Coal, O&M

Over-speed Testing for Centrifugal Fans

Issue 4 and Volume 114.

Les Gutzwiller, Robinson Fans Inc.

Centrifugal fans for large power-plant and heavy-industrial applications are critical elements in the operation of the entire facility. It is, therefore, important to insure a long and safe operating life of the fan rotors without interrupted operation. Over-speed testing of the rotors at the manufacturer’s shop can help provide this insurance.

Shop over-speed testing of large industrial and utility fan rotors offers many benefits to the end user. During over-speed testing the rotor is subjected to stresses that are higher than it will ever see during normal operation.

The objectives of over-speed testing include helping the end user feel more comfortable knowing that these huge rotors have been tested at speeds well in excess of the actual maximum operating speed. A second benefit is that inspection of the rotor welded joints before and after testing (using non-destructive inspection techniques) provides further proof that the rotor can tolerate such high stress levels without cracks developing. And a third benefit is that subjecting the rotor to even one cycle at such high stress levels offers an advantage regarding the rotor’s fatigue life.

When over-speed testing is specified as part of a contract requirement, the designer and manufacturer must design not only with a significant stress safety factor, but also demonstrate that the finished equipment will perform satisfactorily before it ever goes into operation. These factors provide comprehensive proof that the materials and workmanship incorporated into the fan rotor have resulted in equipment that is safe to operate even at speeds above the expected operating range.

Over-speed testing does more than just provide a warm and fuzzy feeling. Centrifugal fan rotors for power plant and other industrial uses are designed using finite element analysis techniques (FEA). This gives the design engineer the information he or she needs to be sure that no stress in the rotor is high enough to result in macro- deformation or tensile failure. For example, assume the maximum allowed steady-state tensile stress (Von-Mises stress) in the rotor at maximum design operating conditions is limited to 90 percent of the material yield strength. For commonly used steel such as A514/A517, the yield at ambient conditions is about 100 kilopounds per square inch (ksi). The example FEA shows that only a few small localized areas of the rotor actually have stress levels near 90 percent of yield. Most portions of the fan rotor have stress levels far lower than this.

Therefore, when such a rotor (which has not been subjected to over-speed testing) is operated in the normal speed range, each start-stop cycle results in an applied tensile stress range of:

Stress Range before over-speed test: 0 ksi minimum to 90 ksi maximum.

Mean stress before over-speed test: 45 ksi.

These parameters are important factors in determining the low-cycle fatigue (LCF) life of the rotor.

Click to Enlarge
Figure 1 Finite Element model (for stress analysis)

However, when an over-speed test is performed, the highest localized stresses in the rotor actually exceed the material’s yield strength. For example, when the rotor is operated at speeds above the maximum design speed, stress increases in all parts of the rotor as the square of the speed ratio. So at 120 percent of design speed, the maximum calculated elastic tensile stress in the rotor would increase to:

selastic = (90% of yield at 100% speed) x (1.20)2 = 129.6% of yield.

This means some localized yielding will occur at the small areas of highest stress in the rotor. These small regions will be permanently “stretched” due to plastic deformation. The stress in the macro regions around these plastically deformed areas will only deform elastically during the over-speed test. As a result, following the over-speed test, the material surrounding the localized plastically deformed areas actually “push back’” against the areas that have “stretched.” The example stress vs. strain curve shows that just one cycle during over-speed testing results in a compressive stress of -15,350 psi at the areas of most interest.

Click to Enlarge


Calculations of true stress and true strain during initial over speed test loading can be done by determining the maximum localized stress in the rotor based on the finite element steady-state. Stress analysis based on 100 percent of rated speed, in this example 90,000 psi).

Click to Enlarge

Solve the Neuber equations simultaneously using an iterative technique.

Neuber Equation #1: σtrue εtrue E= (σelastic)2

Click to Enlarge

The next step is to plot εtrue (ordinate) versus σtrue (abscissa) for all speeds up to the maximum speed to be attained during over speed testing. In this step it is important to refer to the mechanical and fatigue properties of A514/A517 steel. Use Neuber equations (except use stress/2 and strain/2 per Massing) for unloading.

As a result, the cyclic stress range for this rotor’s most highly stressed areas is now:

Stress Range after over-speed test: -15.35 ksi minimum to 74.65 ksi maximum.

Mean stress after over-speed test: 29.5 ksi.

The lower mean stress will result in an improved low-cycle fatigue life for this rotor. In this example, the 2Nf reversals-to-failure is increased by about 23.5 percent as a result of the over speed testing.

Use the Smith-Watson-Topper mean stress correction (which is appropriate for systems that are primarily subjected to cyclic tensile stresses).

Click to Enlarge

During testing, the fan speed is slowly increased until the target over-speed revolution per minute is achieved. The speed is typically held for one minute. This is long enough to achieve all the goals of over-speed testing. The speed is then slowly reduced until the fan rotor comes to a complete stop.

The fan wheel is re-inspected to check for any cracks that may have developed during the over-speed test. If no cracks are found, the test is complete and satisfactory. If any cracking develops during the over-speed test, the cracks must be repaired and the over-speed testing repeated. The test manager ultimately will decide the pass/fail results for each rotor tested and will sign-off on the test documentation.


Quality Assurance


If you are specifying an overspeed test be run on your fan, make sure the facility conducting the test is quality assured. Look for the following to be available at the facility:

  • The test facility must have an ISO 9001 certification for both design and fabrication quality
  • A lateral dynamic analysis is to be completed to verify the rotor will operate below the first bending natural frequency
  • Material certifications are required for all materials used in the fabrication of the fan rotor
  • Fan wheel material must be tested ultrasonically to insure that there are no voids or laminations
  • The shaft that will be used for over-speed testing is to be a steel forging with certified material reports and ultrasonic test reports included
  • All fasteners must have certified material test reports. Fastener torque must be documented by quality assurance personnel to be sure the installation torque falls within acceptable limits
  • Rotor welding is to meet the requirements of AWS D-14.6 (for rotating equipment)
  • Dynamic balancing of the rotor is to meet the requirements of ANSI S2.19 balance quality grade G2.5
  • The weld joint integrity of the rotor must be proven before and after the over-speed test. All non-destructive testing of the fan wheel (both before and after over-speed testing) is to be done by Level II weld inspectors in accordance with SNT-TC-1A.



Safety Considerations


If you are considering being onsite to observe the over-speed test, you will want to consider the facility’s safety procedures as well. Best practices dictate that during the over-speed test, one designated manager must have complete responsibility for the safety considerations and the continuance of the test. The test manager is to constantly monitor all operating parameters and he/she is authorized to abort the test and bring the rotor back to the stopped position immediately if any parameters are found to exceed pre-established acceptable ranges. An emergency shutdown procedure must be developed and approved before testing begins. All parties associated with running the over-speed test must complete training that includes the emergency shutdown procedure.

Click to Enlarge


ID/FD fan rotor over speed testing is available to offer end users an increased fatigue life as well as a feeling of confidence in the design and fabrication of these process-critical fan rotors. The end user can be confident that the rotor has operated satisfactorily at speeds and stress levels that are significantly higher than it will ever be exposed to during actual operation. The manufacturer’s certified test report showing “rotor over-speed test approved” provides the documentation. The integrity of the rotor has been confirmed by rigorous non-destructive test inspections both before and after the over-speed testing. Documentation of these tests is included in the fan supply package. The fatigue life of the rotor has been increased significantly as a result of the over-speed testing. This has been documented by engineering calculations that are a part of the manufacturer’s scope of supply.



More Power Engineering Issue Articles



Power Engineerng Issue Archives



View Power Generation Articles on