Testing Centrifugal Pumps Part 1
Simon Bradshaw
Global Director of Engineering and Technology at Trillium Flow Technologies
So it's been a while since I last visited the topic of testing pumps. While I've covered specific aspects of testing in some past articles (some of which I'll link to later), I figured it was past due to provide readers with a consolidated update on pump testing.
I'd be remiss in mentioning that there are already a few (mostly) excellent pieces of content on this topic. I've linked to them below, asking readers to keep in mind that since they were published, API 610 and HI 14.6 have both been updated.
With that out of the way let's see if we can hit some of the key topics regarding testing.
Why do pump testing ?
The answer is essentially for the same reason that we do testing on any complex mechanical device - to verify it will perform as expected in the application it will be used.
Keep in mind that all complex mechanical devices have many components each with their own tolerances and manufacturing controls. In order to ensure that when assembled together everything performs within acceptable limits, some testing is appropriate.
The level of testing expected tends to follow the criticality of the device. So (for example), a pump that provides the crucial function of circulation coolant around a nuclear reactor core is going to receive far more scrutiny and testing that the pump used for pumping out your basement drainage.
What standards exist to govern pump testing ?
The following current standards exist to guide you (I've linked to their respective sources where possible). These are the primary standards covering (at a minimum), validation of the hydraulic performance.
You may also (less commonly) run across the following test standards
There are also additional standards that govern secondary aspects of testing
The table below summarizes what aspects of testing each of the standards mentioned above will cover.
The important takeaway from the table is that depending on what you want to validate, more than one test standard may be necessary.
What tests should I be performing on my centrifugal pump ?
If you are buying a centrifugal pump that complies with a certain standard - for example ISO 5199 , 9905 , 9908 , ANSI B73.1 or API 610 then the testing requirements will be identified within the standard itself.
For example API 610 has a mandatory requirement for hydrostatic test, and a performance test (which includes vibration and bearing temperature rise measurement).
Keep in mind that a significant number of tests in these standards may be optional - i.e. the purchaser needs to specify them at time of order. To summarize the test requirements in each of the main construction standards:
Additionally, many large End Users of centrifugal pumps (think of companies like Dow or ARAMCO), will have overlay specifications which call out their specific requirements for testing.
However if you are in the position of needing to specify the extent of testing for your pumping equipment, then the next few paragraphs may be beneficial.
Hydrostatic test
This is mandatory in all of the pump construction standards (except ISO 9905) and should always be performed. Nobody wants a leaking pump at site even if the fluid leaking is harmless.
The effects of a pump pressure casing failure can range from a minor annoyance through to catastrophic flooding if a large cast iron pump casing were to fail in service.
The test is not expensive for pumps produced in volume.
领英推荐
Performance test
This in mandatory in API 610. For lower cost volume produced pumps on routine spared services a performance test is going to add limited value.
Where it does add value is when any of the following apply:
Performance test @ site NPSHa
This is an optional API 610 test. Essentially it is conducted just like a normal performance test with a single key difference - the NPSHa at the pump suction is held to <= 110% of the site NPSHa. It requires more careful setup and control of the test loop in order to be able to hold the NPSHa within fairly narrow limits.
This test has value when testing pumps handling water (or aqueous solutions) where the site NPSH margin (NPSHa/NPSHr) is known to be limited.
It will confirm if the pump exhibits any anomalous behavior such as increased vibration under site conditions. An example of where this test did uncover exactly this type of behavior can be found here . (Fortunately for the end user, in that case the pump performance could be corrected.). You can see the effect of site NPSHa in that case below
Typically you'd most likely specify this for larger pumps above a certain power. If you pushed me to say at what power I'd specify the test, for pumps > 500 kW (671 HP), this test should be considered for pumps on water or water like fluid services.
Complete Unit Test
Just as the name implies, this is a test of a complete package.
The primary purpose of this test is to ensure all the unit components function correctly with respect to each other. This provides the customer with assurance that when the unit is installed at site, the need to troubleshoot the package is minimized.
Most commonly the test will comprise of:
Less commonly you may see a complete unit test conducted with VSDs (variable speed devices) such as VFDs or variable fill fluid couplings.
If that potentially sounds like a lot of complexity, then you'd be correct - it can be. The setup for testing really complex packages can take weeks and the setup costs well into six figures.
It is critical that the test parameters and sequence of testing is carefully set out to ensure that each component of the package can be validated under a range of operational scenarios - including startup, shutdown and fault conditions. A close partnership between an experienced OEM and the customer is the best way (in my experience) to ensure that the resulting test program meets the intended goals
At the same time we need to be aware that the unit is being tested in a temporary setup. For example the baseplate will not be grouted in place. The pressure energy imparted to the fluid by the pump will need to be converted to heat through pressure breakdown valves. Some of that conversion will end up as vibrational energy in the test pipework and fed back into the pump.
All this means that the vibration performance of the unit will be worse than the levels achieved at site. It is therefore important that the OEM and the customer agree on how vibration performance will be assessed during testing. This will typically include the agreed levels and measurement points. (I've seen inspectors pull out uncalibrated portable vibration measuring devices and start measuring obscure points of the package then use these readings together with their own pass/fail criteria).
Sound Level Test
Determining the sound level produced by a pump in a test environment is not straightforward. Test environments are noisy with multiple sound sources making it difficult to isolate the pump noise. These sources are commonly:
On high power tests with the test loop pressure breakdown valves in close proximity to the pump, the test loop noise can exceed the noise emitted by the pump. This means taking a simple sound level measurement cannot be relied upon to determine the true pump noise level.
There are more advanced methods available that allow the noise level contribution of the pump to be isolated with a reasonable degree of certainty. ISO 3744 or 3746 can be utilized in a test environment (ISO 3744 is the more accurate, but has more limitations on the test room setup and allowable background noise levels).
With these either of methods, the sound pressure level is measured using microphones placed in a defined pattern around the equipment being tested. The diagram above shows how this is done. A reference "box is defined around the boundaries of the equipment being measured. Then the microphones are places at specific locations in either a rectangular or hemispherical pattern. After taking all the microphone readings the resulting sound power level of the equipment can be computed.
Keep in mind this test requires specialist equipment and personnel to accomplish. It also has an inherent uncertainty of 1.5 dB for ISO 3744 and 3 dB for ISO 3746.
Ok that's it for Part 1. In Part 2 we'll cover a few more niche testing requirements as well as any questions raised. As always, all comments, questions and critiques are most welcome. I always appreciate the feedback.
Until next time Beatus Centrifuga
Sepanta Pishtaz: Expert in Component Repair, Engineering Analysis, and Custom Tool Design for Rotary and Fixed Equipment
2 个月It's great to see the topic of pump testing being discussed!
Retired at Ex-CelerosFt
2 个月It may be of interest to people un-associated with testing or witnessing to include an article on issues encountered, and to be considered, in pumps operating on a different S.G. where contract motors/couplings may not be rated with sufficient power for string testing, and reduced speed testing may be required. Also test seals required if face materials selected for a pumped product may not be suitable for use with water which is the general test pumpage.
Principal / Vice President at Freese and Nichols
3 个月Thank you Simon! Having all of those references together is super helpful. I especially like the summary table of standards. I'm looking forward to Part 2. If you are taking requests, I'd like to see a discussion on applicability of factory vibration measurements to field installation values and any suggestions for successful factory tests where that data is desired.
Overseas Sales Manager
3 个月Good to learn that ??
Something that has bugged me for a while, hopefully some one could clear it up for me. Say, if a pump has an operating point of 60L/s and a head of 45m. It is tested at the manufacturers site, and the performance test results are such that it meets the requirements of 60L/s and 45m. However, on the data sheet, it has Acceptance standard ISO 9906 class 3B which has a +/- on the tolerance. Should it not be an ISO 9906 class 1U that has only + tolerance across all test parameters.