Acoustic Fault Detection in Tactical Flow Meters

Open Access
Patterson, Kevin A
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
May 25, 2017
Committee Members:
  • Karl Martin Reichard, Thesis Advisor
  • David Carl Swanson, Committee Member
  • Daniel Allen Russell, Committee Member
  • acoustics
  • flow meter
  • condition based maintenance
The United States Army is beginning to use turbine flow meters to track bulk fluid fuel transfers and receipts from tanker trucks to storage tanks, bags and end users. Previous research has been conducted to improve the accuracy of measurements used to record large-volume fuel transactions. Installation of this research included implementing temperature, density and viscosity correction algorithms and automated reporting on measured fluid transfer volumes onto the embedded processor. The objective of this research was to use acoustic and vibration data to detect failure modes which can cause flow measurement error. Lab testing on the flow meter’s turbine blades determined key frequency components in the moving parts of the flow meter. Field testing, at a make-shift ARL fuel storage and distribution farm (using water as a surrogate for fuel), determined the characteristics of a normally functioning meter. Several failure modes were detected when field testing these flow meters resulting in flow measurement errors greater than the allowable 0.5% of total volume. After initial analysis, testing was conducted to introduce two faults into the flow meter: air entrainment and debris caught in the flow loop. Further analysis was performed to identify characteristics of these two failure modes. Air entrainment in the flow meter caused a significant decrease in amplitude to a 4kHz tone normally present in vibration data. Debris caught in the flow meter turbine blades resulted in an overall increase in vibration amplitude as well as a drastic increase in amplitude to frequencies between 1.6kHz and 1.8kHz. Based upon both of these observations, preliminary detection algorithms were developed with the intent of Condition Based Maintenance. Future work to improve flow meter accuracy includes detecting calibration drift and internal sensor placement.