Improvements to the 1/3 octave band heavy-hard impact prediction method

Golden Matthew 1 Pliteq Inc. 131 Royal Group Crescent Vaughan, ON, L4H1X9, Canada

Patzke Tim 2 Pliteq LTD. Compass House Vision Park Chivers Ways Histon Cambridge, CB24 9AD, UK

ABSTRACT Previously, the authors presented a 1/3 octave band heavy/hard impact prediction method based on force pulse measurements from a drop tower. Since then, we have been applying the method with a variety of success. Previous papers presented on this topic were based on above grade applications in North American buildings. This paper will be looking at a slab-on-grade construction in the United Kingdom. This paper will be focused on the accuracy of the method and possible modes of error. It is shown that improved methodologies to measure the force pulse increased the accuracy of the prediction. 1. INTRODUCTION

Over the last 8 years, there has been an increased interest in impact noise and vibration due to heavy-hard fitness sources. The authors along with other co-authors have presented many papers at conferences on measuring the noise and vibration in labs designed for ASTM E492 [1], measuring the actual impact force with drop towers [2,3,4], creating a prediction methodology based on the actual impact force and an in-situ measured transfer function [5,6], and finally predicting the one-third octave band sound and vibration levels using data obtained with a reference drop of a 7.26 kg (16 lb) spherical mass dropped from 1 m (39.4 in) onto a thin reference flooring sheet [7]. Both predictions methodologies proved successful below 100 Hz but were less successful at frequencies above 100 Hz. This paper will apply the author’s one-third octave band prediction method to a building in the United Kingdom and use that to examine possible sources or error and how to fix them.

mgolden@pliteq.com 1

tpatzke@pliteq.com 2

tt inter-ncie’” SCOTTISH evENT CAMPUS 3) O , o iLASGOW

2. DROP TOWER

In 2017, Pliteq finished building an automated drop tower, shown in Figure 1, to measure the impact force of heavy hard objects dropped onto fitness flooring. This drop tower is used for measure both single impacts and long-term fatigue performance. Two carriages are supported on low friction rails as to only allow one axis of motion. The upper carriage lifts the lower carriage to a selected height and releases it so that the lower carriage will impact the test specimen with close to free-fall like conditions. The lower carriage can be fitted with various load plates to adjust the mass of the lower carriage and the impact foot shape can be changed to simulate different types of weight impacts. A load cell measures the impact force pulse. Not shown in this photo, a large granite block with four load cells has been place directly below the lower carriage. This was done in an attempt to measure the force below the fitness flooring. The granite block was design to be large enough to avoid possible resonance modes in the frequency range of interest.

Upper Carriage

Force gauge

Impact Foot Lower Carriage

Test specimen

Fig. 1 - Heavy Weight Drop Tower at Rest.

This design has lead to three possible sources of systematic error. The first is that the lower carriage may be slowed down from friction on the rails. We have investigated the velocity of the lower carriage right before impact with displacement sensors and high-speed photography. It has proven difficult to measure the velocity with high enough precision and low enough bias to be certain the lower carriage is in free-fall. However, we have determined that this source of error seems to be well controlled. The second source of systemic error is that some of the mass of the upper carriage is below the load cell, in the impact foot, so the load cell does not experience all of the impact force. Since we can not impact the fitness flooring with the bare load cell, the error can not easily be eliminated, but it can be accounted for. Ideally, the impact foot would be massless. This would allow all of the mass to act on the upper side of the load cell. Since this is not possible we are measuring a reduce force at impact and a negative force while the carriage is rebounding. We have been exploring ways to compensate for this systematic error or to measure the impact force in other ways but have yet to find a superior solution. The mass of the impact foot also has resonances both internal to the foot and due to the mass of the impact foot hanging on sensor with the stiffness of the sensor creating a spring mass system. The third source of systematic error is likely the resonance mode due to the mass of the granite block on the stiffness of the sensors. This created a ringing decay after the impact foot has left contact with the fitness floor. These sensors are currently only used when the upper carriage is not used as the impacting source. As a last minute addition to this paper, a small scale force plate

was designed, built and tested which has produced superior results when measuring small steel shots on thin fitness flooring. To date, any ringing in the data after impact has been address with exponential windowing. The authors are aware of the signal processing issues this creates but the effects can be minimized.

3. ONE-THIRD OCTAVE BAND PREDICTION METHOD

Last year, Golden, et. al. [7] showed that data from the drop tower can be used along with in-situ measurements of lower mass impacts, a steel shot of 7.26 kg, onto a pretested sheet of GenieMat FIT08 used as a reference floor sheet. The method is summarized in Equation 1.

oov sig osz 00% = 091 & sat J itd Microphone | Vertical - roof center GenieMat FIT70 ve Measured 2 8 & & (eavloz a1) jan27 ainssaig punos 100 90 40 |. 30

L ′  I , F , max , DUT = L ′  I , F , max , Ref −( L dt , Ref − L dt , DUT )

(1)

L ′  I , F , max , DUT

Where is the in-situ measured fast-time-weighted maximum vibration level of the device under test, is the in-situ measured fast-time-weighted maximum vibration level of a reference sheet, is the overall impact force levels of the device under test via Pliteq’s drop tower, and is the overall impact force levels of the same reference sheet via Pliteq’s drop tower. This prediction method was very successful at predicting the sound and vibration that would be created by dropping 22.7 kg mass onto GenieMat FIT70 as shown in Figure 2. Note that the predictions were only accurate up to approximately 100Hz. Above that the prediction method over predicted the resulting sound and vibration.

L ′  I , F , max , Ref

L dt , DUT L dt,Ref

Microphone | Lateral - roof center GenieMat FIT70 100 90 8 iS 8 g (eavloz a1) jan27 ainssaig punos 40 |. 00s oov sig osz as sz guess ‘One Third Octave Band Freq (Hz) sig d 24 BE ot H stl 2E oi £9 30

os cov sis osz 00% ont & sag ‘Accelerometer | Vertical - roof wall GenieMat FIT70 120 60) 50 e 3 3 e (;5/uii! 21 ap) Janey uoneseja20y uoReiqIn

‘Accelerometer | Vertical - roof center . GenieMat FIT70 . 120 00s oov sig osz 00% = 091 & sat J 60) 50 2 8 &@ 8 Pe (;5/uii! 21 ap) Janey uoneseja20y uoReiqIn

Figure 2: Measured vs. predicted sound and vibration levels for GenieMat FIT70 at each measurement location.

‘Accelerometer | Lateral - roof center . GenieMat FIT70 . 120 00s oov sig osz 00% = 091 & sat J 60) 50 2 8 &@ 8 Pe (;5/uii! 21 ap) Janey uoneseja20y uoReiqIn

5HFHLYHU URRP

B2

3-13 3-12 3-11

3-M 3-L K

8420 7945

6250 8250 6471

5457 7563 687

5457 8250 5790

2685 3000 225

U

B2

2685 3000

U/S SL 1 0.035

F

'URSSRLQW

5685

FFL 7.350

F

A03-M 00

FFL 4.350

S

A03 -L00

SSL 4.125

L00

3-K 3-F 3-J

6979 5550 6500 6124

685 1819

K2

Business Unit 01

BU 1 -0001

292.6m² (3150ft²) GIA

PV8

'URSSRLQW

B2

2000

Awning Over GF

16087

Awning Over GF

@ Mezz Level

Awning Over GF

@ Mezz Level

Awning Over GF

@ Mezz Level

@ Mezz Level

Seating Area

1900

B2

eS 2 i

Drained Canopy Over

18250

SG

B2

B2

GF @ Mezz Level

Drained Canopy Over GF

@ Mezz Level

B2

003 G SI

B1

-

Seating Area

PV24

K2

8

15000

Figure 3: Plan and elevation view s of measurement location.

4. CURRENT SITE

This paper looked at predicting the soun d le vel in a r ec ei ver room on the sec ond floor while impacts were on a conc rete slab-on-grade on the flo or belo w , as shown in F igure 3. In all previous papers, the impact so u rce was o n a n ab ove grade fl oor an d the receiving l oc a t ions were either on the same slab or dir ectl y below the source. It was unknown if low ma s s im p acts would generate enough structur e-borne s o un d to be a udible on the floor above.

le DU RS

5. STEEL SHOT M EASUREMENTS

5.1 I n-s i tu Measur eme nts A possible standard for measuring impact noise from heavy-hard sources is currently being worked on in the ASTM International E33 Committee on Building and Environmental Acoustics in work item WK57850.[8] There is currently a debate as to what the appropriate mass and drop height should be for investigations such as this. To that end, three different mass steel shots (3 kg, 6 kg, 7.26 kg) were dropped from two different heights (0.5 m and 1 m) onto a thin reference sheet.

!

+,
& +,& +,
& +,& +,
& +,& #$+,- 

"#$%&'()* 
)




































!














      

Figure 4: In-situ impact sound pressure level as measured on the 2nd floor due to 6 different mass- height combinations.

The sound generated for each mass-height combination drop as well as the maximum background noise is shown in Figure 4. The steel shot impacts were above the background noise above 63 Hz. The general trend of the data is that the higher the mass and the higher the drop, the louder the structure-borne noise was on the 2nd floor. However the lowest mass-height combination, 3 kg at 0.5 m, seemed to slightly depart from this general trend.

5.2 Drop Tower Measurements The impact force levels on the reference sheet due to the same mass-height combinations as used above were measured with the sensors below the granite mounting block on the drop tower, shown in Figure 5 on the left. As with the in-situ measurements of the impact sound pressure levels, the general trend of the data is that the higher the mass and the higher the drop, the impact force level is higher. Note, that the frequency of roll-off does not change. Only the levels. At 400 Hz and above, the impact force levels are no longer smooth. As a late addition to this paper, a new lower mass force plate was designed, built and tested. This data is shown in Figure 5 on the right. Note that in general the force levels start to roll off at a higher frequency compared to those measure using the granite mounting block. Two of the mass-height combinations force levels roll off sooner than others. This was observed to be due to the rocking of the force plate. The new force plate will be improved and the data will be re-collected before the InterNoise conference. The analysis in the rest of this paper will use the force levels measured with the new lower mass force plate.

!

!





"#$% &  ' 

(

"#$% &  ' 

(













)*# )*
# )*# )*
# )*# )*
#

)*# )*
# )*# )*
# )*# )*
#















































!









!



























      

      

Figure 5: Impact force measured on the drop tower due to 6 different mass-height combinations.

6. IN-SITU - DROP TOWER DELTA STEEL SHOT

L ′  I , F , max , Ref L dt,Ref

The delta between the , in-situ fast max sound pressure level, and the , impact force level from the drop tower, was calculated to see if all mass-height combinations resulted in the same level difference. This delta is shown in Figure 6. The region below 63 Hz was shaded to emphasize the range where the in-situ fast max sound pressure levels were below the background noise as shown in Figure 4. From 100 Hz up to 250 Hz the deltas from 5 of the 6 mass- height combinations are nearly identical. The outlier, the 7.26 kg at 1 m mass-height combination, has known issues with the force plate measurement. The 3 kg at 0.5 m mass-height combination was also an outlier from 63 Hz to 100 Hz. This was noted in the field measurements in section 5.1.

This agreement is very reassuring. If there are any systematic errors in the in-situ and drop tower measured steel shot data, they are at least linear and repeatable. The deltas do not agree above 200 Hz. This appears to be an issue with the data collection in the field. However, investigations are continuing.

!

%&' %&
' %&' %&
' (%&' (%&
'





" #$  #
































!














      

Figure 6: Delta level difference between in-situ measured fast max sound pressure level and drop tower measured force level.

7 DROP TOWER DELTA OF DROP TOWER TESTS The delta between the , impact force level due to steel shots onto the reference sheet from the drop tower, and the , impact force levels of the device under test from the drop tower was calculated to examine any possible systematic errors, Figure 7. There are clear inflection points in the data at 250 Hz, 200 Hz and 125 Hz for the deltas between the steel shots on the reference sheet and the 22.7 kg dropped onto the three different fitness flooring respectively. This suggests there is a possible issue with the data collection of the force levels measured with the drop tower at those frequencies. Further study will be needed.

L dt,Ref

L dt , DUT



  
 

 







%&' %&
' %&' %&
' (%&' (%&
'

%&' %&
' %&' %&
' (%&' (%&
'

%&' %&
' %&' %&
' (%&' (%&
'







! "  

! "  

! "  



































#



$





#



$





#



$












#












#












#














































      

      

      

Figure 7: Delta level difference between tower measured force level between the steel shots on the reference sheet and the desired mass and device under test.

8 PREDICTION Using the data from the new lower mass force plate, data from the load cell on the lower carriage of the drop tower, and the data collected in the field, a prediction of was calculated using the data for each mass-height combination of steel shot drops. This prediction was then compared to the field measured values of and the max background noise measured on site. These predictions and measurements are shown in Figure 8 for all three fitness flooring tested. The predicted levels using data from each mass-height combination steel shot drop were within a few dB of the measured sound pressure levels.

L dt,Ref L dt , DUT L ′  I , F , max , Ref L ′  I , F , max , DUT

L ′  I , F , max , DUT





 


  
 

     
       
       
   

     
       
       
   

     
       
       
   




























































































































Figure 9: Predictions of sound pressure level due to a steel shot drop vs measured levels and maximum background noise levels.

9 CONCLUSTION In conclusion the heavy-hard impact noise level prediction method the authors have previously published continues to be valid even when the impact source location is on a slab-on- grade and the receiver location on a floor above. The prediction accuracy was increased at frequencies above 100 Hz and up to 250 Hz by improving the methodology of the measurement of force levels using a low mass force plate. Future work will be performed to further improve the force level measurement methodologies and will be shared at the InterNoise conference.

9 AKNOWLEDGEMENTS The authors would like to thank several of their co-workers who have helped collect data including last minute measurements on the lower mass force plate. These individuals include, Hamza Bakheet, Faiz Musafere, Navid Shariati, and Paul Gartenburg. Their help was greatly appreciated.

11. REFERENCES

1. Paul Gartenburg and Paul Downey, “Comparing various fitness flooring assemblies using heavy/

soft and heavy/hard impact”, INTER-NOISE 2015, San Francisco, CA, US. 2. Paul Gartenburg and Matt Golden, “Comparing force impulses for various fitness flooring

assemblies”, NOISE-CON 2017, Grand Rapids, MI US. 3. Matthew V. Golden, “ Drop Tower Testing—The next step in heavy weight impact testing”,

The Journal of the Acoustical Society of America 143, 1726 (2018); Presentation

4. Paul Gartenburg and Tony Nash, “Fitness Noise Control Products – Should we be Measuring

System or Material Properties?”, NOISE-CON 2016, Providence, RI, US. 5. Matthew V. Golden and Paul Gartenburg, “Prediction of Heavy Weight Drops on Resilient Sports

Floors in Existing Buildings”, INTER-NOISE 2018, Chicago, IL, US. 6. Matthew V. Golden and Faiz Musafere, “Continuing Prediction of Heavy/Hard Impacts on

Resilient Sports Floors in Existing Buildings” International Commission for Acoustics 2019, Aachen, Germany. 7. Matthew V. Golden, John LoVerde, Rich H. Silva, Wayland Dong and Samantha Rawlings,

“Prediction of one-third octave band sound and vibration levels from heavy-hard impacts”, INTER-NOISE 2021, Washington, DC, US, Virtual 8. ASTM Work Item WK57850, “ New Guide for Field Measurement of the Reduction of Impact

Sound from Heavy Impact Sources When Using Floor Coverings”, ASTM International, West Conshohocken, PA, 2003