Richard Boyd Acting Director, Office of Defects Investigation 10/12/2010 1

Transcription

1 Richard Boyd Acting Director, Office of Defects Investigation 10/12/2010 1

2 ODI Overview Mission Safety Defect Conditions Early Warning Reporting Overview Data Specifics EWR data uses in screening Defects Assessment Overview Vehicle Owners Questionnaires VOQ uses in screening Defect Investigations Overview Data Usage 10/12/2010 2

3 ODI Mission ODI Process / Information Flow Data Sources Early Warning Reporting Vehicle Owners Questionnaires Safety Defect Conditions Unintended Acceleration (UA) Conditions Electronics concerns 10/12/2010 3

4 Identify defects that relate to motor vehicle safety o To order a safety recall, NHTSA must be able to demonstrate that: A defect exists (as shown by a significant number of real world failures), and The defect poses an unreasonable risk to safety Assure that defects are remedied effectively and promptly Assure that information relating to investigations and recalls is readily available to a variety of customers 10/12/2010 4

5 Sources Consumer Complaints Screening DAD Investigation Proposals Investigative VID Uninfluenced Recalls EWR VCD Influenced Recalls RMD TSBs Tips EWD MHDVD Closing Reports 10/12/2010 5

6 Early Warning Data Prompted by the TREAD Act Collected from the manufacturers quarterly Consumer Complaints to NHTSA Furnished by individual consumers Received continuously Each one is reviewed as it is received 10/12/2010 6

7 Electronic control systems permeate the vehicle fleet ODI routinely investigates related defects Engine Stall Electric Power Steering Failure Inappropriate Electronic Stability Control Activation Since 2000: Over 30 related investigations conducted Causes include: Sensor failures PCM manufacturing flaws Software issues Engine stalling is the most frequent condition examined 10/12/2010 7

8 Traditionally reported High initiation speed / long duration Throttle or accelerator pedal stuck at high power setting Low speed initiation / high power Traditional SUA Failure to return to idle (lower power) Throttle or accelerator pedal stuck at low power setting More recently reported Modest, intermittent increases of power Cruise control Resume Surging Transmission / torque converter behavior Engine power adjustment (accessory loads) 10/12/2010 8

9 Christina Morgan Early Warning Division Chief Office of Defects Investigation 10/12/2010 9

10 Origin Requirements Overview of EWR Data EWR Screening Process Identification of Electronic Vehicle Controls and UA in EWR 10/12/

11 Ford/Firestone crashes led Congress to conclude NHTSA needed information that may provide early warning of defect Tread Act (2000): required NHTSA to issue rule establishing early warning requirements Rule became effective /12/

12 Requirements vary depending on size and type Large Manufacturers Production Volume (cumulative) Aggregate data (warranty claims, consumer complaints, property damage claims, field reports) Death and Injury claims and notices Non Dealer field reports Small Manufacturers death claims and notices 10/12/

13 Large Manufacturers Light Vehicles (5,000 + annually) Medium/Heavy Vehicles (5,000) except Emergency Vehicles (500 +) Buses (100 +) Motorcycles (5,000 +) Child Restraint System Tires (15,000 +) Small Manufacturers Other Original or Replacement Equipment Manufacturers Low Volume Manufacturers 10/12/

14 10/12/

15 Quarterly Report Production, Aggregate, Death and Injury Data due 60 days Non Dealer Field Reports due 75 days Scope Vehicle Manufacturers 10 Model Years Child Restraint and Tire Manufacturers 5 Production Years Make (Tire line), Model (Tire Size), and Model year (production) Component Codes 10/12/

16 10/12/

17 10/12/

18 Make, Model, Model Year Vehicle Type and Platform Annual Production (Cumulative) 10/12/

19 10/12/

20 10/12/

21 10/12/

22 10/12/

23 Make, Model, Model Year, VIN Incident Date Number of Deaths and Number of Injuries State or Foreign Country Component A through Component E Claims often made well after incident so considerable time lag 10/12/

24 10/12/

25 10/12/

26 Average about 20,000 documents a quarter Most are submitted by Light Vehicle Manufacturers 2010 Qtr 1 17,319 total 14,263 Light vehicle 2,103 Medium Heavy Vehicles 542 Child Restraint Manufacturers 403 Motorcycles 2 Trailers Make, Model, Model year and Component Most can be searched using keyword 10/12/

27 Identify potential safety related defects Request more information from manufacturers on some death and injury incidents Variety of different analytical methods Aggregate and Production Data Mahalanobis Distance, Probability, Time Series (Crow AMSAA) Identify high counts or high rates, increasing trend, and outliers Review important products identified using these methods Bayesian Filter to identify important Non Dealer Field Reports Data mining tool Each quarter retrain the filter by reviewing both important and non important Field Reports Refer any potential safety related defects to Defects Assessment Division Evaluate the EWR data and methodologies used to identify safety related defects 10/12/

28 Aggregate data Component Code Death and Injury data Component Code Non Dealer Field Reports Component Code Keyword search of text 10/12/

29 Steering Electrical Speed Control Suspension Exterior Lighting Tire Related Service Brake Visibility Wheels Parking Brake Air Bags Seats Engine & Engine Cooling Seat Belts Fire Related Fuel system Structure Rollover Power Train Latch 10/12/

30 Electronic Vehicle Control Found under a number of component codes Steering, Service Brakes, Electrical, Air Bag, Etc. Planning a Rulemaking to add components for fuel/propulsion systems and crash avoidance technologies. Unintended Acceleration Found under a number of component codes Speed Control, Powertrain, Engine, Electrical, Service Brakes Speed Control means the systems and components of a motor vehicle that control vehicle speed either by command of the operator or by automatic control, including, but not limited to the accelerator pedal, linkages, cables, springs, speed control devices (such as cruise control) and speed limiting devices. 10/12/

31 10/12/

32 10/12/

33 10/12/

34 Limitations Data is received after the end of Quarter Production, Aggregate, Death and Injury data 60 days Non Dealer Field Reports 75 days Sometime VOQ data actually shows possible trends sooner Data alone lacks details to identify safety related defects; can simply identify possible trends Benefits Provides a large volume of data quarterly Identifies spikes, outliers, and trends Non Dealer Field Reports give technical description of cause and remedy Overall EWR data has prompted some defects investigations by itself but it is most useful in supporting defects investigations by supplementing our traditional ODI data 10/12/

35 Gregory E. Magno Defects Assessment Division Chief Office of Defects Investigation 10/12/

36 Introduction Defects Assessment What is a VOQ Where do VOQs come from VOQ Traffic Publicity Effects Identifying UA in VOQs Component Codes Narratives Screening Process General Electronics UA Issues 10/12/

37 Mission: Screen field data for indications of safety defect trends Verify underlying field data Present the issue to the investigative staff as a potential safety defect investigation Review all incoming consumer complaints VOQ 10/12/

38 Who When How What 10/12/

39 Consumers file their own VOQ with NHTSA Used to report safety defect concerns Filed on a continual basis Annual traffic: ~35k / year Redacted versions available on NHTSA web site Complaints contain: Complainant & vehicle identity Complainants direct narrative Primarily used for defect screening and investigations Complaints are not altered or recoded after they are submitted by the complainant UA incidents identified via the narrative 10/12/

40 IVOQ: Internet VOQ 56% of total volume VOQ: OMB Form % of total volume EVOQ: Electronic VOQ Auto Safety Hotline 37% of total volume Consumer Letters Mapped into VOQ database 5% of total volume Consumer Complaints Rcv d since Jan /12/

41 Oracle database ~482k VOQs since 2000 Relational & image data ARTEMIS Search VOQs Manage correspondence Store investigative & recall files Receive / store EWR data 10/12/

42 CY 06 Present 10/12/

43 Professional Judgment What is the safety hazard of the concern? Frequency / Trend Are the complaints closely spaced over time? Can a trend be discerned? Failure Rate (rate = failures / production) Does the subject failure rate stand out from peers? Severity Were any fires, crashes, injuries, or deaths reported? ODI History Has this condition been recalled or investigated in other vehicles? Special Considerations Special vehicles, new technology, unattended vehicle roll away / fire 10/12/

44 Where appropriate, complainants are called / visited: Verify complaint content Additional details re: Vehicle history and incident Repair history Prior symptoms Driving conditions Equipment Obtain evidence Parts Photographs Documents 10/12/

45 Repair Invoices Police Accident Reports Insurance Cause / Origin Reports Correspondence Other documents as appropriate 10/12/

46 YTD 10/12/

47 Relevant complaints are identified via manual review during screening or an investigation No condition codes to indicate UA UA incidents may be filed under a number of component codes: Speed Control, Powertrain, Engine, Electrical, Brakes Other conditions such as stalling, hesitation, and brake concerns may overlap these codes 10/12/

48 Component: Powertrain Condition: UA Component: Brakes Conditions: Hesitation Harsh Shifting Component: Component: Engine Speed Control Conditions: Engine Stall Hesitation Condition: ABS concerns 10/12/

49 Broad review of VOQs received Jan 2000 Mar 2010 initiated earlier this year Identifying possible UA VOQs Boolean keyword search of narrative Manual review of results Is the complaint UA related or not? Sort the complaints Factual conditions Location, circumstances Inferences Where facts not all reported in narrative Analysis still in progress 10/12/

50 10/12/

51 10/12/

52 VOQs filed with NHTSA from 1 Jan 00 5 Mar 10 10/12/

53 A number of consistent VOQs are identified Follow up narrows down the circumstances Common repair op (less likely to find root cause than in other areas) Vehicle Control Electronics defects generally show consistent incident circumstances Interrelationships between three areas complicate matters: Control hardware Software Sensors A sensor failure may be mechanical and then a software safeguard may generate the condition reported in the VOQs Investigations generally need a higher number of complaints and complaint rate before opening Crash rate is lower than in non electronics vehicle control investigations 10/12/

54 Mid 2003: noted increases in Camry UA VOQs Multiple circumstances No prolonged high speed incidents New throttle technology Consumer follow ups initiated Exact root cause not known ~ 50% of incidents involved crashes None were fatal Issue Evaluation proposed early December consumer complaints Pre dated publicity or advice from external parties Screening activity and a defect petition led to the opening of an investigation 10/12/

55 Late 2006: noted prolonged high speed incidents in 2007 ES350 Consistent circumstances Consumer follow ups initiated Exact root cause not known ~ 50% of incidents involved crashes Issue Evaluation proposed late Feb consumer complaints Further follow up & field work identified the accelerator pedal entrapment by all weather floor mats 10/12/

56 Electronic Vehicle Controls Data give us info on trends Help us focus on the condition And also on the causal component New EWR improvements aimed at better focus on advanced electronics controls Unintended Acceleration VOQs helpful in spotting trends Not very helpful in identifying a vehicle based cause or component 10/12/

57 Jeffrey L. Quandt Vehicle Control Division Chief Office of Defects Investigation 10/12/

58 Purpose of a defect investigation Stages of a defect investigation Information request letters Complaint analysis Complaint review Important factors in UA investigations Throttle questions Brake questions Examples of UA complaint categories Examples of investigations of electronic defects Closing comments 10/12/

59 Collect data to assess the following for an alleged defect: Frequency Warranty claims Complaints Surveys Other (e.g., crash databases, insurance data) Statistical models (e.g., Weibull, Log Normal) Safety consequences Crashes, injuries, deaths or potential for these events to occur Complaint interviews Testing Scope Design/process factors Problem experience 10/12/

60 Opening the investigation: opening resume summarizes basis Preliminary evaluation (PE): obtain additional information from manufacturer via information request (IR) Engineering Analysis (EA): if PE has not resolved issue either way, EA involves deeper examination, usually a new IR, and possibly testing or surveys Closing: final resume explains basis for closing the investigation (either recall has occurred or agency concluded it could not establish safety defect) 10/12/

61 Defines: Subject vehicles Subject component Alleged defect Requests: Production data Complaints, field reports, claims, lawsuits Warranty claims Mfr field communications (e.g., technical service bulletins) Design & manufacturing process changes Mfr studies, surveys, investigations Sample parts (exemplar, warranty returns) Part sales Manufacturer s assessment 10/12/

62 Analysis of all UA complaints Useful as a first step to identify trends Includes many different conditions Publicity effects hamper rate analysis/comparisons Complaint categorization/binning Must assign complaints to appropriate categories or bins Assess scope, frequency & severity for each bin independently Examples of bins Pedal entrapment (09V 388, 10V 023) Sticky pedal (10V 017) Dual pedal application Acceleration during braking More detailed information collected during complaint review to aid in binning Owner interview a fundamental step in complaint review 10/12/

63 Interview Driver Witnesses (repair technicians, police officers, passengers, ) Collect evidence Repair records Parts Photographs/videos Police reports Insurance reports Field investigation Throttle controls evaluation Brake system evaluation Snapshot of electronic data Event data recorder (only since March 2010) 10/12/

64 UA involves throttle opening originating either from unintended application of accelerator by driver or unintended throttle opening due to vehicle malfunction Vehicle malfunctions that may cause UA can be mechanical, electronic, or obstruction of pedal Malfunctions related to mechanical issues or pedal obstruction generally result in stuck throttle not throttle opening (i.e., sudden acceleration) Vehicle characteristics (e.g., pedal configuration) may influence likelihood of pedal misapplication or dual application When UA occurs, information on brake application is important; normally functioning brake should stop UA incident; pressing brake alone cannot cause acceleration 10/12/

65 Did the engine power increase from idle or did it fail to decrease after the accelerator pedal was released? Engine power level (high/low, fixed/changing)? Duration (short surge or sustained increase)? Initiation speed? Environmental conditions (ambient temp, moisture)? Engine conditions (cold or warm)? Cruise control status? What equipment was being operated? Post incident inspection/repairs to throttle system? Throttle system service history? 10/12/

66 What was the vehicle response to brake application? Did the engine power increase begin when the brake was applied? Did the engine power change with braking force? Did the engine power change after brake release (in P or N )? Was the brake system inspected after the incident? Were any problems found? Did the brake components display signs of overheating? Did the driver apply the brake pedal more than once during the event? Were there any brake system service issues before or after the incident? 10/12/

67 VOQ No. Vehicle Description Camry 3.5L V6 We have a six cylinder, 2007 Toyota Camry. We were passing a semi truck on an interstate highway when the car suddenly demonstrated unintended acceleration. Applying the brake and shifting into neutral did not stop the acceleration. Our only recourse was to get off the interstate and we were able to get the car to stop by turning off the key. If there had been a vehicle in front of us there would have been a serious accident. 10/12/

68 VOQ No. Vehicle Description Avalon 3.5L V6 I own a 2007 Avalon. In the past 6 months I have experienced 5 events where the car accelerated on its own prior to the final events leading up to this report. Several times I experienced the car accelerating without my foot on the gas pedal as I drove through town. The car would go back to its correct rpms after driving a few miles or after the car was stopped and put into Park. After experiencing the sudden acceleration the third time I took my vehicle to be checked at my local auto shop. They could not find anything wrong with the vehicle. After 2 more incidents I brought my car to a Toyota dealership to be checked. After keeping my car for 2 days they found no unintended acceleration problems and confirmed that the factory mats were installed properly. On 12/28/09 I was driving to work on a major highway. The car began to accelerate without my foot on the gas pedal. As I pushed on the brake, the car continued to accelerate. At that time I was not able to stop my vehicle by pressing hard on the brake. The only way I was able to slow the car down was to put the car into Neutral. I took the next exit which was the exit for the Toyota dealership. I called the dealership and told the service manager to meet me outside because I was experiencing acceleration problems. I drove approximately 5 miles by alternating from neutral to drive and pressing very firmly on the brakes. As I pulled into the front of the dealership I put the car into Neutral and exited the car. With the brakes smoking from the excessive braking and the car's rpm's racing the manager entered my car. He confirmed that the mats were properly in place and confirmed the rpm's were very high. They contacted a Toyota tech to look at my car, after 1 1/2 weeks with my car they were told by Toyota to replace the throttle body and accelerator assembly including both sensors. I no longer feel safe driving this car and I am not convinced that Toyota has found the root cause of this problem. 10/12/

69 VOQ No. Vehicle Description Camry unk Both brake pedal and accelerator pedal are designed too close together. Feel free to provide any further information Camry 2.4L L Camry unk When the brakes were applied, the gas pedal was pressed accidentally, which caused unexpected acceleration. The brake pedal had gone to the floor, below the gas pedal which had felt spongy. The consumer had taken the vehicle to a service dealer which stated that the brakes were out of adjustment. After the repair was done, the brakes still remained the same Toyota Camry incidents happened before current acceleration hoopla. Twice to my wife, twice to me. When quickly applying brake, foot touches gas pedal and car accelerates. Never on any other car, including a 2005 Camry has this happened. I believe that the relation of both pedals to each other is too close. Also the design of both does not help. I drive autos, not design them, but I think this is more of a concern than floor mats. Thanks for your consideration. 10/12/

70 VOQ No. Vehicle Description Camry 2.2L L Camry 2.4L L Camry 3.5L V6 When approaching a parking space, the driver pressed the brake pedal and the vehicle accelerated. The vehicle crashed into a store window. Vehicle was taken to the dealer. The consumer contacted the manufacturer who replied by saying that the incident was not the result of any type of manufacturer design or defect. The manufacturer wants to close the case. When the brake pedal was depressed vehicle suddenly accelerated. Consumer heard the engine motor racing. This occurred while pulling into a driveway. Vehicle hit a parked vehicle's bumper, a post, and a fence before stopping. Insurance claims department said that a number of complaints regarding on the vehicle on the electronic system which gives information to the throttle control system. The contact owns a 2009 Toyota Camry XLE. The contact was backing out of a driveway. There was pressure applied to the brake pedal, which was followed by unintended acceleration. The vehicle crashed into the utility pole. There was no personal injuries. The vehicle sustained minor damages. The police was not called to the scene. The failure mileage was 13, /12/

71 Alleged defect: electronic throttle model failure Wear out condition with potentiometer type TPS Safety consequences: limp mode & engine stall Population 331,301 Complaints 15,631 Engine stall 479 Limp mode 729 Crashes 12 Warranty claims 82,478 Claims with codes associated with limp mode 14,965 Investigation included assessment of mfr fail safe strategy Recall 06V /12/

72 Alleged defect electronic stability control malfunction Fretting corrosion of steering wheel position sensor connector Safety consequences inappropriate ESC activation Inappropriate braking with no brake lights Risk of lane departure from braking pull Population 40,028 Complaints 58 Crashes 4 Warranty claims 2,424 (SWPS connector repairs) Testing to simulate fault condition: Fault detection normally occurs in less than 1 sec (ESC deactivated) Fault injection produced range of sensor voltages where fault detection may be delayed by several seconds Recall 10V /12/

73 Alleged defect electronic stability control malfunction Yaw rate sensor voltage drift results in erroneous yaw signal Safety consequences inappropriate ESC activation Inappropriate braking with no brake lights Risk of lane departure from braking pull Population 95,900 Complaints 10 Crashes 0 Warranty claims 462 (yaw rate sensor claims) Testing to simulate fault condition: Fault detection normally occurs in less than 1 sec (ESC deactivated) Fault injection identified voltage drift condition that may take as long as 10 sec to detect (later reproduced in complaint vehicle) 10/12/

74 Investigation of UA Important to understand driver experience for each incident Proper categorization/ binning of complaints Each category requires independent analysis Pedal entrapment incidents are different from incidents alleging sudden increase in engine power while parking and/or applying brake Analysis of UA requires understanding throttle & braking effects Publicity can significantly influence complaint rates & trends Investigation of electronic defects Experience has shown that electronic defects provide evidence for investigation or can be reproduced by testing Fail safe strategies can be evaluated through fault injection or testing with faulty parts 10/12/