Advanced Automotive Component Failure Prediction and Simulator Control System
Objective
This project will develop an intelligent system to aid and accelerate the task of developing new designs for complex mechanical automotive components by simulating the crankshaft dynamics of various automotive engines in order to detect incipient faults in automotive accessory drive components designed to be driven by the crankshaft of the engine in question.
Deliverables
A torsional vibration simulator for an automotive accessory drive capable of serving as an analog for a real automobile for long duration tests of rotary components including start-up, nominal operation and deceleration phases of the duty cycle. This system will have the ability to test real-world components at an automotive manufacturing facility.
Technology & Benefits
The fault detection functionality of the proposed system represents a totally new application of AI-based novelty detection classification algorithms to identify faults in the testing of prototype and production automotive components. A unique feature of this fault detection system will be its combined use of real-time control data to augment the sensory data measured from the part itself. This will allow for measurements to be placed in context of what duty the component is being subject to during the test. This integration of real-time operating mode data from the control system with the fault detection system represents a new level of sophistication in parts testing. In addition to the novelty detection based fault detection, the system will employ a parallel flexible expert system. This expert system will allow for hard rules to be applied for known fault states that can be readily defined (such as a rolling element bearing temperature reaching a pre-determined alarm state).
The system proposed aims to reduce powertrain parts development cost and increase the quality of powertrain parts development. By reducing engine and drivetrain part torsional testing costs and increasing the level of confidence that these tests accurately reproduce real life conditions, auto parts manufacturers will in turn be able to better compete in the global marketplace.
Precarn - Small Company Automotive
Participants
- Quanser Inc., Markham ON (Lead)
- Laurentian University, Sudbury, ON
- Litens Automotive Group, Woodbridge ON
Project Duration
Start: January 2008
End: March 2010
A Precarn-APMA Initiative
Email this to a friend
