The “Smart Hit Selection” (SHS) automatically selects the best impact measurements to use for the Frequency Response Function (FRF) calculation. It eliminates double hits and overloads while selecting the measurements that yield the highest coherence or consistency.
With “Smart Hit Selection”, the time data of the individual hits of each impact location are recorded. More impacts are acquired than needed so that the “Smart Hit Selection” can choose the best measurements to use for the final averaged Frequency Response Function (FRF).
How does it help? Consider an example case. During a test, an operator will impact repeatedly close to the same location. Due to human variance, the exact impact location can be slightly off each time (Figure 1).
Figure 1: Example of human error on an FRF during impact testing (left). The hammer hits are at slightly different locations (red dots, middle). This causes small shifts in the FRF measurements (right).
The misalignment of the hits causes slight shifts which affect the average FRF. These shifts due to the misalignment of hits will adversely affect analysis performed on the resulting FRFs. For example, force estimations that use FRF inversion can have spurious peaks as a result.
These issues, as well as overloads and double hits, are prevented using Smart Hit Selection.
Benefits of “Smart Hit Selection” include:
Operator can focus on testing, the algorithm will discard bad hits
Less need to manually scan all data after the test
Yields highest quality FRF data from the available impacts
The settings used for selecting the impact measurements used in the FRF calculation can be adjusted by the user. This article explains the “Smart Hit Selection” process and the influence of different settings: 1. What is Smart Hit Selection (SHS) 2. How Does Smart Hit Selection Work? 2.1 Number of Hits 2.2 Overloads and Double Hits 2.3 Peak Force Limit (Optional) 2.4 Pairwise Similarity versus Pairwise Coherence 2.5 Target Value 2.6 Taking Measurements 3. Reviewing and Adjusting Smart Hit Selection (SHS) 4. Saving the FRF
1. What is Smart Hit Selection (SHS)?
Smart Hit Selection (SHS) is a feature that uses the best impacts and discards outliers to give the operator the cleanest set of frequency response functions (FRFs). Smart selection acts as an assistant during data collection and scans through the data to discard bad hits.
A flow chart of the Smart Hit Selection process is shown in Figure 2 below.
Figure 2: Smart Hit Selection process flow chart.
As repeated impact measurements are performed at a given location:
Overloads and double hits are flagged and discarded
Targeted force levels are enforced (optional)
Impact hits for averaging are selected based on their consistency or coherence to produce the best FRF possible from the individual hits
2. How Does Smart Hit Selection (SHS) Work?
Smart Hit Selection (SHS) begins after an impact is performed and the resulting force and acceleration is measured. It is part of the Measure step of Simcenter Testlab Neo Impact Acquisition (Figure 3).
After the accelerometer and modal impact hammer is entered in the Instrumentation tab and Hammer setup, the following settings are part of the Smart Hit Selection (SHS):
2.1 Number of Hits
One of the first things to define in the Smart Hit Selection process is how many impact hits to perform at a given location. More hits will be acquired than actually used in the final averaged FRF data. For example, 10 hits total may be specified to yield an average consisting of between 3 to 6 hits.
There are two places where the number of hits is specified:
Automatic Impact Selection: Specify “Minimum impacts” and “Maximum impacts” for calculating the average FRF
Campaign Settings: Specify the total number of hits to acquire with the “Target impacts per hammer”
An example measurement with corresponding number of hits settings is shown in Figure 4.
Figure 4: Result of Smart Hit Selection with “Minimum impacts” set to 3 and “Maximum impacts” set to 6. The number of hits used to calculate the average will be between 3 (minimum) and 6 (maximum) hits. Ten hits total are recorded because “Target impacts per hammer” is set to 10.
These settings mean that total number of hits used to calculate the averaged FRF are between the “Minimum impacts” and “Maximum impacts” setting. The total number of hits that is recorded will be set by the “Target impacts per hammer” (which should be the same as or higher than “Maximum impacts”). The software will select FRFs from the total number of hits to get the best FRF based on all specified criteria.
The similarity/correlation percentage is only shown if the minimal number of impacts is reached. If there are less hits than the minimal number of impacts (perhaps due to double hit rejections for example), no percentage is calculated.
Note: After changing settings in “Automatic Impact Selection”, just press “Close” on the dialog box to use them. The “Save” in automatic dialog is to create a preset list of criteria if desired.
2.2 Overloads and Double Hits
The first checks performed in SHS after an impact hammer hit are:
Overloads: Overloads can cause distortions in the frequency domain, and not capture the full amplitude of the signal, so these must be discarded.
Double Hits: Double hits with the measurement time can greatly affect the data.
Double hits and overloads are automatically excluded by the Smart Hit Selection algorithm as shown in Figure 5.
Figure 5: Out of eight total hits shown in the Impact Acquisition screen, four total are accepted. Two were excluded due to double hits and two were excluded due to overloads.
If the “Auto reject with overload” checkbox is ON then measurements with overloads will not show up at all. The “Auto reject with overload” is found in the “Measurement advanced” settings of the Campaign settings.
2.3 Peak Force Limit (Optional)
If activated by the user, the Smart Hit Selection checks if the force from each hit is outside the target level specified. The force check is helpful when working with non-linear structures whose response changes as a function of force level. The force limit check is optional.
Figure 6 shows one example of how inconsistent hammer hit force levels can alter FRF and coherence results.
Figure 6: A comparison of the coherence and averaged FRF with peak force limit selection on (yellow) and peak force limit selection off (blue). With peak force limit selection on, the coherence is improved, especially in the noisier regions of the FRF.
The difference between FRFs measured with and without a targeted force will vary depending on the linearity of the test structure.
To do the force check, a “Target force level” can be set in the “Hammer setup” tab in the Impact. This will only keep hits that are within a certain range of the target.
In Figure 7 below, the target force level is set to four Newtons.
Figure 7: In the Hammer Setup tab, click on the “More” to set the “Target force level”.
A range around the target force level can also be set. This set in the “Automatic Impact Selection” of the Measure tab of Impact. Click on the “More” to be able to enter a “Target force level”.
An upper threshold and lower threshold can be individually set as shown in Figure 8.
Figure 8: Upper and Lower limits for force threshold are set in the “Automatic impact selection” menu (which looks like a wand). The peak force levels are colored either orange (potential exclusion) or red (excluded). This is part of the Measure tab of Impact.
In this example, the upper and lower thresholds were both set to 20%, so the range for each hammer is plus or minus 20% of the target. During the measurement, the peak force level will be colored based on the force target level and the threshold:
Black: Within the “Target force level” and specified deviation range.
Orange: If the “Auto exclude force level” option is NOT ON in the “Automatic impact selection” menu, force levels outside the force deviation limits will be colored orange for potential exclusion. It will be up to the user to click in the “Selected” column to manually exclude the measurement.
Red: If the “Auto exclude force level” option is ON in the “Automatic impact selection” menu, force levels outside the force deviation limits will be colored red and excluded from the average.
The upper and lower thresholds that can be adjusted during impacts at any time of the measurement. If the threshold percentage is 100% (the default), no hits will be flagged or rejected based on achieving the target force level. In other words, a 100% threshold essentially does nothing, as all hits will pass the force threshold check.
After checking overloads, double hits, and force deviation checks, the FRFs of the individual hits are evaluated for consistency or coherence in the next step of the Smart Hit Selection process. This is described in the next section.
2.4 Pairwise Similarity versus Pairwise Coherence
Across the impact measurement hits, a comparison of the consistency and coherence of the individual hits is used to select which hits to include in the averaged FRF.
Under the “Automatic Impact Selection” menu (with the magic wand icon) either “Pairwise similarity” or “Pairwise coherence” can be selected as shown in Figure 9.
Figure 9: Select either “Pairwise coherence” or “Pairwise similarity” as the method in Smart hit selection of Simcenter Testlab Neo Impact Acquisition.
The two methods work in a similar manner, but use a different basis for comparing the acquired FRFs over multiple hits:
Pairwise Similarity: This compares each hit’s instantaneous FRF to one another and computes the similarity between each pair. It evaluates how similar the 2D curve (amplitude versus frequency) shape of the two FRFs are to each other.
Pairwise Coherence: This compares the coherence of the final results of FRFs to create a statistical evaluation. Coherence is a measure of the repeatability of the FRF from hit to hit. Coherence reflects how much of the output acceleration is due to the input force.
A target value is set for the pairwise similarity or pairwise coherence.
2.5 Target Value
“Target Value” is another parameter that can be set under the “Automatic Impact Selection” menu (with the magic wand icon). The target value can be set between 1% and 100%. Targeting as close to 100% as possible is desired. The menu is shown in Figure 10.
Figure 10: With a “Target Value” for the pairwise coherence of 98%, Smart Hit Selection found six hits (the minimum number of hits required) that exceeded the targeted coherence with a value of 99%.
Whether the Method is set to “Pairwise coherence” or “Pairwise similarity”, the “Target Value” can be set the same. Setting the “Target Value” to 100% always yields best result possible results.
3. Reviewing and Adjusting Smart Hit Selection (SHS)
The individual hits can be reviewed in the displays. Highlighting the row of one of the measurements shows it overlaid with the averaged data as shown in Figure 11.
Figure 11: Individual hits can be viewed against the averaged data by highlighting individual rows in the measurement area.
If desired, the user can manually override the automatic selections made by the Smart Hit Selection algorithm. Clicking in the “Selected” cell of the individual hit will toggle between being included in the average or not as shown in Figure 12.
Figure 12: User can override the automatic Smart Hit selections by clicking in the “Selected” cell. Cells with the "Person" symbol indicate manual user selections while cells with the "Gear" symbol indicate automated selection.
After manually selecting/deselecting, the “Mode” cell symbol changes from a “Gear” (indicating automatic selection) to a “Person” (indicating manual selection).
Across the upper right of the Measure are “Smart Hit Selection” (SHS) buttons. These buttons can be used to change the settings and recalculate the FRFs (Figure 13).
Figure 13: Action buttons in the Smart Hit Selection.
The action buttons (from left to right) perform the following actions:
First Button (orange): Adds a DOF (Degree of Freedom). A DOF is the name of a measurement (combination of Point name and Direction). This button would be used to initiate a new measurement.
Second Button (Yellow): Change settings of Smart Hit Selection.
Third Button (Blue): Recalculate the FRFs after changing the settings. Run on all impacts, accelerometers, and hammers for every Degree of Freedom (DOF). If data and DOFs are large, then this method could take a while to calculate.
Fourth Button (Red): Recalculate the FRFs after changing the settings for only the selected excitation DOF.
Fifth Button (Green): Removes all manual user selections of hits and returns to fully automatic selections.
Sixth Button (Purple): Refreshes campaign.
Once any final adjustments have been made to the impact hit selections, the FRF can be saved.
2.6 Taking Measurements
After the settings are in place, press the "Start" button in the lower left of the screen. After each hit a new measurement will appear in the list.
After each hit, the status will update in the lower left of the screen (Figure 14).
Figure 14: Four possible status after each hit.
The status shown after each hit can be one of four possible values:
Overload: Hit rejected due to overload.
Double Hit: Rejected due to double hit.
Deviation: Hit rejected due to deviation compared to FRF from other impacts.
OK: Measurement accepted. This status could change to rejected if the FRF is ultimately inconsistent with the other FRFs as they are being collected.
If the minimal/maximal hits required for the average FRF was three, and the fourth hit was more consistent that the previously accepted hits, the accepted status of one of the three previously collected hits can become rejected so the latest fourth hit becomes accepted.
4. Saving the FRF
To save the FRF based on the selected hits, press the “Save” button at the bottom of the screen as shown in Figure 15.
Figure 15: Press the Save button to store the FRF. A campaign named can be assigned at the same time.
The individual time histories of the hits are stored in the Measurement Run.
The FRFs are stored in a separate file structure in the project called a “Campaign”. The name of the Campaign can be changed in the box at the bottom of the Measure screen.
If a new measurement is started and the “Save” button was not pressed, no FRF is stored in the project. To save the FRF after the fact, restored the “Archived Settings” of the run (from the Desktop tab) and then go back to Measure in Impact and re-save it.
