Random Vibration Notching

What Is It?

Notching is the practice of reducing the control PSD at specific frequencies during a random vibration qualification test to limit the response at a critical location. The notch reduces the input at a resonance where the test specification would otherwise produce an unrealistic over-test. Notching is based on analysis predictions of interface loads or response limits that should not be exceeded.

Why It Matters

A qualification test specification is often an envelope of the expected environment — it is designed to be conservative. At some frequencies, particularly at structural resonances, the envelope PSD may be significantly higher than the actual service environment. Testing at the envelope level at these frequencies can produce unrealistic over-test — response levels far exceeding anything the article would see in service. This can cause unrealistic failures that are not representative of service conditions. Notching prevents this.

Notching prevents unrealistic over-test at resonances where the qualification PSD exceeds the actual service environment. Without notching, the test may produce failures that would never occur in service — misleading and costly.

Purpose

The purpose of notching is to limit the test article response to a realistic level at frequencies where the test specification is over-conservative. This ensures the test is a meaningful representation of the service environment — not an unrealistic over-test. The notch allows the test to demonstrate survival at representative levels without destroying the article at unrepresentative levels.

Interface Loads

Notching is often based on interface loads — the forces and moments at the mounting interface. The analysis predicts the interface loads for the service environment. During the test, the interface loads are monitored (using force gauges or derived from acceleration and mass). If the test interface load exceeds the predicted service interface load at a frequency, the control PSD is notched (reduced) at that frequency to bring the interface load back to the service level.

Notching based on interface loads:

1. Analysis predicts service interface loads:
   F_service(f)  from service PSD

2. Test monitors interface loads:
   F_test(f)  from force gauges or
   F_test ≈ m × a_test(f)

3. If F_test(f) > F_service(f) × safety_factor:
   → Reduce control PSD at frequency f (notch)
   → Until F_test(f) ≤ F_service(f) × safety_factor

This limits the test to realistic interface loads.

Response Limits

Notching can also be based on response limits — acceleration, displacement or stress at a critical location. The analysis predicts the maximum expected response for the service environment. During the test, if the response exceeds this limit, the control PSD is notched. Response-based notching is used when interface load monitoring is not available or when a specific response quantity is the limiting factor.

• Response-based notching — limit acceleration, displacement or stress at a critical location
• Analysis predicts the maximum expected service response
• If test response exceeds the limit, control PSD is notched
• Used when interface load monitoring is not available or when response is the limiting factor

Avoiding Unrealistic Over-Test

The key engineering judgement in notching is determining what constitutes an unrealistic over-test. The test specification is an envelope — it is intended to be conservative. But if the envelope is far above the service environment at a resonance, the test is no longer a meaningful representation of the service condition. Notching brings the test back to a representative level. The basis for the notch — the service environment prediction — must be well-documented and defensible.

Notching must be based on a defensible analysis of the service environment. The notch limits the test to realistic levels. The basis — the predicted service loads or response — must be documented and justified. Do not notch without a clear engineering basis.

Analysis/Test Correlation

Notching requires accurate pre-test analysis. The notch levels are based on the predicted service response or interface loads. If the analysis is wrong, the notch levels are wrong — too high (insufficient notching, over-test) or too low (excessive notching, under-test). The pre-test analysis should be correlated with the test article's actual modal properties (from a pre-test modal survey) to ensure the predictions are accurate.

• Notching requires accurate pre-test analysis
• Notch levels based on predicted service response or interface loads
• If analysis is wrong, notch levels are wrong
• Correlate pre-test analysis with actual modal properties (modal survey)
• Document the basis for all notch decisions

Implementation

Notching is implemented in the shaker controller. The controller monitors the limiting channel (interface load or response) and adjusts the control PSD to keep the limiting channel within the specified limit. The notch is visible in the control PSD as a dip at the notched frequency. The depth and width of the notch depend on the limit and the article dynamics. The controller should be set up to notch smoothly — abrupt notches can cause control instability.

Key Takeaways

• Notching reduces the control PSD at resonances to prevent unrealistic over-test
• Based on interface loads or response limits from analysis of the service environment
• Ensures the test is a meaningful representation of service, not an unrealistic over-test
• Requires accurate pre-test analysis — wrong analysis produces wrong notch levels
• The basis for notching must be well-documented and defensible

Engineering judgement — what can change the conclusion

For Random Vibration Notching, the harmonised review should concentrate on the interaction between control strategy, sensor location and local structural response. A control channel can protect the fixture while allowing another location to exceed the intended environment. The engineering value comes from identifying the assumptions that can move the governing margin or failure mode, then testing those assumptions deliberately rather than adding complexity indiscriminately. Where simplified and high-fidelity methods coexist, the simpler method should be used as an independent trend or magnitude check so that agreement is based on physics rather than shared modelling assumptions.