the case of Q=0, the adiabatic condition, then u 2 =u adiabatic, and since u is a state property, for a given density and pressure we can directly calculate the adiabatic temperature. Adiabatic ( T, P) T adiabatic depends only on station enthalpy and initial tank conditions (Temp, Press). m u 1 1 ( T, P ) 1 1 ( m m m1 ) h i 0 ) Q ( T, P)

15 MC Method Development 15 X Q? m C ( T T 2 v adiabatic final Density, T adiabatic to T final ) Now let heat transfer occur again, and let the tank hydrogen in the tank cool to some final state, T final. Density is constant as the tank cools. m 2 =final mass of hydrogen in the control volume Q=heat transferred across the control volume boundary Density=target density at end-of-fill T adiabatic =adiabatic temperature if no heat was transferred T final =end of fill temperature C v =specific heat capacity of hydrogen at constant volume The heat transfer can be described as: Q m C ( T T ) 1 2 v adiabatic final

16 Tank Temp (C), Pressure (MPa) MC Method Development 16 Use of T adiabatic to Calculate Q: T vs Time for 35MPa Type 3 Fill at 25C T Adiabatic T Final (3min) Q End of Fill Time = 3min Time (s) m C T ( T? 2 v adiabatic final T Final (30min) T Adiabatic Fill from 2MPa start at 25C using no precooling T ) The total heat transfer from the hydrogen can be described by T adiabatic to T final T adiabatic is maximum possible temperature in the tank.

17 MC Method Development 17 Q T H 2Inside The temperature distribution inside the liner is very complex. T LinerWall T CFRPOuter m 2 T LinerW/CFRP Individual layer mass, specific heat capacity of liner, tank valve assy, carbon fiber, epoxy, etc Model Temp Distribution Time Domain Q Environment Heat Capacitance T Environment Q Environment Actual Tank 1D or 2D Heat Transfer Complex All Time Need to solve Each material M, C

18 MC Method Development 18 Q T H 2Inside T LinerWall T CFRPOuter The temperature distribution inside the liner is very complex. Simplify m 2 T LinerW/CFRP Individual layer mass, specific heat capacity of liner, tank valve assy, carbon fiber, epoxy, etc Temp Distribution Time Domain Q Environment Heat Capacitance Actual Tank T Environment Q Environment Model Model 1D or 2D Heat Transfer Lumped Heat Capacitance Complex All Time Need to solve Each material M, C

19 MC Method Development Q T H 2Inside The temperature distribution inside the liner is very complex. Q T H 2Inside Characteristic Volume (Mathematical entity Not actual mass or volume) 19 T T CFRPOuter LinerWall T m LinerW/CFRP 2 Simplify m 2 =T final= M cv T Characteristic Volume T Environment Individual layer mass, specific heat capacity of liner, tank valve assy, carbon fiber, epoxy, etc Actual Tank T Environment Q Environment =0 Model Model 1D or 2D Heat Transfer Lumped Heat Capacitance Temp Distribution Complex T=T H2Inside Time Domain All Time 3 min + Dt (final condition) Q Environment Need to solve 0 Heat Capacitance Each material M, C Combined MC Combined mass and specific heat capacity = MC Q Environment Adiabatic Boundary

20 MC Method Development 20 Q T H 2Inside m 2 =T final= M cv Combined mass and specific heat capacity = MC Characteristic Volume T Characteristic Volume T Environment =0 Q Environment Adiabatic Boundary When we make these simplifications, the heat transfer into the Characteristic Volume can be described as: Q MC T final T ) 2 And from before, the heat transfer from the hydrogen can be described as: Model Temp Distribution Time Domain Q Environment 0 Heat Capacitance ( initial Q m C ( T T ) 1 2 v adiabatic final Model Lumped Heat Capacitance T=T H2Inside 3 min + Dt (final condition) Combined MC

21 MC Method Development 21 Q T H 2Inside m 2 =T final= M cv Combined mass and specific heat capacity = MC Characteristic Volume T Characteristic Volume T Environment =0 Q Environment Adiabatic Boundary These equations can be combined: MC( Tfinal Tinitial ) m2cv ( Tadiabatic Tfinal ) And a direct analytical expression for T final can be achieved: T final Model Temp Distribution Time Domain Q Environment 0 Heat Capacitance m2cvtadiabatic MCT MC m C Model Lumped Heat Capacitance T=T H2Inside 3 min + Dt (final condition) Combined MC 2 v initial

22 Outline 22 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

23 Determining Tank Specific MC from Test Data 23 To determine MC from test data, combine equations 1 and 2, and apply the heat transfer calculated from the test data: Heat transfer from the hydrogen Q Q Q m C m 2 2 MC v ( u ( T adiabatic adiabatic u Heat transfer to the MC T final ( T final Tinitial final h i =mass averaged enthalpy at the inlet u adiabatic =adiabatic internal energy in the control volume m 1 =initial mass of the control volume m 2 =final mass of hydrogen in the control volume u 1 =initial internal energy in the control volume u final =final internal energy in the control volume Q=heat transferred across the control volume boundary T final =final temperature of the hydrogen (at 3 minutes) T initial =initial temperature of the system ) ) ) Adiabatic Internal Energy (from data) u Adiabatic ( T, P) MC m2( u ( T adiabatic final m u 1 T 1 ( T, P) Resulting MC (for the fill) u initial final ) ( m ) m 2 2 m ) h 1 i ( T, P) Units: kj/k Analyze test fill data to determine MC for that fill. Use a few fills to characterize the tank.

24 Tank Temp (C), Pressure (MPa) MC (kj/k) Determining Tank Specific MC from Test Data 24 T Adiabatic = 160C MC vs Time for 35MPa Type 3 Fill at 25C min80C MC(t) increases with time as the tank cools 10min67C 30min60C End of Fill MC = 62 m2( u MC ( T End of Fill Time = 3min adiabatic final T u initial final ) ) T P MC Basic Theory: For a given tank, the MC vs Time curve will be a similar shape for each fill, with the magnitude depending on conditions. Time (s) MC at end of fill, MC vs Time is characteristic for a given tank.

25 Determining Tank Specific MC from Test Data 25 Find a curve fit that represents the physics. MC varies with fill time, initial fill pressure, temperature, and precooling level Use Internal Energy and Time MC( Conditions, time) Constant (at 3min fill target) U U adiabatic C A g 1 init e kdt j Time adjustment of MC for >3min filling Adjustment dependent on Fill Amount (at 3min fill target), initial conditions, and precooling temp 5 Coefficients: C - Constant A - Initial Conditions g, k, j - Time Response (There are other solutions that might work better. This one is simple.) Coefficients of a standard equation are communicated to the station Basis of ID Fill

26 Temperature Condition Determining Tank Specific MC from Test Data 26 25C, No Precooling 25C, -20C Precooling Test Matrix T(180) T(300) P(180) P(300) Initial Fill Amount 2MPa 1/2 Tank For each test, Stop at 100%SOC. Take data of Temp and Pressure for 3600s. T(600) P(600) T(1200) P(1200) 180s 300s 600s 1200s At Powertech, we conducted a series fill tests. The goal was to determine the constants of the equation using 4 tests, and then use the equation to predict the outcome of a 5 th test. Tanks tested: 35MPa Type 3, 70MPa Type 4, and 70MPa Type 4 tank filled to 50MPa.

27 MC (kj/k) DMC (kj/k) Determining Tank Specific MC from Test Data 27 Conduct a few test fills at different conditions to find the constants of the equation U adiabatic kdt j MC( U, t) C A g 1 e MC DMC U minute fill >3minute fill MC vs Uadiabatic/Uinit 100 MC vs (t-180s) A C y = x R² = Fills at 50% SOC 2 Fills at 0% SOC init Uadiabatic/Uinit Result of 4 tests to characterize the tanks can determine all of the coefficients Data Model t-180s g * (1- exp -kt ) j g = 6.33E+04 k = 3.66E-09 j = 5.83E-01

28 Outline 28 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

29 Applying MC Method 29 Step 0) Toss the vehicle information to the station Data Toss (Static Communication RFID, Barcode, HVAS, ID-Fill, etc. Not integrated to ECU) Photo courtesy of Air Products and Chemicals Inc Knowable by the Vehicle: -NWP -Tank Volume -Max Hot Soak Temp -Max Cold Soak Temp -Tank MC Characteristics -Other? Knowable by the Station: -Ambient Temperature -Initial Tank Pressure -Enthalpy of Fill -Precooler Output -Starting Pressure -Ending Pressure

30 Applying MC Method 30 Step 1) Set the fill time based on Hot Soak T init =T ambient +DT Hot Soak (Similar to SAEJ2601) T T DT Set hot soak temp Vehicle Station Optimizes Fill Time for Hot Soak Yes Dt=Dt+10 Tfinal >85C No init Dt 0 m m m u h u T T init cv add initial average adiabatic adiabatic Final ambient V V m u cv initial ( T T ( m m cv initial target m m add m initial ( T initial initial add u target U MC C A U Step 2) h, P e ( T, P) initial, P adiabatic init hot initial m, P initial m cv adiabatic CvTAdiabatic MCT ( MC m C ) ), u ) g 1 e cv initial add v h average adiabatic kdt Initial ) j Calculate initial mass Set target SOC=100% Calculate additional mass Calculate initial internal energy Estimate average enthalpy to be delivered Calculate adiabatic internal energy and temperature Calculate MC Optimizes T Final Close to 85C

31 Applying MC Method 31 Step 2) Set the Pressure Target based on T init =Cold Soak (Similar to SAEJ2601) Vehicle Station T m m u h u T U MC C A U T P init init add initial Final Target T average adiabatic adiabatic CPRR V m u ambient T ( m cv m cv P( P DT m initial initial ( T m m initial ( T u h CvTAdiabatic MCT ( MC m C ) target T arget initial initial add add target cold, T, P init P initial 180s Dt, P e ( T, P) initial adiabatic init m m Final, P initial cv adiabatic h, u g 1 e cv initial ) add v ) ) and average adiabatic kdt Initial target ) j 100% SOC Set cold soak temp Calculate cold initial mass Calculate additional mass Calculate initial internal energy Estimate average enthalpy to be delivered Calculate adiabatic internal energy and temperature Calculate MC Calculate T final Calculate Pressure Target Resulting Pressure Ramp Rate

32 Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Applying MC Method to J Lookup Table: 70MPa, -20 C, 1-7kg B kg Average Pressure Fueling Target Pressure, P target (MPa) Ramp Rate, APRR Initial Tank Pressure, P0 (MPa) Step (MPa/min) 1 Information Simulation 15 20Input Simulation 60 Results 70 > 70 > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature ( C) Ramp Rate (MPa/min) Ramp Rate (min/87.5mpa) Peak Flow (g/s) no fueling B kg no fueling no fueling Hot Cold 10% 10% 157l 174l Lookup Lookup Step 2 Soak Information Fill slower slower 6.3kg 7kg > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature Ramp 1.8 Rate no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling B-70 B-70 Look-up tables from J2601 Actual Fueling Duration (min) no fueling no fueling ( C) Add intermediate leak check times: up to 10 sec after every 25MPa increase in fueling pressure no fueling5.3 no fueling kg Cold Cold (MPa/ 64.7 (min/ kg no fueling 6.5 no fueling7.2 no fueling Initial Tank Pressure, P 0 (MPa) Soak min) 87.5MPa) no fueling no fueling no fueling Step Fill 4.03 Information > > 50 no 66.4 fueling 65.0 no fueling63.8 no fueling 62.6no fueling > no fueling no no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling < no fueling no fueling 50 no fueling 25 no fueling -22.5no fueling 1.6no fueling54.0 no fueling no fueling no 50 fueling no 41.0 fueling no fueling 39.1 no fueling no fueling Hot Case 14.4Final State of Charge, SOC no fueling B (Hot Soak - No History) no fueling kg Initial 22.8 Tank 14.4 Pressure, 13.2 P (MPa) no fueling > no 70fueling no fueling no fueling no fueling > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling 50 91% 91% 90% 90% 90% 90% 90% 90% 90% 91% no fueling no fueling no fueling % 7.290% 6.490% % % 89% 90% 90% 91% 92% no fueling no fueling no fueling % % -1089% % % % % % 1.792% % 0.1 no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling % % -2089% % % % % % 1.693% % no fueling no fueling no fueling no fueling % % -3089% % % % % % 1.493% 0.5no fueling no fueling no fueling no no fueling % % -4089% % % % % % 1.493% 0.5 no fueling no fueling no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling 20 88% 88% < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling 88% 89% 89% 90% 90% 92% 93% no fueling no fueling 10 88% 88% 88% 88% 88% 89% 90% 91% 93% no fueling no fueling 0 87% 87% 87% 88% 88% 89% 90% 91% 93% no fueling no fueling % 86% 86% 86% 87% 87% 88% 89% 91% no fueling no fueling % 85% 85% 85% 85% 85% 86% 87% no fueling no fueling no fueling % 84% 83% 83% 83% 84% 84% 85% no fueling no fueling no fueling % 83% 83% 83% 83% 83% 84% 85% no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling

33 Ambient Temperature, Tamb ( C) MC Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Applying MC Method to J Lookup Table: 70MPa, -20 C, 1-7kg B kg Average Pressure Fueling Target Pressure, P target (MPa) Ramp Rate, APRR Initial Tank Pressure, P0 (MPa) Step (MPa/min) 1 Information Simulation 15 20Input Simulation 60 Results 70 > 70 > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature ( C) Ramp Rate (MPa/min) Ramp Rate (min/87.5mpa) Peak Flow (g/s) no fueling B kg no fueling no fueling Hot Cold 10% 10% 157l 174l Lookup Lookup Step 2 Soak Information Fill slower slower 6.3kg 7kg > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature Ramp 1.8 Rate no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling B-70 B-70 Actual Fueling Duration (min) no fueling no fueling ( C) Add intermediate leak check times: up to 10 sec after every 25MPa increase in fueling pressure no fueling5.3 no fueling kg Cold Cold (MPa/ 64.7 (min/ kg no fueling 6.5 no fueling7.2 no fueling Initial Tank Pressure, P 0 (MPa) Soak min) 87.5MPa) no fueling no fueling no fueling Step Fill 4.03 Information > > 50 no 66.4 fueling 65.0 no fueling63.8 no fueling 62.6no fueling > no fueling no no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling < no fueling no fueling 50 no fueling 25 no fueling -22.5no fueling 1.6no fueling54.0 no fueling no fueling no 50 fueling no 41.0 fueling no fueling 39.1 no fueling no fueling Hot Case 14.4Final State of Charge, SOC no fueling B (Hot Soak - No History) no fueling kg Initial 22.8 Tank 14.4 Pressure, 13.2 P (MPa) no fueling > no 70fueling no fueling no fueling no fueling > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling 50 91% 91% 90% 90% 90% 90% 90% 90% 90% 91% no fueling no fueling no fueling % 7.290% 6.490% % % 89% 90% 90% 91% 92% no fueling no fueling no fueling % % -1089% % % % % % % % 0.1 no fueling no Ufueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling adiabatic % % -2089% % % % % % 1.693% MC % no Cfueling no fueling bno fueling no fueling % % -3089% % % % % % % 0.5no fueling no fueling no fueling no Uinit no fueling % % -4089% % % % % % 1.493% 0.5 no fueling no fueling no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no 70fueling no fueling no fueling no fueling no fueling 20 88% 88% 88% 89% 89% 90% 90% 92% 93% no fueling no fueling 10 88% 88% 88% 88% 88% 89% 90% 91% 60 93% no fueling no fueling 0 87% 87% 87% 88% 88% 89% 90% 91% 93% no fueling no fueling % 86% 86% 86% 87% 87% 88% 89% 50 91% no fueling no fueling % 85% 85% 85% 85% 85% 86% 87% no fueling no fueling no fueling % 84% 83% 83% 83% 84% 84% 85% no fueling no fueling no fueling % 83% 83% 83% 83% 83% 84% 85% 30no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling 20 MC Method can also be used to set the target for a communications fueling. Look-up tables from J2601 can be fully described by an equation with default coefficients using the MC Method. MC vs Fill Time, SAE J2601 Type B Station 10 0 g 1 e T Final kdt j mcv CvTAdiabatic T MC mcv ( 1 Cv) MC Fill Time (minutes) Initial MCCalc

34 Ambient Temperature, Tamb ( C) MC Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Ambient Temperature, Tamb ( C) Applying MC Method to J Lookup Table: 70MPa, -20 C, 1-7kg Average Pressure Fueling Target Pressure, P target (MPa) Ramp Rate, APRR Initial Tank Pressure, P0 (MPa) Step (MPa/min) 1 Information Simulation 15 20Input Simulation 60 Results 70 > 70 > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature Ramp Rate Ramp Rate Peak Flow no fueling 40 B ( C) (MPa/min) (min/87.5mpa) (g/s) 72.3 no fueling Benefits of MC Method for no fueling SAE J2601 Hot 157l 174l kg Cold 10% 10% no fueling no fueling B kg No need for Station Types Lookup Lookup Step 2 Soak Information Fill slower slower 6.3kg 7kg > 50 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling Temperature Ramp 1.8 Rate no fueling no fueling no fueling no fueling Actual Fueling Duration (min) B-70 B-70 Allows refueling at any level of precooling no fueling no fueling Look-up tables from J2601 can be fully described by an equation with default coefficients using the MC Method no fueling no fueling ( C) Add intermediate leak check times: up to 10 sec after every 25MPa increase in fueling pressure no fueling5.3 no fueling kg Cold Cold (MPa/ 64.7 (min/ kg no fueling 6.5 no fueling7.2 no fueling Initial Tank Pressure, P 0 (MPa) Soak min) 87.5MPa) no fueling no fueling no fueling Step Fill 4.03 Information > > 50 no 66.4 fueling To 65.0 no fueling63.8 meet no fueling 62.6no fueling 60.6local 58.7 > climate no fueling no no fueling no fueling no fueling needs no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling < no fueling no fueling 50 no fueling 25 no fueling -22.5no fueling 1.6no fueling54.0 no fueling no fueling no 50 fueling no 41.0 fueling no fueling 39.1 no fueling no fueling Best efficiency operation for location Hot Case 14.4Final State of Charge, SOC no fueling B (Hot 18.7 Soak No History) no fueling kg Initial 22.8 Tank 14.4 Pressure, 13.2 P (MPa) no fueling no fueling no fueling Best 2 station cost 20 for 30 application > no fueling no fueling > 50 no no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling 50 91% 91% 90% 90% 90% 90% 90% 90% 90% 91% no fueling no fueling no fueling -30 Allows fueling for 6.4 out-of-tolerance situations 45 90% 90% 90% 89% 89% 89% 90% 90% 91% 92% no fueling no fueling no fueling % % -1089% % % % % % % % 0.1 no fueling no Ufueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no fueling adiabatic % % -2089% % % % % % 1.693% MC % no Cfueling no fueling bno fueling no fueling % % -3089% % % % % % % 0.5no fueling no fueling no fueling no Uinit no fueling % % -4089% % % % % % 1.493% 0.5 no fueling no fueling no fueling no fueling no fueling Highest SOC < -40 no fueling no fueling no fueling no fueling < -40 no fueling no fueling no fueling no fueling no fueling no fueling no 70fueling no fueling no fueling no fueling no fueling 20 88% 88% 88% 89% 89% 90% 90% 92% 93% no fueling no fueling 10 88% 88% 88% 88% 88% 89% 90% 91% 93% no fueling no fueling Fastest fill time given conditions % 87% 87% 88% 88% 89% 90% 91% 93% no fueling no fueling % 86% 86% 86% 87% 87% 88% 89% 50 91% no fueling no fueling MC vs Fill Time, SAE J2601 Type B Station With ID-Fill, fuels to the maximum capability of the tank g 1system e No need to redevelop the tables if a new pressure standard, tank type, % 85% 85% 85% 85% 85% 86% 87% no fueling no fueling no fueling % 84% 83% 83% 83% 84% 84% 85% no fueling no fueling no fueling mcv or station type -40 is 84% developed. 83% 83% 83% 83% 83% 84% 85% 30no fueling no fueling no fueling CvTAdiabatic TInitial < -40 no fueling no fueling no fueling no fueling no fueling no fueling no fueling no no no fueling no fueling TFinal MC 20 mcv Sets the bounds for communications fueling, so the vehicle ( 1T signal Cv) MC 10 does not need to control the station (just a check). 0 kdt j Fill Time (minutes) MCCalc

35 Outline 35 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

36 MC (kj/k) DMC (kj/k) Testing MC Method at Powertech 36 Using the test fills that we conducted, we generated the constants for the equation below, and applied the algorithm. U adiabatic kdt j MC( U, t) C A g 1 e MC( U) DMC( Dt) U minute fill >3minute fill MC vs Uadiabatic/Uinit 100 MC vs (t-180s) A C y = x R² = Fills at 50% SOC 2 Fills at 0% SOC init Uadiabatic/Uinit Data Model t-180s Result of 4 tests to characterize the tanks can determine all of the coefficients g * (1- exp -kx ) j g = 6.33E+04 k = 3.66E-09 j = 5.83E-01

37 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 37 (Using results of 4 fills to predict fill 5. 35MPa Type 3 Tank Result Fill 5 at different conditions) Vehicle Data Toss Station Data Read Result of 35MPa Fill 5 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure MPa 35MPa Type Type 3 3, 5MPa Start, 35C 35C Ambient, 4.8C 4.8C Nozzle Nozzle 41.3MPa Hot Soak Bound 74.3C Target 69.2C Cold Soak Bound 62.3C 196Temp Model Note: This test confirmed how difficult it is for a station to control the nozzle temperature in the real world. J2601 tolerance is unrealistic s Fill Time (s) 99.7% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

38 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 38 (Using results of 4 fills to predict fill 5. 35MPa Type 3 Tank Result Fill 5 at different conditions) Vehicle Data Toss Station Data Read Result of 35MPa Fill 5 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure MPa 35MPa Type Type 3 3, 5MPa Start, 35C 35C Ambient, 4.8C 4.8C Nozzle Nozzle 41.3MPa Hot Soak Bound 74.3C Target 69.2C Cold Soak Bound 62.3C Tank Temp Tank Pressure 196Temp Model Temp Model Note: This test confirmed how difficult it is for a station to control the nozzle temperature in the real world. J2601 tolerance is unrealistic s Fill Time (s) 99.7% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

39 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 39 70MPa Type 4 Tank Result Filled to 50MPa Vehicle Data Toss Station Data Read Result 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure MPa Type 50MPa 4 Type 2MPa 4, Start, 30C Ambient, -14.8C -14.8C Nozzle Nozzle 61.4MPa (Using results of 4 fills to predict fill 5. Fill 5 at different conditions) Hot Soak Bound 89C Target 86.7C Cold Soak Bound 83C T Model Note: Target was based on -20C Precooling. Station actually delivered -14.8C gas. So the targets shown here are adjusted to show -15C gas target s Fill Time (s) 99.6% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

40 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 40 70MPa Type 4 Tank Result Filled to 50MPa Vehicle Data Toss Station Data Read Result 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure Note: This test confirmed how difficult it is for a station to control the nozzle temperature in the real world. J2601 tolerance is unrealistic MPa Type 50MPa 4 Type 2MPa 4, 2MPa Start, 30C Ambient, -14.8C -14.8C Nozzle Nozzle 61.4MPa Note: Original targets were based on expected -20C nozzle temp with 2.5C tolerance. Targets shown here are adjusted to see what they would have been based on actual hydrogen nozzle temp of -14.8C. (Using results of 4 fills to predict fill 5. Fill 5 at different conditions) Hot Soak Bound 89C Cold Soak Bound 83C 183s Target 86.7C Tank Temp Tank Pressure T Model T Model Fill Time (s) 99.6% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

41 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 41 70MPa Type 4 Tank Result Vehicle Data Toss 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Station Data Read Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure MPa (Using results of 4 fills to predict fill 5. Fill 5 at different conditions) Result of 70MPa Fill 5 70MPa Type Test 4-17MPa Start, 25C 25C Ambient, -7.5C -7.5C Nozzle Nozzle Fill Time (s) Hot Soak Bound 81C Cold Soak Bound 70C 185s Target 76.6C T Model 95.6% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

42 Temperature (C), Pressure (MPa) Testing MC Method at Powertech 42 (Using results of 4 fills to predict fill 5. 70MPa Type 4 Tank Result Fill 5 at different conditions) Vehicle Data Toss Station Data Read Result of 70MPa Fill 5 5 MC Parameters (From Testing) Tank Volume NWP Max Hot Soak DT=7.5C Cold Soak DT =-10C Ambient Temperature Initial Tank Pressure Expected Average Nozzle Temperature Nozzle Pressure MPa MPa 70MPa Type Test 4-17MPa Start, 25C 25C Ambient, Ambient, -7.5C -7.5C Nozzle Nozzle Hot Soak Bound 81C Cold Soak Bound 70C 185s Target 76.6C T Model Tank T Model Temp Tank Pressure Fill Time (s) 95.6% SOC 1) Calculate Hot Soak Fill Time (Based on 85C) 2) Calculate Cold Soak Pressure Target (100% Density or MAWP) 3) Calculate Expected Result

43 MC (kj/k) DMC (kj/k) Testing MC Method at Powertech 43 Using the test fills that we conducted, we generated the constants for the equation below, and applied the algorithm. U adiabatic kdt j MC( U, t) C A g 1 e MC( U) DMC( Dt) U minute fill >3minute fill MC vs Uadiabatic/Uinit 100 MC vs (t-180s) A C y = x R² = Fills at 50% SOC 2 Fills at 0% SOC init Uadiabatic/Uinit Data Model t-180s Result of 4 tests to characterize the tanks can determine all of the coefficients g * (1- exp -kx ) j g = 6.33E+04 k = 3.66E-09 j = 5.83E-01

44 MC (kj/k) DMC (kj/k) Testing MC Method at Powertech 44 Additional data from verification at other conditions closely follows tank characterization data the model works. Uadiabatic kdt j MC( U, t) C A g 1 e MC( U) DMC( Dt U minute fill >3minute fill MC vs Uadiabatic/Uinit 100 MC vs (t-180s) A C y = x R² = Data of Additional Verification Fills 2 Fills at 50% SOC 2 Fills at 0% SOC init Uadiabatic/Uinit ) 0 Data of Additional Verification Fills Data Model t-180s The model describes data outside of the conditions used to generate the model! g * (1- exp -kx ) j g = 6.33E+04 k = 3.66E-09 j = 5.83E-01

45 T Final Error (K) Error in MC Method Use T Definition MC Method Model final Error T Error in Calculated Temp Using MC Method final Type 3 Tank Overestimate T final = conservative temp but slight overshoot target density Measured Temp Underestimate T final = slightly over target temp but undershoot target density. Type 4 Tank Max Error in modeling of SAE J2601 development Thermocouple Placement and Time Lag Error Thermocouple Standard Error 35MPa Fill 1 2MPa 25C 35MPa Fill 2 17MPa 25C 35MPa Fill 3 2MPa -20C 35MPa Fill 4 17MPa -20C 35MPa Fill 1R 2MPa 25C 35MPa Fill 2R 17MPa 25C 35MPa Fill 3R 2MPa -20C 35MPa Fill 4R 17MPa -20C 35MPa Fill 5 5MPa 0C 70MPa Fill 2 17MPa 25C 70MPa Fill 4 17MPa -20C 70MPa Fill 2R 17MPa 25C 70MPa Fill 3R 2MPa -20C 70MPa Fill 4R 17MPa -20C 70MPa Fill 5 22MPa -20C 50MPa Fill 1 2MPa 25C 50MPa Fill 2 17MPa 25C 50MPa Fill 3 2MPa -20C 50MPa Fill 4 17MPa -20C 50MPa Fill 1R 2MPa 25C 50MPa Fill 2R 17MPa 25C 50MPa Fill 3R 2MPa -20C 50MPa Fill 6 2MPa -20C 3K error in T final 1% error in Density MC Method yields accurate results for Type 3 and Type 4 tanks at any fill pressure

46 Tfinal Error (K) Tfinal Error (K) Sensitivity Analysis of MC Method Tfinal Error vs MC for Various Input Errors for Type 3 Tank MC (kj/k) +10% Error in MC -10% Error in MC +10K Error in TH2 Station -10K Error in TH2 Station +10K Error in Tinit -10K Error in Tinit Type 3 most sensitive to: Initial Temperature Error (large heat mass) 10K error in initial temp 6K error in final temp Tfinal Error vs MC for Various Input Errors for Type 4 Tank MC (kj/k) +10% Error in MC -10% Error in MC +10K Error in TH2 Station -10K Error in TH2 Station +10K Error in Tinit -10K Error in Tinit Type 4 most sensitive to: Inlet Temperature Error (poor heat transfer) 10K error in station nozzle temp 8K error in final temp Robust Result T Final is not sensitive to errors in MC, or Input Error

47 Outline 47 Need for Advanced Fueling Algorithm What An Advanced Algorithm Should Be MC Method Development Determining Tank Specific MC From Test Data Applying MC Method Testing MC Method at Powertech Summary and Conclusions

48 Summary An advanced fueling algorithm 48 -should be based on the actual conditions of a hydrogen station enthalpy of delivered hydrogen (any condition) -should be applicable for all initial fill temperature, pressure internal energy in the tank (any condition) -should be based on the actual capabilities of a tank system actual volume, NWP real thermodynamic characteristics hot soak, cold soak based on test results and conditions expected by the OEM and the Station -should make use of available technology to communicate these parameters to the station. ID-Fill? IRDA? RFID? etc + Fast, High SOC, Safe Fills

49 Conclusions 49 MC Method shows excellent promise to enhance SAEJ2601, through the use of additional information to improve vehicle fueling (fast fill, high SOC). MC Method enables the use of thermodynamic parameters of the vehicle tank, plus actual station conditions, to allow the best possible fill for a given vehicle (shortest fill time and highest SOC). MC Method allows fueling at conditions that operate outside of current J2601 tables, such as new stations with -10C precooling, or already existing stations, and works at any fill pressure, for any tank system. The equation with coefficients derived from J2601 tables can replace the look-up table-based approach of SAE J2601. MC Method provides for station design flexibility and fill optimization for each vehicle.

50 Thank you for your attention! Questions? Please contact us: Ryan Harty Steve Mathison (Meet us at the Honda Booth at 5:35pm!)

51 Summary An advanced fueling algorithm Default 51 Fueling Methods 70MPa) using -20C max pre-cooling Non-Comm J2601 Type B (-20C) Non-Comm using MC Method (-20C max but variable) Full-Comm J2601 Type B (-20C) Full-Comm using MC Method (-20C max but variable) ID Fill with MC Method (-20C max but variable) Type 3 Fill Time X X Type 3 SOC Type 4 Fill Time X X * Type 4 SOC Station Cost Station Reliability Station Design Flexibility * Majority of tank systems ** Station throughput improves since customer can still fuel if pre-cooling capability falls outside the tolerance allowed in J2601. Cost also improves since pre-cooling system can be downsized and operated at optimal temperature. * X X ** * ** ** J2601 fill using MC adjustment for fill temp Station Design Flexibility! MC Method fill gives Full Comm SOC and Fill Time, without active temperature and pressure monitoring, at lower cost stations. X = No Good = Ok (acceptable) = Good = Excellent

52 Characteristics of MC 52 MC is not directly a physical constant such as the mass and the specific heat capacity of the liner material. Rather it is a composite of many things. All of the intricacies of the temperature distribution of the wall of the tank get washed out by using a parameter such as MC, especially over a time scale such as a hydrogen tank refueling. MC model cannot be used for phenomena that occur over very short time Systems with slow heat transfer characteristics (convection or conduction) will result in small values of MC (such as Type 4 tanks), because they absorb small quantities of heat in the 3 minute time domain. Systems with faster heat transfer characteristics (convection or conduction) will result in high values of MC (such as Type 3 tanks), because they absorb small quantities of heat in the 3 minute time domain. MC is a function of many things; MC=MC(time, fill conditions, tank materials, etc). But for a given tank, fill time, and conditions, MC will always be the same. The trend of MC with time is predictable. The trend of MC with fill conditions is predictable.