International Symposium on DIAGNOSTIC TOOLS FOR FUEL CELL TECHNOLOGIES ONLINE PEMFC STACK MONITORING WITH “THDA“ Erich RAMSCHAK, AVL List GmbH Trondheim, 23rd – 24th June 2009
BASICS OF THE APPROACH “AVL THDA” EFFECTS RESULTING FROM VOLTAGE DRIFTS ARE ANALYZED INSTEAD OF VOLTAGE DRIFT MEASUREMENT ITSELF If defects or critical conditions occur in one or few cells, time variant conditions or local non-linearities in the transfer function distort a superimposed signal and form harmonics • Extra spectral components (i.e. harmonics) are detectable in the stack voltage – no cell voltage measurement required • Reduced measurement effort: stack voltage & stack current only • COST EFFICIENT APPROACH “THDA THDA™ ™” ” – – Total Harmonic Distortion Analysis* Total Harmonic Distortion Analysis* “ *AVL patent, registered trademark �������� �� Trondheim, 23rd – 24th June 2009 �
BRIEF BACKGROUND OF HARMONIC DISTORTION 0 amplitude [dB] -50 ���������������� ������������������� -100 ���������������� 0 2 4 6 8 10 12 � �� ������ � �� ������ ����������� normalised frequency scale ≡ order of harmonics �������� �������� ��������� � Signal distortion forms extra components in the frequency spectrum. Harmonics can be detected at integer multiplies of the fundamental frequency � Occurrence of extra spectral components is measured � Quantification: “Klirrfactor” or “THD”, [%] [dB] �������� �� Trondheim, 23rd – 24th June 2009 �
THDA INSTRUMENTATION PRINCIPLE battery pack inverters, converters, etc. DC fuel cell system DC DC DC DC +12V AC stack FC control FC control current i AC THDA device stack voltage M µC µ C CAN 2.0 ≈ ≈ ≈ ≈ ≈ ≈ ≈ ≈ signal de- signal analysis source coupling STAND ALONE APPROACH � Superimposition of a small AC signal * � (Spectral-)Analysis of voltage in terms of extra spectral components i.e. harmonic distortion * i AC = typ. 1A ( � 1mV/cell, sinusoidal, low frequency) �������� �� Trondheim, 23rd – 24th June 2009 �
FULLY INTEGRATED INSTRUMENTATION PRINCIPLE inverters, converters, ect. battery pack & signal modulation DC fuel cell system ≈ ≈ ≈ ≈ DC a. DC i FC i AC DC DC +12V AC FC control w/ THDA b. M COST EFFICIENT APPROACH a. Modulation of specific current signal pattern by converter b. Embedded signal distortion analysis by existing FC controller (SW function) �������� �� Trondheim, 23rd – 24th June 2009 �
THDA OPERATION EXAMPLE (PEMFC SYSTEM) 5kW PEMFC System: Varying Air Supply Conditions Load = 2.5kW, AVL List GmbH 2007 2.0 60 cell voltage [V] air lambda 1.5 45 air lambda [-] THDA level 1.0 30 CVM (~100 channels) 0.5 15 THDA „THDA level“ 0.0 0 200 220 240 260 280 300 320 340 360 380 400 is indicating on- time scale [s] line critical 5kW PEMFC System: Dead End Operation with Delayed Anode-Purge load = 2.4kW; AVL List GmbH 2007 voltage drifts 2.0 48 H2 purge H2 purge H2 purge (somewhere) in CVM (~100 channels) 1.0 36 cell voltage [V] THDA level the stack 0.0 24 THDA -1.0 12 -2.0 0 25 50 75 100 125 150 175 200 225 250 275 300 325 time scale [s] �������� �� Trondheim, 23rd – 24th June 2009 �
EXTENDED DIAGNOSIS FUNCTIONS: CLASSIFICATION OF CRITICAL CONDITIONS On-line determination of causes for detected critical cell voltage drops/drifts: � Water issues (flooding – drying) � Low media supply issues (cathode – anode) Usage of same hardware and same instrumentation principle i.e. continuing cost efficient 2 channel approach �������� �� Trondheim, 23rd – 24th June 2009 �
ON-LINE DETECTION OF CRITICAL WATER ISSUES Extended On-line Diagnosis "AVL-THDA" E.R., AVL List GmbH, Dec.08 160 2.5 membrane issue cathode issue water issue 140 temperature air stoichiometry 2.0 C] 120 intensity of issue [-], temperature [° 100 1.5 stoichiometry 80 1.0 60 40 0.5 20 0 0.0 0 500 1000 1500 2000 2500 3000 time [s] continous air reduction too fast air reduction temperature high air flow ISSUES: >> insufficient O2 supply >> blocking through increase >> dryout remaining water >> dryout air supply issue liquid water drying issue �������� �� Trondheim, 23rd – 24th June 2009 �
DETERMINATION INTO ANODE – CATHODE ISSUES 5kW PEMFC System: Varying Air Supply Conditions Load = 2.5kW, AVL List GmbH 2007 2.0 40 air lambda cell voltage [V] air lambda [-] 1.5 30 THDA level “THDA” indicates CVM (~100 channels) 1.0 20 critical cell conditions 0.5 10 cathode relevance THDA cathode relevance 0.0 0 200 220 240 260 280 300 320 340 360 380 400 time scale [s] 5kW PEMFC System: Dead End Operation with Delayed Anode-Purge PLUS: load = 2.4kW; AVL List GmbH 2007 TH DA level & cathode relevance 2.0 48 “CATHODE H2 purge H2 purge H2 purge CVM (~100 channels) RELEVANCE” 1.0 36 cell voltage [V] FOLLOWS “THDA” 0.0 24 cathode relevance cathode relevance (cathode issue) THDA -1.0 12 OR STAYS CLOSE -2.0 0 ZERO (anode issue) 25 50 75 100 125 150 175 200 225 250 275 300 325 time scale [s] �������� �� Trondheim, 23rd – 24th June 2009 �
SYSTEM VIEW: RELIABILITY & ROBUSTNESS plausibility check wire break, out of range, device overload EMI electromagnetically interference criterion #1 signal noise e.g. thd @f1 THDA index higher frequency switching signals criterion #2 index for critical stack operation overlapping signals … blower control, motor frequency, valves ~40 parameters criterion #3 … cathode load changes impacting spectral components relevance criterion #4 … correlation to cathode issues water status stack voltage stack current 2kHz/22bit 2kHz/22bit Criteria for stack monitoring with extended diagnosis functions are � ROBUST within entire system environment due to special failure detection and compensation algorithm � Still detectable without cell voltage monitoring (main criteria are harmonic distortion based) �������� �� Trondheim, 23rd – 24th June 2009 ��
ROBUSTNESS: THDA OUTPUT EXAMPLE IN AN AUTOMOTIVE SYSTEM Automotive FC System: THDA during temporary critical air supply issues AVL List GmbH, 2007 60 THDA index 50 thd w/o compensation stack voltage stack current 40 Voltage & Current THDA level [-] 30 20 10 0 0 50 100 150 200 250 time [s] Within automotive environment usually occur electromagnetic interferences (grey plot). AVL has developed compensation algorithm for reliable THDA measurement (red plot) � critical issues are clearly indicated �������� �� Trondheim, 23rd – 24th June 2009 ��
RELIABILITY: AUTOMATED ERROR DETECTION IN THE THDA CALCULATION dynamic load test of stationary PEMFC system with THDA Monitoring THDA Messungen am Range Extender: Einfluss von Lastschwankungen THDA Messungen am Range Extender: Einfluss von Lastschwankungen cut Lastschwankungen ca. 20A/s; Pe=1.8kW & 3.0kW. 94 Zellen PEMFC Stack, Inverter ist abgeschalten. Lastschwankungen ca. 20A/s; Pe=1.8kW & 3.0kW. 94 Zellen PEMFC Stack, Inverter ist abgeschalten. Log-Datei="lastschwankungen.txt"; E.R. 25.Juni 2007 Log-Datei="lastschwankungen.txt"; E.R. 25.Juni 2007 64 80 voltage 80 70 56 70 thd1 thd 48 60 60 thd2 V_fc [V] thd@f 1 V_fc [V] I_fc [A] current & voltage I_fc [A] Strom [A], Spannung [V] Strom [A], Spannung [V] current 50 40 50 Klirrfaktor [%] thd [%] 32 40 40 24 30 30 20 16 20 THDA thd@f 2 8 10 10 0 0 0 25 50 75 100 125 150 175 200 0 25 50 75 100 125 150 175 200 Zeit [s] Zeit [s] time [s] time [s] cut harmonic distortion measurement THDA Index at two different frequencies: with automated blower controller causes error error compensation (controller uses same frequency as set for THDA) �������� �� Trondheim, 23rd – 24th June 2009 ��
SPIN-OFF EFFECT: ONLINE FAST ELECTRICAL IMPEDANCE MEASUREMENT Niquist Plot for Air Supply Variations air lambda varies from 2.1 to 1.6; frequencies=6/14/26/96/1000Hz; P=500W AVL List GmbH, July 2008 -0.03 Z1 Z2 -0.025 Z3 Z4 Continously EIS Z5 Z1typ scan every 0.5 sec. -0.02 Z2typ imaginary* [Ohm] 15Hz Z3typ Z1crit. 5Hz shows even Z2crit. -0.015 Z3crit. 25Hz time variant conditions e.g. -0.01 during water 100Hz blocking issues -0.005 1kHz 0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 real [Ohm] Conclusions from this example: only imaginary fraction changes at lower frequencies � douple layer capacity = constant & not important � membrane characteristics not changed � activation resistance to be considered � harmonic distortion effects for air supply issues below ~20Hz �������� �� Trondheim, 23rd – 24th June 2009 ��
Recommend
More recommend