Vibration Analysis ~ Condition Based Maintenance

Aviation Condition Based Maintenance

In 2005 a unique opportunity presented itself, one that consisted of studying a 17 component Apache (AH-64) tail rotor drive train (TRDT) system's vibrations for the purpose of improving the Army's Condition Based Maintenance (CBM) program, a program established to monitor aircraft components via accelerometers in order to detect and diagnose failures. Since then, three test stands were constructed and operated in order to acquire aircraft component data for the purposes of improving vibration monitoring algorithms. The original intent of this project was to monitor seeded fault bearings and differentiate those vibration signatures from the baseline and non-seed faulted bearings. The analysis was broken up into two sections baseline testing and seeded fault testing. It was quickly realized that studying vibration signatures for bearings surrounded by so many non-master components was going to be a daunting task due to such a largely cross-coupled system. By not utilizing “Golden Suite” parts for components not under test, the frequency data for the seeded fault tests were clouded by surrounding components which would later be seen as, non-intentionally inserted fault components. In addition to noise from neighboring components, vibration signatures were affected by loading...(read more)

Incorporating Flight Regime Recognition into Frequency Sorting and Analysis Processes for the Purposes of Condition Based Maintenance

This research identified a methodology for locating nontraditional frequencies of deteriorating rotating components and their locations as related to AH-64 tail rotor components as mounted to the TRDT test stand. For these selected frequencies, this research also illustrated the importance of load monitoring, the influences of shaft misalignment, shaft's imbalances and their effects on vibration signatures and the importance of flight regime recognition. The plan for this research was to devise a way of extracting useful power train information from the test data in order to reduce the amount of false alarms while increasing the ability to detect faults for the last three stages of bearing fault progression.

The desired outcomes were reduced false triggers and misdiagnosis and a reduction in the amount of data monitored by eliminating frequencies that provide no insight on the state of a components health. This also reduced post processing time and storage space required to handle large amounts of data by focusing on frequencies that yielded the most information in reference to a bearings state. Another desired outcome was the incorporation of the affects of loading into the bearing life plot in order to normalize the acquired in flight data. It was also desirable to provide information that improved the ability to create more robust CIs, modification of the existing bearing life algorithms and assist in creating a more successful condition based maintenance program...(read more)

Frequency Sorting When working vibration data related to 17 simultaneous components it can be difficult to establish what vibrations are due to which component, much less which part of what component. Difficulties in establishing what frequencies belong to which components can be due to noise and/or due to components, subassemblies or parts sharing similar frequencies. When manipulating vibration data, depending on the method of manipulation, similar frequencies can have constructive or destructive properties and is known as the superposition of waves; this superposition of waves can complicate locating the correct source or sources.

For the purpose of this research, as much as possible, processes will be used as if no other AH-64 vibration information exists. The only component information used will be component locations, locations relative to each other and specific geometrical information. Beginning with this assumption, some method of sorting the data must be used in order to dig into the vibration information for the lowest part of a subassembly. This first step will be a large scaled sorting, this is where the NL TRDT stand provides great insight. Excluding the empennage, the NL TRDT has all of the components of the TRDT. This huge difference provides a vehicle to test similar hanger bearing assemblies and drive shaft components on either stand and enable us to separate vibrations due to gearboxes from hanger bearing assemblies and drive shaft components...(read more)

Shaft Misalignment The second step in minimizing stand to stand variability was the alignment procedure. This step was a pain staking procedure that, in the beginning stages, consumed as many as eight hours. As this procedure was repeated the times were reduced to a few hours. The original method of choice, as suggested by AED, was to use a typical industrial laser alignment device. The problem with this equipment would have been in how it mounted. In order to use these industrial devices, the equipment would have required that they be mounted to the shafts. Due to the flex coupling in the AH-64 tail rotor drive shaft the weight of any device hanging on the shafts would have thrown off the alignment. Although these devices may have not thrown off the alignment a great deal, the max misalignment angle was only set to be 1.3 degrees and would not have required a great deal to be out of tolerance. Using some retired physics optical equipment scavenged from the basement and some stainless steel mirrors another method was devised.

Using an optics indexing table to measure the shaft and flange angles, the non-rotating base was placed half way between the input flange of the intermediate gearbox and 1st flex coupler of the tail rotor stand. Next, the stainless steel mirror was attached to the input flange of the intermediate gearbox and the last hub of the isolation shaft. with three wing nuts a low powered laser was mounted to the rotating portion of the table...(read more)

Shaft Imbalance Beginning with the five aforementioned baseline bearings, it was important to perform repeatable tests that would produce data that would minimized the amount of variables. Being that the baseline tests were going to be run on two different stands, there were going to be some inherent differences. There were specific shafts that were used for each of the stands and the phasing of these shafts would be important for analysis purposes. Prior to baseline testing, practice tests were run on both stands so that the balanced shafts could be phased relative to each other such that the lowest vibrations were produced by them. Once the phases of the balanced shafts were located, all of the components were marked on either stand...(read more)

Torque Monitoring Without a doubt it can be stated that loading has a great impact on frequency's amplitudes. For some frequencies the amplitudes increase with loading, while others decrease with loading increases. Some work has been done in this area to quantify the influences of loading on various frequencies, to resolve this issue more work needs to be performed in software writing and post processing. There may even be a better way to determine the onset of failures by comparing signatures at another load range other than flat pitch ground 101 (FPG-101), because it may be possible to reduce noise in the tail rotor drive train system at other loads...(read more)

NL-TRDT Baseline Observations Beginning with the NL-TRDT stand all of the baseline bearings were installed one at a time and run on the NL-TRDT stand in the aft hanger bearing position for a half an hour. For the NL-TRDT test stand all bearings were run without having a load applied to the drive shafts and were run aligned. During each half hour test, accelerometer data was acquired every two minutes. Including the first vibration measurement taken at time zero there were a total of 16 acquisitions of vibration data for each baseline bearing...(read more)

NL-TRDT Seeded Fault Observations From the beginning of this project statements were made that suggested that the location of the test article would have no affect on the purity of the test and that it would be unwarranted to run only a single seeded fault component at a time. Depending upon the type of data analysis, not seen thus far, this may be true, but for the most part it would false. From the beginning of baseline and seeded fault testing it had been suggested that only one article of interest be run at a time, with the remaining components being Golden Suite components, until more information was known...(read more)

Vibrats.com

Home

Related Link Category

Aviation Condition Based Maintenance

Incorporating Flight Regime Recognition into Frequency Sorting and Analysis Processes for the Purposes of Condition Based Maintenance