Abstract : This paper documents the investigation of fuel injector fault detection methods for a seven liter diesel engine. This effort was conducted for the Tank Automotive Research Development Engineering Center (TARDEC) Condition Based Maintenance (CBM) team. The task was to develop algorithms capable of real-time detection of injector misfire events. The purpose of this task was to enable TARDEC's Engine Control Management (ECM) research and development efforts to evaluate the technical feasibility of integrating automated on-board condition monitoring algorithms with future ECM monitoring and control operations. During this investigation, it was shown that multiple techniques can correctly detect and identify injector cylinder misfiring. Each individual technique has its own advantages, and this investigation focused on low computational power signal processing methods that would be suitable for embedding in an engine controller or processor. The list below includes six injector fault analysis approaches that were evaluated for this effort 1. Injector signal based analysis 2. Vibration based analysis 3. Time domain speed encoder analysis 4. Order domain speed encoder analysis 5. FFT classifier selection techniques 6. Time domain classification techniques This paper highlights the first four diagnostics techniques. On-platform tests are suggested for technique validation and future development of these initial findings.
[1]
K. Reichard,et al.
Diagnostic fault detection for internal combustion engines via pressure curve reconstruction
,
2003,
2003 IEEE Aerospace Conference Proceedings (Cat. No.03TH8652).
[2]
Matthew A. Franchek,et al.
An investigation of crankshaft oscillations for cylinder health diagnostics
,
2005
.
[3]
Robert Campbell,et al.
Failure Modes and Predictive Diagnostics Considerations for Diesel Engines
,
2001
.
[4]
K. Reichard,et al.
A Non-Intrusive Technique f or On-Line Shaft Crack Detection a nd Tracking
,
2005,
2005 IEEE Aerospace Conference.