Transcript Document
Double Layer Capacitors: Automotive Applications and Modeling David A. New John G. Kassakian Joel E. Shindall David J. Perreault Massachusetts Institute of Technology, Laboratory for Electromagnetic and Electronic Systems Validation of DLC Model DLC Use in Automobiles Abstract •Charge the DLC with a constant current •Discharge DLC through a resistive load •Charge/discharge, charge/hold/discharge capability •Supplement to Battery ─ Ease of starting ─ Reduction of weight and volume of the current “oversized” battery ─ Increase lifetime of automotive battery Prevention of frequent battery SOC dips that will decrease the lifetime of the battery •Supply/accept high peak power demands efficiently Computer: LabView ─ Regenerative braking Captured energy generated by the braking system ─ Brake-by-wire ─ Steer-by-wire ─ Integrated starter/generator 100A No Delay 4 74.67 74.70 Cycles Lab (%) PSpice (%) Modeling of DLCs HP 6011A Current Source TDS 754D Voltage Meter •Various experimental/simulation profiles conducted –100A No Delay, 100A Delay –10A No Delay, 10A Delay AM 503 Current Amplifier P6139A V-Probe ADA 400A Differential PreAmplifier 100A With Delay 3 74.72 72.67 10A No Delay 3 93.52 90.82 Discharge Resistance A6303 I-Probe NESSCAP 2500F P6139A V-Probe 10A With Delay 1 77.41 76.59 100A Charging No Transistion Delay 0.5 0.4 Voltage [V] This research project has concentrated on the modeling of double layer capacitors (DLCs) and the validation of those models. Several experiments have been conducted to subject the device under test to a variety of charging/discharging profiles in an effort to extract model parameters from the device’s performance and to simulate the various conditions such a device might encounter in an automotive type application. High and low current charging profiles were performed for both charge/discharge and charge/hold/discharge type experiments. The derived DLC model was then used in PSpice® simulations to determine how accurately the model could predict the performance of the device. This poster gives a look into the modeling of DLCs, the experimental setup and testing procedure used to test the devices, the simulations for the comparison, and presents the results of the comparison. As a result, this poster documents the conclusion that this simple model very adequately predicts the performance of the device under these various performance profiles. Finally, the lowtemperature trends of these devices are illustrated by experimental results. 0.3 0.2 0.1 0 •“Short-term” Model: Fast, Medium, and Slow Branches •Device quickly charged to ~0.5V, voltage decay due to charge diffusion •Model parameters extracted from observed behavior •Very high capacitance per unit volume and weight •Low voltage devices (~3V) •Superior charging/ discharging efficiency •Cycleability of device exceeds the lifetime of most applications Conventional Capacitor Note: Does not investigate non-linear capacitance Supercapacitors (carbon based) Rf 300 350 400 450 500 0 50 100 150 200 250 300 350 400 450 500 0 50 100 150 200 250 Time [s] 300 350 400 450 500 Current [A] Power [W] 40 30 20 10 0 0.3 0.25 0.2 0.15 Rs Cs -10 0.1 0.05 0 500 1000 1500 2000 2500 Low-Temperature Trends 3000 Time [s] Case •First experiment determines 6 of the 7 model parameters •Leakage resistance is not a dominant term when dealing with “short-term” time periods (i.e. a few hours) •Leakage resistance important for long-term behavior of device (i.e. 30 day “airport test”) •Second experiment to determine leakage resistance macropore (>500Å) •Investigate the effects of low-temperature on DLC model parameters •Constructed low-temperature setup that utilizes a liquid nitrogen coolant •Device held at low-temperature for ~3hrs •Experiment performed at ~-30ºC Plexiglas NessCap 2500F Electrolyte DLC micropore (20-8Å) Average Ionic Center sub-micropore (<8Å) mesopore (500-20Å) Activated Carbon Solid/Liquid Interface Single Time Constant Capacitor Model •“Base-line” room-temperature model determined •Device charged to and held at 1V for extended period of time •Minimize voltage decay due to ionic diffusion •Periodic measurements of voltage decay V ≅ 0.77V @ t ≅ 30 days Actual DLC Behavior Heat sink 0.44m Voltage Decay 2700F 0.57 220F Styrofoam 2.9 490F Potentiostatic Charge •Model fitted to low-temperature data v(t) Voltage Decay Leakage Resistance 1V 1 Cm 250 50 0.35 0 1.1 Cf 200 60 0.4 Battery Electrode Bulk Rs 150 50 0.45 Electrodes ion Rm Cm 100 0 Separator Helmholtz Layer (few Angstroms) •DLC are complex devices that are modeled by multi-time constant networks Rm NESSCAP 2500F, 100A Pulse 0.5 Rlk Cf C A/d •Small charge separation •High surface area (1500-2400 m2/g) Rf 50 100 Voltage [V] What is a Double Layer Capacitor? 0 Cs 0.45m 2700F 1.7 140F 20 300F Voltage [V] 0.9 0.8 0.7 3k 0.6 0.5 0.4 0 20 40 60 80 Time [days] 100 120 ~1 Week t Model Rf Rm Rs Cf Cm Cs Room-temperature model 0.44m 0.57m 2.9 2700F 220F 490F Low-temperature model 0.45m 1.7m 20 2700F 140F 300F 140 Advantages of Double Layer Capacitors •Carbon/carbon DLCs can be fully discharged without reducing the lifetime of the device •Cycleability of DLC exceeds lifetime of automobile 0.68m •Superior charging/discharging efficiency •Higher specific power rating than Pb-acid battery •Higher specific energy rating than conventional capacitor technologies 2.9 3k 2600F •Minimal (if any) maintenance 0.8 250F 560F Model: NessCap 2500F, 2.3V •Derived model fits well to experimental data Acknowledgements •MIT/Industry Consortium on Advanced Automotive Electrical/Electronic Components and Systems for its funding of this project. •Ness Capacitor Co., Ltd (NCC), Maxwell Technologies, and Panasonic for supplying our devices. •Thomas Keim, John Miller (J-N-J Miller Design), and Steven Leeb for their technical support and guidance. •Ivan Celanovic, Woo Sok Chang, Tushar Parlikar, Joshua Phinney, Juan Rivas, John Rodriguez, Wayne Ryan, Sai Chun Tang, and David Wentzloff for their help in the lab and all the late night coffee. MIT/Industry Consortium on Advanced Automotive Electrical/Electronic Components and Systems