Forkenbrock & GarroTt Presentation 2002 SAE Govt/Industry meeting SLIDE 1: LIGHT VEHICLE DYNAMIC ROLLOVER PROPENSITY PHASES IV, V, AND VI - RESEARCH ACTIVITIIES SLIDE 2: OVERVIEW OF NHTSA ROLLOVER RESEARCH PHASES PHASE 1-A - spring 1997 - exploratory in nature - emphasized maneuver selection and procedure development PHASE 1-B - fall 1997 - evaluation of test driver variability - introduction of the programmable steering machine PHASE 2 - spring 1998 - evaluation of 12 vehicles using maneuvers researched in phase 1 PHASE 3-A - spring 200 - introduction of 'roll rate feedback' PHASE 3-B - summer 2000 - pulse brake automation PHASE 4 - spring 2001 - response to TREAD act - consideration of many maneuvers PHASE 5 - spring 2002 - research factors that may affect dynamic rollover propensity tests - rollover and handling rating development PHASE 6 - evaluation of 26 vehicles using phase 4 recommendations SLIDE 3: PHASE 4 BACKGROUND Tread Act / Congressional requirements - develop dynamic rollover propensity tests to facilitate a consumer infromation program - consumer information methodology released by november 2002 - national academy of sciences report SLIDES 4-6: ADDITIONAL BACKGROUND In their assessment of NHTSA's exising rollover resistance rating system (January 2002), the national academy of sciences recently recommended: "NHTSA should vogiorously pursue the development of dynamic testing to supplement the information provided by SSF" - NHTSA is presently providing rollover resistance rating - based on vehicle measurements and real world crash data - vehicle measurement is static stability factor - 5 star ratings are similar to NCAP crash ratings SLIDE 7: Static stability/ rpbability of rollover per single vehicle crash (implementation of star system, as explained via line chart) SLIDE 8: MANEUVER RECOMMENDATIONS - alliance of automobile manufacturers - consumers union - ford motor company - heitz automotive, inc. - ISO 3888 pt 2 consortium * VW, BMW, Daimler Chrysler * Porsche, Mitsubishi - MTS systems corporation - nissan motors - toyota motor company - UMTRI SLIDE 9: PHASE IV TEST CONDITIONS SLIDE 10: TEST VEHICLES 2001 Chevrolet Blazer 4x2 - one star static rollover rating - high sales volume 2001 Ford Escape 4x4 - three star static rollover rating - smaller, car-like SUV 1999 Mercedes ML320 4x4 - "less aggressive" stability control intervention - two star static rollover rating - first SUV with available stability control (ESP) 2001 Toyota 4Runner 4x4 - "aggressive" stability control intervention - two star static rollover rating - relatively high sales volume SLIDE 11: VEHICLE CONFIGURATIONS - instrumented - fully fueled - front and rear mounted aluminum outriggers - performed with and without stability control - multiple configurations * nominal vehicle * reduced rollover resistance SLIDE 12: REDUCED ROLLOVER RESISTANCE - roof-mounted ballast - designed to reduce SSF by 0.05 - increased roll inertia from Nominal condition * Escape = 8.0% * Blazer = 11.5% - longitudinal C.G. preserved - maneuver sensitivity check SLIDE 13: REDUCED ROLLOVER RESISTANCE (measurements taken without instrumentation) 4RUNNER - 180 lbs ballast - C.G. raised 1.3" - SSF[nominal] = 1.11 (2 stars) - SSF [RRR] = 1.06 (2 stars) ESCAPE - 120 lbs ballast - C.G. raised 1.0" - SSF[nominal] = 1.26 (4 stars) - SSF [RRR] = 1.21 (3 stars) BLAZER - 180 lbs ballast - C.G. raised 1.3" - SSF[nominal] = 1.04 (2 stars) - SSF [RRR] = 0.99 (1 star) ML320 - 180 lbs ballast - C.G. raised 1.2" - SSF[nominal] = 1.14 (3 stars) - SSF [RRR] = 1.09 (2 stars) note: nominal SSF differ from those measured without outriggers SLIDE 14: TEST VEHICLE SSF SUMMARY [bar charts comparing blazer, 4runner, ml320, escape] SLIDE 15: TIRES - OEM specification (as installed on vehicle when delivered) * make * model * DOT code * inflation pressure - frequent tire changes - innertubes used during some maneuvers to prevent debeading - manuver speed iterations selected to minimize tire wear within a given test series SLIDE 16: TEST SURFACE - all tests performed on TRC's VDA (a dry, high-mu asphalt surface) - tests performed 04/01 to 11/01, 02/02 - stable friction coefficients * peak mu: 0.94 to 0.98 * slide mu: 0.81 to 0.88 SLIDE 17: PHASE IV MANEUVER REVIEW SLIDE 18: CHARACTERIZATION MANEUVERS - used to define NHTSA's dynamic rollover propensity maneuvers * constant speed, slowly increasing steer - used to characterize transient response * pulse steer * sinusoidal sweep * J-turn response time tests SLIDE 19: DYNAMIC ROLLOVER PROPENSITY MANEUVERS AUTOMATED STEERING - NHTSA J-turn - fixed timing fishhook - roll rate feedback fishhook - Nissan fishhok - Open loop pseudo-double lane change DRIVER-BASED STEERING - ISO 3888 part 2 - CU short course DRIVER-BASED STEERING; COMPUTER CORRECTED - ford PCL LC SLIDE : SLIDE 20: NHTSA J-TURN AND FISHHOOKS - Steering magnitude based on vehicle response 1. determine the handwheel angle at0.3g from slowly incresasing steer results 2. multiply rate based on successful Phase II testing - steering rate based on successful phase II testing 1. J-turn = 1000 PER SET 2. fishhok = 730 deg/sec SLIDE 21: NHTSA J-TURN VEHICLE HANDWHEEL INPUT (degrees) Blazer 401 4Runner 354 ML320 310 Escape 287 SLIDE 22: NHTSA FIXED TIMING FISHHOOK (SYMMETRIC) VEHICLE HANDWHEEL INPUT (degrees) Blazer 326 4Runner 287 ML320 252 Escape 233 SLIDE 23: NHTSA ROLL RATE FEEDBACK FISHHOOK (symmetric) VEHICLE HANDWHEEL INPUT (degrees) Blazer 326 4Runner 287 ML320 252 Escape 233 SLIDE 24: NISSAN FISHHOK - adjusts timing to maximize roll motion - 270 degree initial steer - vehicle-dependent reversal magnitude (for fishhooks) * blazer = 570 degrees * escape = 505 degrees - all rates = 1080 deg/sec - response-dependent dwell times * iterative determination SLIDE 25: CLOSED-LOOP, PATH-FOLLOWING LANE CHANGES [2 overhead diagrams] - Consumers Union Short Course - ISO 3888, Part 2 course SLIDE 26: FORD PCL LC - comprised of a suite of closed-loop paths (double lane changes) - data is processed to remove driver effects and factilitate comparison at a constant severity * all vehicles taken to follow the same path * all vehicles subject to the same lateral acceleration demands - test output is an overall dynamic weight transfer metric SLIDE 27: FORD PCL LC [3 comparative diagrams] SLIDE 28: COMMENTS BASED ON THE PHASE IV ROLLOVER RESISTANCE MANEUVERS SLIDE 29: NHTSA J-TURN - lowest speed of two-wheel lift is metric - uses programmable steering controller - simple step-steer (one cycle) - handwheel magnitude dependent on vehicle rsponse SLIDE 30: J-TURN WITH PULSE BRAKING - lowest speed of two-wheel lift is metric - uses programmable braking and steering controller - addition of braking controller makes maneuver substantially harder to perform - timing of brake pulse dependent on vehicle response (roll rate feedback) - results significantly influenced by whether vehicle has working ABS SLIDE 31: FIXED TIMING FISHHOOK - lowest speed of two-wheel lift is metric - dwell time independent of vehicle response - handwheel magnitudes dependent on vehicle response - handwheel inputs within ranges established during - handwheel inputs within ranges established during ISO and CU double lane change testing - timing may be better for one vehicle than another SLIDE 32: ROLL RATE FEEDBACK FISHHOOK - lowest speed of two-wheel lift is metric - handwheel magnitudes dependent on vehicle response - handwheel inputs within ranges established during ISO and CU double lane change testing - dwell time also dependent on vehicle response - timing should no longer favor one vehicle over another SLIDE 33: NISSAN FISHOOK - lowest speed of two-wheel lift is metric - iterative procedure requires additional testing time - large number of tests required many tire changes (to reduce tire wear concerns) - reversals are harch; increases steering machine wear SLIDE 34: FORD PATH CORRECTED LIMIT LANE CHANGE (PCL LC) SLIDE 35: FORD PCL LC - metric dynamic weight transfer at 0.7g based on one of four standard paths (DWTM) - method removes driver dependence by normalizing data - extra tire testing required (tire measurements) - concerns about 0.40 second window used for metric calculation (mitigates dynamic weight transfer observed) - metric now measured during tests performed with a driving robot SLIDE 36: ISO 3888 PART 2 DOUBLE LANE CHANGE - suggested rating metric is maximum achievable "clean" run speed * "clean" run: no cones struck/bypassed - test driver generated steering inputs - not as repeatable as programmable steering controller inputs - tests are straightforward to perform - course adapts to vehicle width SLIDE 37: CONSUMERS UNION SHORT COURSE DOUBLE LANE CHANGE - suggested rating metric is maximum achievable "clean" run speed * "clean" run: no cones struck/bypassed - test driver generated steering inputs - not as repeatable as programmable steering controller inputs - tests are straightforward to perform - course does not adapt to vehicle size SLIDE 38: OPEN-LOOP PSEUDO-DOUBLE LANE CHANGE - uses programmable steering controller - having three major steering moves slightly degrades repeatability - straight-forward to perform - uses programmable steering controller - additional development required SLIDE 39: REPORTING OF PHASE IV FINDINGS Draft of Phase IV NHTSA Technical Report has been written - reviews in progress - anticipated release late Spring '02 SLIDE 40: PHASE V RESEARCH SLIDE 41: PHASE V OVERVIEW - investigate potential use of a centrifuge - improved test equipment * alternative outrigger development * quantification of two-wheel lift - resolution of existing matters * cold and hot weather testing * surface effects testing - finalize methodology for Phase VI * loading SLIDE 42: CENTRIFUGE - metric could be lateral acceleration at wheel lift or weight transfer - quasi-static test - may be demonstrated by NHTSA using a NASA facility SLIDE 43: OUTRIGGER DEVELOPMENT - reduce effects of outrigger installation without compromising driver safety - use wheel load transducers to evaluate dynamic load transfer and cornering forces - compare three designs * existing VRTC Design - aluminum - 78 lbs per outrigger * new VRTC design - titanium - 68 lbs per outrigger * Carr engineering - carbon fiber - 58 lbs per outrigger - testing complete SLIDE 44: CARBON FIBER - manufactured by Carr Engineering - light weight (58 lbs) - strong - expensive ($25K/set) SLIDE 45: TITANIUM - designed at VRTC using finite element analysis - light weight (68lbs) - less roll inertia than aluminum or carbon fiber - strong - 1/3 cost of carbon fiber - 6AI-4V of a common Ti alloy - low-mu hemispherical skid pads replace heavier casters SLIDE 46: QUANTIFICATION OF TWO-WHEEL LIFT - objective methodology required - laser-based height sensors on each wheel * elimiates video data analysis subjectivity SLIDE 47: COLD AND HOT WEATHER TESTING - will reserach the effects of temperature extremes on dynamic rollover propensity - all testing to be performed at TRC - cold weather tests performed during January '02 - hot weather tests to be performed Summer '02 SLIDE 48: SURFACE EFFECTS TESTING - determine effects of different test surfaces on dynamic rollover propensity - testing performed in Arizona * DaimlerChrysler Arizona proving grounds (APG) * GM Desert proving grounds * performed with the Blazer and 4Runner - testing complete - results from Arizona will be compared with those produced at TRC SLIDE 49: PHASE VI SLIDE 50: PHASE VI OVERVIEW - maneuvers based on phase iv findings - three load conditions - titanium outriggers - 26 vehicles - will include a wide range of make/models for whch state rollover rate data is available