Slippage Detection and Traction Control System Reliability Report October 19, 2004 Sponsors Dr. Edwin Odom U of I Mechanical Engineering Department Advisors Dr. Jim Frenzel Dr. Richard Wall Team Members Nick Carter Kellee Korpi Brian McConnel
I. FAILURE RATE AND MTBF Using the Relex 7.7 Reliability Software we were able to find the failure rate for the majority of the systems parts and of the system as whole. The Relex software allows you to input and specify different components, electrical and mechanical, as well as different connections. Some parts could not be specified under Relex s descriptions so they were specified by the closest matching description. The list of components, connections and their failure rates are presented in Appendix I. Based on Relex s library of statistical information the failure rates were calculated and given in percent failure per 10000 hours. After obtaining the failure rate from Relex, the mean time between failures (MTBF) was calculated by taking the reciprocal of the failure rate. The failure rate and the MTBF for some of the more crucial parts of the system are presented in Table 1. As the table shows, the overall system failure rate is very dependent upon the inductive sensors. This indicates that a problem would most likely occur with the sensors. TABLE 1: Reliability Scores Part Failure MTBF Rate (hours) Overall System 1.692 590,716 Microcontroller 0.0192 52,083,333 Inductive Sensors 0.0944 10,593,220 The MTBF values may in actuality be lower due to the mobile environment the system will be operating in. II. FAILURE MODES, EFFECTS AND CRITICALITY ANALYSIS (FMECA) FMECA identifies possible failure modes and assigns a rating in the form of a number. First a list of possible system failures was established and is shown in Table 2. Next, these possible system failures were given ratings in three categories. The three categories were severity, occurrence and probability of detection of each failure. A. Rating Scale The rating scales chosen for the three categories were adapted from [1]. The rating scales from this paper are presented in Appendix II. Severity was rated on a 1 to 10 scale. A 1 indicates that the problem can remain unnoticed and the system continues to function and a 10 indicates that the problem is hazardous and the system is inoperable. The probability of occurrence was rated on a similar scale with 1 indicating an extremely remote chance of occurrence and 10 indicating an extremely high chance of occurrence. The probability of detection of one of the failures was also ranked on a 1 to 10 scale. A 1 being an almost certain detection and a 10 being almost undetectable. B. Risk Priority Number Once a reasonable ranking was agreed upon for each of the possible failures in the three categories, the Risk Priority Number (RPN) was calculated. The RPN is the product of the three rankings for each possible failure. A high RPN indicates a significant risk of system failure that should be corrected to reduce effect severity, the probability of occurrence and to increase the probability of detection, according to Dr. Anderson. Possible failures with a high RPN should be addressed in the design phase. Table 3 presents the list of possible system failures along with their ratings in the three categories and their RPN. They are listed by decreasing RPN. TABLE 2: Possible System Failures 1) Switch Connections 2) Sensor Damage 3) Sensor Reading Error 4) PCB 5) Encoder 6) Counters 7) MOSFETS 8) Microcontroller 9) Software Failure 10) Power Loss 11) Component Disconnection 12) Contained System Damage 13) Reset of ECU
C. Results As Table 3 shows, software failure is the highest liability. This will need to be taken care of in the design phase with an abundant amount of testing to ensure the code will not fail. The sensors are also a high risk. This will be addressed with testing to determine the ideal mounting position and number of teeth for the toothed wheel target. The switch, as a mechanical device, is at a higher chance for failure. This problem can be addressed in the trouble-shooting manual, with a recommended switch replacement, if this becomes an issue. TABLE 3: FMECA Table of Ratings Severity Occurrence Detection RPN Software 5 5 8 200 Failure Sensor Reading 5 5 7 175 Error Switch 4 4 6 96 Connections Counters 6 3 4 72 Microcontroller 6 3 4 72 PCB 8 2 4 64 MOSFETS 8 2 4 64 Component 4 6 2 48 Disconnect Sensor Damage 6 6 1 36 Encoder 4 2 4 32 Reset of ECU 5 2 3 30 Physical 4 7 1 28 Damage Power Loss 3 3 3 27
APPENDIX I COMPONENTS, CONNECTIONS AND FAILURE RATES
APPENDIX II RATING SCALES
REFERENCES [1] D.O. Anderson, Ph.D., Hazard Analysis in Engineering Design, Louisiana Tech University, 2001.