Hydraulic Pump Durability Testing 1

[JDG1977]

I work in hydraulics in off-highway equipment. I am curious what other people in industry do to verify the durability of hydraulic pumps (piston pumps, gear pumps). Our testing leans toward "we've run this test for years and haven't had any problems in field, so if we pass it again it must be ok." We have a couple of durability tests that are run. The frustrating part to me is there is really no correlation between the tests and field hours.

On the other end of the spectrum with fatigue testing of say a valve we do a very good job of defining the S-N curve for the component and measuring the load history and combining the two to get a life prediction. With this method we are able to predict what effect a pressure increase or other change would have or look at what it would take to double the life goal quite easily.

With pump durability a pressure increase or other change would require another round of durability testing and there really isn't a good answer for doubling the life goal other than double the length of the test. I'm not sure if it is possible to predict things like wear and other durability issues as well as fatigue, but I wanted to get an idea of what other people do.

 

[JohnRobinsonOz]

I'm not sure if this is of any help but I too have worked for many years designing mobile off road hydraulic systems (explosives manufacturing trucks for mines in Australia, Asia & Africa).

My personal experience is that mobile equipment life expectancy rarely depended much on continuous wear conditions, rather it depends on unusual loads i.e. a hose rubs through and the operators run the pump dry (fails very quickly maybe a fraction of a second on a dry suction). Intermittent cavitation, excessive heating (i.e. poor power efficiency design), external corrosion, strange maintenance practices (I've seen examples of valves being hit with hammers to "free them up&quot all tend to destroy mobile equipment, particularly in the mining environment.

Conversely I've had some experience with fixed installations in steel works rolling mills. These are continuous duty 24hr operations. Here life expectancy does have correlation with run time, however once again its more dependent on other things than manufacturers advised design life. Usually filtration and oil (glycol/water in hot areas) quality are far more important.

In all of the above the important thing is to design with the operating practices in mind.

If you truly just need standards for testing pumps I'd be looking around at what the bearing / gear industries do and how they arrive at statstical life expectancy for antifriction bearings. Afterall the wear in a pump will be a similar mechanism. Spalling for fatiguing surfaces and errosion where there is only fluid contact.

 

[vanstoja]

To design an endurance test for critical pumps that must operate for considerable periods of time without mauintenance it is necessary to know or find out:
1. How the pump and its driver is really operated over a typical period, a year for example. How many start/stops? How many pressure surges from check valve closures and what magnitude of pressure surges? How many hours at fast speed, slow speed? What are the pumped fluid temperature transients and how many of each kind?
2. What are the pump and driver failure modes and how are they affected by the operating steady state and transient conditions.
3. What are the environmental conditions and their variation over the selected time period? Ambient temperatures, vibration levels at the driver and pump bearings, crud levels in the pumped fluid, etc.

With the above kind of data in hand you can decide what fraction of the entire life you want expend in the endurance test. Say simulation of 20% lifetime transients will carry you safely thru "infant mortality" failures and significantly enhance prospects of meeting full life for most of the units in service. Combine those fractions for all of the important transient events you can realistically run in an endurance test, select a unit for test that you are willing to expend, or repair as necessary for return to service. Ideally, devise a single test cycle that includes all of the conditions you want to simulate in a one-day, perhaps two-workshifts timeframe and dedicate as many days as necessary to run up your 20% simulated life-cycle test.
This process is regularly done for some vital motor-driven centrifugal pumps that are expected to run
maintenance-free for hundreds of thousands of hours. It has proven to be a valuable endurance test that is specified for and applied to every new pump and driver design for the same operating service. Even when designs pass such a test, they often show signs of distress in some area that suggests where design improvements would be worthwhile.

 

[JDG1977]

Thanks to both of you for your input. We do run into similar poor maintenance type failures like JohnRobinsonOz mentioned, but I wouldn't say that is the majority of our failures.

Vanstoja's comments are quite similar to what we do today, but the direct tie to machine hours isn't there. We do take what we call a pressure/load history which would give all of the application information that vanstoja mentioned. I think that for gear pumps or vane pumps (constant displacement), I have a good feel for how to design a test. Most of our applications run at fairly constant speeds so the only major variables become pressure and environment. We have enough data that we could probably come up with typical environmental conditions to run in. The pressure/load history could then be used to define a cycle.

I am having more difficulty with piston pumps and the real issue seems to be what vanstoja mentioned in number 2 above. I'm not sure we have a good feel for how the transients effect the life of the pump. For example in some of our applications the displacement changes a great amount, but fairly smoothly. In other applications the displacment changes very quickly, but over a shorter range. What we don't know is if long smooth changes over the entire displacement range or many short quick changes near neutral are worse. We do have the data to show typical pressures, displacements, temperatures, speeds, and environment over the life of the pump. The issue becomes reducing that down into a reasonable cycle. Our applications see about every combination of those variables. Some combinations may be rare, but they may have a big effect on life.

Any way again thanks for your input. I am afraid the real answer to the question is to run several durability test to both understand all the effects (so far pump suppliers haven't been much help in this area and we work with most of the major pump suppliers) and to compare to field results. Either that or stay on a similar path to where we are today. With the investment involved in testing several pumps I assume we won't have a choice and will have to stick with close to what we have today.

 

[automatic2]

Your best test is your actual field operations. It includes the dynamics of application, maintenance, enviroment. My years in mobile have shown me that contaminated fluids are the largest problem. Leaking rod seals allowing contaminates in, poor maintenance practices, overheating of oil. Cold starts with subsequent cavitation, hard actuations or misuse by operators is close behind. A strong pump with a bad alignment has limited life. Measuring leak down on your piston pumps once a month will keep you abreast of there life cycle.