How Flow and Vibration Affect Filter Performance

Gregory LaVallee, Donaldson Company, Inc. Philip Johnson, Donaldson Company, Inc.

Standard hydraulic, lube and fuel test procedures do not always reflect real-world filter performance. Laboratory tests have shown that when vibrations and unsteady or cyclic flow conditions are present, finer filtration of the oil becomes necessary in order to achieve the same efficiency ratings as are found in the lab.

Tests also indicate that the viscosity of the fluid itself affects the filter’s efficiency rating, under dynamic conditions.

Several authors have reported the degradation of performance for hydraulic filters under unsteady flow conditions. Also, the National Fluid Power Association (NFPA) 1994 has worked to develop a multipass test method using unsteady flow conditions.

The effect of engine operation on fuel filtration has also been studied. Both engine vibration and fuel pump pulsation are known to cause degradation of filter performance.

This article presents some test system modifications and test results under unsteady conditions. It also relates these results to real-world systems and suggests alternatives for filter selection and placement in the system.

Test Systems and Procedures

Two test systems were used for this study. First, all cyclic flow tests and some vibration tests were performed on an ISO 4572 standard multipass system for the evaluation of hydraulic filter performance.

This system allowed continuous recirculation of contaminated MIL-H-5606 hydraulic fluid through the main system and test filter. Fine test dust was continuously injected into the main flow from an injection system. A sampling and dilution system with two online particle counters was used to measure particle penetration through the test filter.

Additional vibration tests were performed on a J1985 fuel bench (a single-pass system) designed to measure the particle removal characteristics of fuel filters. The contaminant was not allowed to recirculate through the filter more than once. It used ViscorTM fluid (specially developed fuels with properties resembling diesel fuel and/or gasoline) with similar injection and sampling systems.

The test conditions were as follows:

  • Flow rate: 3 and 7 gallons per minute (gpm)
  • Cyclic frequency: 0.1 and 0.5 hertz (Hz)
  • Cyclic amplitude: 75 percent and 50 percent reduction
  • Vibration frequency: 15 Hz
  • Vibration acceleration: 2 g, 4 g and 6 g (g = gravitational units, 1 g = 9.8 m/s2)

The test housing, filter element sizes and flow rates were selected to accommodate the range of conditions seen in typical filters for diesel fuel, lube oil and hydraulic applications. Cyclic flow conditions were selected based on the NFPA proposed standard (1994) and what was judged reasonable for the test system.

Vibration conditions were based on conditions that would likely be encountered by some of the filters considered. Items related to element design were held constant.

The basic format for testing involved running normal steady tests. Subsequent cyclic flow and vibration tests were performed and compared to the steady tests.

Cyclic Flow

Cyclic flow was achieved by adding an actuated ball valve triggered by an electronic relay and timer system. The ball valve was positioned in an existing bypass line of the multipass main flow system as shown in Figure 1.

Cyclic Flow Main Flow System
Figure 1. Cyclic Flow Main Flow System
Showing Addition of Actuated Ball Valve

The rated flow was set with the ball valve closed. When the valve was opened, the flow was bypassed to a level controlled by a needle valve.

Thus the flow through the main line and test filter was cycled at a controlled amplitude and frequency as shown by the square wave forms in Figure 2.

Flow Cycle Wave Forms
Figure 2. Flow Cycle Wave Forms for Two
Frequencies with 75 Percent Cyclic Reduction

The graphs show two cyclic flow frequencies. The high flow was generally set at the rated flow for the filter, and the flow was reduced by a specified percentage. The graphs shown are for a 75 percent cyclic reduction from rated flow.


For vibration testing, the test filter was isolated from the rest of the test system and connected by flexible tubing. Filters were vibrated using an electrodynamic shaker system. The vibration input was sinusoidal with a known frequency and peak acceleration. Input frequency was set at 15 Hz for all tests because this was found to be a natural frequency for the system. The peak acceleration input was varied by test. Although the filter could vibrate in all directions, the primary direction was horizontal or perpendicular to the filter axis.


Cyclic Flow Results

The results are shown in units of particle penetration. Particle penetration is the percentage of particles that are not captured by the filter. Lower particle penetration is better. These laboratory tests confirmed that cyclic flow conditions result in an increase of particle penetration. However, this effect is not significant until some contaminant has collected on the filter.

Figure 3 shows the results for the initial portion of the test where the filter could still be considered clean.

Cyclic Flow Results for Clean 3-Micron Filter
Figure 3. Cyclic Flow Results for Clean 3-Micron Filter.
Tests Performed in MIL-H-5606 Hydraulic Fluid at 3 gpm.

The first number in the legend is the percentage reduction in flow rate. The last number is the flow cycle frequency. Note there is little difference between steady and cyclic flow penetration, demonstrating that cyclic flow has no effect on initial penetration for the range of tests.

Figure 4 shows the results of loading, after contaminant has collected on the filter and pressure drop has increased to approximately 75 percent of terminal pressure drop.

Cyclic Flow Results for 3-Micron Filter Loaded to 75 Percent of Terminal Pressure Drop
Figure 4. Cyclic Flow Results for 3-Micron Filter
Loaded to Approximately 75 Percent of
Terminal Pressure Drop. Tests Performed in
MIL-H-5606 Hydraulic Fluid at 3 gpm.

Here, the increased level of particle penetration is obvious. In fact, the greater the reduction in flow during a cycle, the greater the increase in penetration.

Vibration Results

These experiments clearly show that vibration increases penetration. Vibration tests were performed using low particle concentrations and therefore should be considered clean. No vibration tests were performed with significant numbers of particles captured on the filter.

Vibration tests revealed that the test fluid had a significant impact on particle penetration. Tests using Viscor produced substantially higher penetration than the tests using MIL-H-5606 hydraulic fluid.

Figure 5 is a plot of penetration vs. particle size for some representative tests with these two fluids.

Vibration Results
Figure 5. Vibration Results Showing Fluid
Dependence for 10-Micron Filter. Tests
Performed with Vibration Frequency of 15 Hz.

The legend shows fluid type, flow rate and peak acceleration. Notice the high penetration resulting from the 4 g acceleration test in Viscor.

This increase due to vibration was fairly typical in Viscor, but was never observed in MIL-H-5606 hydraulic fluid. The obvious difference between these two fluids is the viscosity: 14.4 centistokes for MIL-H-5606 vs. about 2.5 centistokes for Viscor.

When vibration occurs, the increasing flow rate causes penetration to decrease. Figure 6 shows the results of three tests with no vibration and increasing flow rates. Note that the penetration did not change.

Vibration Results Showing Flow Rate
Figure 6. Vibration Results Showing Flow Rate
Dependence for 10-Micron Filter. Tests Were Performed
in Viscor Fluid at a Vibration Frequency of 15 Hz.

Tests with vibration showed very different results. While there was higher penetration at 1 and 2 gpm, maintaining vibrations at 2 g acceleration and increasing the flow rate decreased the penetration to steady levels.

Increasing the vibration acceleration increased penetration. Figure 7 shows a plot of initial penetration vs. particle size with different levels of vibration acceleration. The penetration increased nearly two orders of magnitude over the range tested.

Vibration Results Showing Acceleration
Figure 7. Vibration Results Showing Acceleration
Dependence for 3-Micron Filter. Tests Were Performed
in Viscor Fluid at a Flow Rate of 2 gpm and
Vibration Frequency of 15 Hz.

Vibration Combined with Cyclic Flow Results

The final test series combined vibration with cyclic flow. Each filter media was tested under the following conditions: steady, vibration only, cyclic flow only, and vibration combined with cyclic flow.

Results for penetration vs. particle size are shown in Figure 8. The combined tests had significantly higher penetration, which indicates the mechanisms were amplified or possibly synergistic.

Vibration Combined with Cyclic Flow
Figure 8. Vibration Combined with Cyclic Flow
Results Showing the Synergistic Effect and
Consequent Increase in Penetration for
3-Micron Filter. Tests Performed in MIL-H-5606
Hydraulic Oil at 2 gpm with a Cyclic Frequency
of 0.1 Hz and a Vibration Frequency of 15 Hz.

Recommended Strategies

  1. Filters should be positioned in lines with steady flow. The results from the lab are clear: unsteady or cyclic flow will result in higher particle penetration through the filter. Therefore, positioning the filters, where possible, in lines with steady (or steadier) flow should result in improved filtration and overall system cleanliness.

    An example from a real-world application is shown in Figure 9.

    Hydraulic Shovel
    Figure 9. Hydraulic Shovel

    This hydraulic shovel was experiencing high iron levels, an indication of wear. A filter was positioned in a steady line before an oil cooler. The results showed a dramatic reduction in iron particles, which clearly demonstrates how this strategy can improve cleanliness and reduce wear.
  2. More frequent filter changes may be beneficial. Based on the lab results, one should expect that loaded filters in the field could release particles under unsteady conditions. If the filters cannot be positioned in lines with steady flow, a simple solution to this problem is to change the filter more frequently. This removes contamination from the system entirely.
  3. Filters should be positioned to avoid vibration. The results from the lab are clear: vibration will result in higher particle penetration through the filter. Positioning filters away from vibration can improve performance.

    In the Donaldson labs, this was proven on a diesel engine as shown in Figure 10. The fuel filter was alternately mounted and removed from the engine during operation, and downstream particle counts were monitored.

    Particle Counts vs. Time
    Figure 10. Particle Counts vs. Time in an
    Engine Fuel Filter Application. If Accepted
    That Particle Counts Increase Wear, There
    is Evidence to Justify Some Attempt to Isolate
    the Filter from the Source of Vibration.

    When the filter was attached or removed from the engine, there was generally a brief surge in particle counts, followed by decay to lower counts. Small particles hardly decayed at all, and tended to remain at high particle counts as long as the filter was mounted to the engine. Large particles experienced almost a full recovery to pre-vibration count rates.
  4. Positioning a filter in a line with a higher flow rate may be justified. Results show that filters will allow more particles to penetrate under vibration. Increasing the flow rate diminishes this effect. While we do not have real-world data to document this, it is worth considering. At the very least, lower contamination levels result in a faster cleanup rate.
  5. More efficient filters may be needed for less viscous liquids. Tests also indicate that the viscosity of the fluid itself, under dynamic conditions, affects the filter’s efficiency rating.

    For most hydrocarbon-based oils, a 10ºC change in temperature on either side of 40ºC can increase or decrease viscosity by approximately 50 percent. If a system is likely to run hot, it would be prudent to use finer filtration compared to a similar system installed with an oil cooler.

    By contrast, finer filtration may be required for fuel filtration vs. oil filtration in order to obtain similar system cleanliness levels. Keep this in mind when designing bulk lube and fuel handling facilities, or when designing filter carts for handling multiple fluids.

    Filter performance may degrade substantially under dynamic conditions found in the field. Increased penetration, and therefore reduced system cleanliness, will result in increased wear. System performance can be improved by carefully applying and positioning filters in hydraulic, lube and fuel systems.
Subscribe to Machinery Lubrication

About the Author
About the Author