Engine Optimization Methodologies: Tools and Strategies for Diesel Engine Design George Delagrammatikas Dennis Assanis, Zoran Filipi, Panos Papalambros, Nestor Michelena The University of Michigan May 24, 2000
BACKGROUND: VEHICLE AND ENGINE FUNDAMENTALS ENGINE TUNING INVERSE DESIGN NOVEL TECHNOLOGIES CVT, INJECTION TIMINGS MAP SHAPING AND MATCHING ENGINE FLEXIBILITY: COMPRESSION RATIO, CVT,HYBRID TARGET CASCADING
Motivation Federal Regulations Fuel economy (CAFE) Emissions (NOx, smog, and other pollutants) Public Awareness Green movement Global warming scare Decrease Dependence on Foreign Oil Avoid another oil crisis
Objectives Develop an engine optimization framework 21 st century conventional and hybrid heavy truck Implement techniques to conventional vehicle Define a problem analytically Apply suitable driving cycle(s) Investigate location of use on engine map Extend lessons learned to hybrid vehicle How different are demands on engine and transmission? Need for a systems approach
Euro III Steady State Test Procedure
Power Demands on Engine WHEEL Torque, Speed at Wheel Torque, Speed at Engine ENGINE
HEV System Simulation Framework Matlab-SIMULINK environment FUEL (g/s) ADVISOR TORQUE RPM Parallel HEV ENGINE FUEL DELIVERY LOOK-UP TABLE
Baseline Vehicle Parameters Cummins M11-330 (246 kw) Diesel Engine Wheel/axle assembly for heavy truck Kenworth T400 Vehicle Standard heavy vehicle accessory loads Standard catalyst for CI engine Eaton Fuller RTLO-12610B 10-Speed Transmission Generic 10-spd constant efficiency gearbox Heavy Vehicle Powertrain Control Eaton Fuller RTLO-12610B 10-Speed Transmission Generic 10-spd constant efficiency gearbox
Baseline Cummins Engine Map
Driving Cycles Investigated US06 REP05 FUDS FHDS
Engine Use Points for Various Cycles US06 REP05 FUDS FHDS
Power Frequency for Each Cycle US06 REP05 FUDS FHDS
Ideal BSFC Line Generation Output Torque Engine Speed
Benefits of Flexible Engine Designs Actual Transmission Case Determine cumulative fuel throughput for each cycle investigated Interpolate BSFC from engine map for every torque/speed combination for that given cycle Integrate all BSFC s from above step Find the total time that engine is producing power Mean effective BSFC = total BSFC/engine on time
Ideal BSFC vs.. Power Level BSFC (g/kw-hr) Power
Benefits of Flexible Engine Designs Ideal Transmission Case Find ideal BSFC transmission line on engine map used for a given cycle Interpolate BSFC for every visited power level on the BSFC vs. Power graph Sum of all BSFC s is cumulative fuel throughput Mean effective BSFC = numerical average of total fuel throughput during time steps when engine is active
Potential Benefits of Ideal CVT Design 140 Increase in Mean BSFC Per Cycle 120 100 80 60 40 20 0 4*BSFCmin 2*BSFCmin US06 REP05 FUDS FHDS
Optimum Injection Timing Method Using an optimization framework Vary injection timing for every torque/speed combination (over 200 map points, ~100 executions per point) Computationally prohibitive Parallel computer framework Run as many maps as you want at different injection timings Coalesce data with a Matlab-based routine One map = ~5 minutes
Injection Timing Maps Output Torque Engine Speed
Optimum Injection Timing Map Timing Fueling Rate RPM
Variable Compression Ratio Engine Hypothetical investigation of novel engine design Find the ideal fuel consumption benefit Apply ideal transmission techniques from previous slides Determine how BSFC can be optimized at various power levels First maximize power density to find engine s power upper bound Allow engine controller to change parameters that are not normally variable
Problem Formulation For each power level : 50, 100, 200, 300, 400, 500 kw Minimize BSFC, subject to: overall phi < 0.6 20% < percent premixed burn < 40% peak cylinder pressure < 150 bar Variables: Inlet manifold pressure Compression ratio Injection timing Fuel Engine speed
Same Engine - Different Maps
VCRE - Combined Map TORQUE RPM
240 Ideal BSFC Line vs. Power Level 230 BSFC (g/kw-hr) 220 210 200 190 vcre_bsfc base_bsfc 180 170 50 100 150 200 250 300 350 400 450 500 Power (kw)
Hybrid Powertrain Investigations
Demands on Engine - CVT vs. 5-Speed
Demands on Motor - CVT vs. 5-Speed
Battery SOC - CVT vs. 5-Speed
Zero Delta SOC - CVT vs. 5-Speed
Additional Hybrid Clustering Scenarios 40 60 20 Power-Assist Battery Recharge
Future Directions How realistic is the variable compression ratio engine for the driving cycles and vehicles we are considering? How can we better quantify the benefits of increased flexibility in transmission parameters? What are the effects of injection timing and variable valve timing on engine map characteristics? Is it better to cluster points around optimum power levels on an engine map or disregard that an engine can be described by a single map?
Integration with Target Cascading Parameterize engine torque curve Maximize engine turndown ratio while meeting mobility constraints Match maximum torque curve with a real engine defined by high fidelity model Use engine visitation points in conjunction with control strategies to meet BSFC and emissions targets Send results back to top level for verification and subsequent iteration
Parameterized Max Torque Curve - I
Parameterized Max Torque Curve - II
Sample Torque Curves Min cluster area Min rated torque Min rated power
Engine Matching Subproblem
Conclusions High fidelity engine model can be used at different levels in the design process Methods have been illustrated on a variety of different engines Continue feed-backward work with ADVISOR and extend methodology to VESIM