CONTENT SLIDING RESEARCH AT THE UNIVERITY OF CANTERBURY

by Gregory MacRae, Trevor Yeow, Rajesh Dhakal

Third International Workshop on Seismic Performance of Non-Structural Elements (SPONSE)

31st March 2016, University of Canterbury (UC), Ilam,Christchurch, New Zealand

Sponsored by the UC Quake Centre

Contents damage can contribute significantly to loss (in dollars) as measured in terms of 3Ds (damage, death/injury, and the resulting downtime). While methods to restrain contents may be effective, many contents in buildings are not appropriately restrained and it is possible that significant movement can occur, which results in loss. A significant volume of research on building contents are focused on the rocking performance of rigid body structures. Recent studies at UC instead investigated the sliding response of unrestrained contents in structures. The studies performed, and findings to date, are summarized below.

Demand Estimation:

The demands on contents in a building depend on the location of the contents over the building height. Since contents are considered to be acceleration sensitive components, the acceleration profile over the height, as well as the drifts over the height, were investigated using shear and flexural type structures by Dantanarayana et al. (2011-2012). This approach used was based on that of Sadashiva et al. (2009) where simple stick models, calibrated to the behaviour of more complex structures, were used to represent multi-storey structural behaviour. An algorithm was written to automatically (i) design a range of code compliant structures (i.e. varying number of stories, periods, design drifts, and lateral force reduction factors), (ii) perform inelastic structural response history analyses using a suite of ground motion records, and (iii) extract and plot specific results. Using this approach, the behaviour of a large range ofstructural systems was evaluated. The study indicated that accelerations were affected by the structural form, ductility demand, andhysteretic characteristics. Also, it was shown that the NZS1170.5 acceleration profile for multi-storey structures often conservatively estimated the acceleration demands.

Numerical Studies related to Sliding:

The first numerical study of sliding contents at the UC was by a final year undergraduate student, English (2011, 2012), who:

a)Used springs exhibiting elastoplastic behaviour to mimics Amonton’s and Coulomb’s dry frictionlaws for modelling content movementin structural analysis software.

b)Showed that the three key parameters affecting the movement of contents, all with the same friction coefficient (μ), within single-storey buildings for a particular record are (i) the natural period of the structure, T, (ii) the structural strength reduction factor, R, and (iii) the hysteretic shape.

c)Showed that the ratio of AFT/(μg) (where AFT is the peak total floor acceleration,and g is acceleration due to gravity) needed to be greater than 1.0to initiate sliding.

d)Showed that a content-to-structural mass ratio of 1:1000 had almost no influence on the structural response.

e)Explained the sliding mechanicsof contents subjected to sinusoidal floor excitations.

f)Showed that sliding was greatest in elastic structures and decreased with increasing structural R.

g)Showed that single-storey structures with hysteresis curves with high hysteretic energy dissipation had lower contents movement than those with lesser energy dissipation and the same backbone. This was consistent with the motocycle analogy by MacRae (2010) that structures with hysteresis loops giving a sudden increase in stiffness at high structural velocity, such a flag-shaped loops for rocking structures, will have more damage to contents and people inside the building than inelastically responding bilinear buildings (e.g. steel frames) with the same backbone hysteresis curve.

h)Showed sharp corners in the hysteresis loop caused high frequency ringing in the response.

i)Developed the first contents sliding spectra for both (i) impulse excitation and (ii) a suite of earthquake records, where the structure oscillated in a sinusoidal manner after the initial excitation.

The nextphase of the research, conducted by postdoctoral candidate Lin during 2012-2103, verified the work by English (Linet al, 2012).In addition, aclosed-form analytical solution was derived for the first sliding excursion (i.e. back and forth motion) displacement of contents on a single-storey buildingsubjected to baseimpulse loading(Lin et al, 2013). This was a function of the building’s natural period, itsdamping ratio, the spectral acceleration at the period considered, , and the times that sliding started and stopped. There were two issues with this:

a)While the first sliding excursion’s displacement was shown to be equal to the maximum sliding displacement when g/Sa 0.35, the maximum sliding displacement was greater than that of the first sliding excursion for g/Sa < 0.35. Thus, this equation is non-conservative for predicting the maximum sliding displacement when g/Sa < 0.35.

b)The equation was difficult to solve directly because the times of starting and stopping of sliding have no closed-form solutions and are difficult to obtain.

However, when sliding displacements from earthquake records were computed, it was shown that the first sliding excursion’s displacement computed as a result of base impulse loading matched the median maximum sliding displacementwell. This was consistent with previous studieson other topics such as building torsion (e.g. MacRae et al., 2008). Lin (2015) also developed an empirical expression to consider variation in thesliding displacement, and provided a design example for sliding displacement assessment.

Another undergraduate study on content sliding was performed by Cain and Rileyin 2013 (Cain et al. (2013), Riley et al. (2014)). They:

a)Verified the study by English (2011) and Lin (2012).

b)Showed that greater structural damping could reduce the response of contents on single storey structures. For example, changing the structural damping from 0% to 5% reduced the contents movement by more than 50%in some cases.

c)Considered cases when contents were beside a wall or other barrier which means that they could not slide past the wall or barrier. This is referred to as one-way sliding or obstructed sliding. Momentum was considered to be conserved on impact in the initial model used. They explain the impact effect for impulse analysis.

d)Showed that considering a lower μ(i.e. greater AFT/(μg)) resulted in greater contents movement in general, thoughthe change in contents movement was less for one-way (obstructed) sliding than for two-way (or unobstructed) sliding.

e)Demonstrated that contents displacements from obstructed slidingwas at times almost an order of magnitude greater thanif unobstructed sliding is permitted, thoughthe median ratio between obstructed and unobstructed sliding displacements considering a suite of records was generally not more than two.

f)Found that when  = 0.25, with the SAC LA10in50 records, the 50th percentile one-way contents movement for a design level suite of 20 records was 0.6m for one-way sliding at some structural periods, and 0.3m for two way sliding. The 84th percentile one-way contents movement for these records was 3m for some structural periods for one-way sliding, and 0.8m for two-way sliding.

g)Identified that sliding displacements were sensitive to peak total floor velocity (VFT) and developed a parameter containing total VFT and AFT terms to best relate to contents movement.

Yeow, a PhD student who had been involved with all previous studies, conducted further analyses and compared methods to assess sliding capacity. Considering a wider range of cases, he further emphasized Cain and Riley’s findings thatVFT, as well as AFT, wereneeded to get a good estimate of the likely sliding movement. He used this concept to improve upon Lin’s work and produce a more efficient and sufficient parametric equation that does not require the times at which sliding initiates or terminates to be obtained first (Yeow et al. JAE, 2016).

Yeowincorporated this parametric equationinto building seismic performance assessments (Yeow et al. 2015)to estimate the likely impact of content sliding on loss, and to also investigate how a building’sstiffness may affect the extent of content movement. While building contents are traditionally considered to be acceleration sensitive, it was shown that higher accelerations in stiffer buildingsoften did not result in greater contents displacements because the velocities were relatively low. This resulted in stiffer buildings having lower content sliding-related losses in general. Stiffer buildingsare generally more likely to result in lower losses than more flexible buildings during strong earthquake shaking if drift-damage, need for full-replacement, chance of collapse, and possibility of pounding against other buildings were considered. While much empirical evidence supports the concept that stiffer buildings have lower damage in general, this work provides a theoretical anchor for this observation.

Content movement was then considered in a newly developed injury prediction framework by Yeow (2016) for use in loss estimation. He found that current methodologies either employ the use of predefined injury rates, or predict the occurrence of peoplebeing struck by toppling contents without considering its consequence. Also, while models do existto predicting the occurrence of people falling, and the severity and costsof an injury based on impact conditions,these have not been incorporated into seismic risk frameworks.Yeow’s framework considers the spatial distribution of peopleto account for non-uniform occupant densities and body position, and predictsimpact conditions due to people being struck by contents or falling to estimate injury severity and subsequent costs. Case-study injury numbers were consistent with injury severity data from the 1994 Northridge and 2010-11 Canterbury events. Simplifying the spatial distribution model,or not considering injuries due to people falling or being struck by contents, resulted in injury rates differing by 30 to 80%; demonstratingthe need to consider these aspectsfor producing outputs consistent with historical data.In addition, cost-benefit analyses of anchoring contents was performed, where the reduction in expected costs after 50 years of the building being in service was twenty times that of the implementation costs, highlighting that it is economically justifiable and sensible to restrain contents.

Experimental Testing:

A number of experimental studies have been undertaken related to sliding of realistic office furniture on realistic flooring surfaces by Yeow (2016). Several types of test were conducted:

a)Pull tests. Here it was shown that Amonton’s and Coulomb’s dry frictionlawsare a good approximation of the actual sliding hysteretic behaviour.

b)Dynamic 2D shaking table tests of unobstructed contents subjected tosinusoidal excitation. Comparisons of these tests’ findings with analytical approachesbased on Amonton’s and Coulomb’s dry frictionlawsshowed that analyses can provide a good prediction of content sliding response(Yeow, NZSEE2014), which validates the modelling approaches adopted in all analytical studies described previously.

c)Dynamic 2D shaking table tests of obstructed contents subjected tosinusoidal excitation. These tests indicated that obstructedsliding displacements were approximately 3 times that for the unobstructed sliding tests (Yeow, 10NCEE 2014).

Additional Related Studies:

A study related to contents behaviour was related to conducted by a visiting undergraduate student Bruyere (2014) and it was supervised by MacRae and Yeow. This study focused on modelling the fall response of bricks and contents. This was modelled in three steps; (i) sliding or rocking response, (ii) projectile response after the content dislodges or falls from the building, and (iii) bounce response. Information from such studies may be useful to estimate the magnitude of shaking from bricks or other items which are found lying a specific distance from a damaged building. It is also useful for those who wish to establish cordons around a structure as a precaution against falling items due to shaking from an earthquake or its aftershocks, or to model the effect of small items (i.e. books) falling off shelves inside the building.

A further study was conducted by a visiting researcherIihoshi (Iihoshi et al. 2014, 2015) on the viability of allowing slabs to slide in single storey structures to reduce structural demands. This case is different from that considering contents alone as the contents’ weight is large enough to affect the structure’s response. A special element in OpenSEES was found to consider this effect appropriately. It was shown that for long period structures, the frame forces could be considerably reduced by allowing the frame to slide, and the total displacements of the slab on the frame were similar to those of a fully connected slab in an elastically responding frame. The implication of this is that frames could be more economically designed and may remain elastic and undamaged under seismic shaking. Empirical equations were developed to determine the average peak sliding displacement for variousperiods and mass ratios as a function of friction factor.

Acknowledgements:

The authors are grateful to all people who have contributed to these studies. Their names are listed in the reference material. They also wish to acknowledge the Quake Centre for coordinating this workshop.

Reference Material:

  1. Bruyere, R., Yeow T. Z., and MacRae G. A. "Beware of Falling Contents - Modelling for Seismic Regions", Research Paper, University of Canterbury, 2014.
  2. Cain, E. S. and Riley-Smith, H. “Building Contents’ Sliding Demands in Elastic, Single and Multi Degree of Freedom Structures”, Final Year Project, 2013, Dept. of Civil and Natural Resources Engineering, University of Canterbury. Supervised by MacRae, G.A., Dhakal, R.P., Yeow, T.Z.
  3. Dantanarayana H., “Quantifying Building Engineering Demand Parameters under Seismic Loading”, supervised by G.A. MacRae, R. Dhakal and T.Z. Yeow, Final Year Project, 2011, Dept. of Civil and Natural Resources Engineering, University of Canterbury.
  4. Dantanarayana H., MacRae G. A., Dhakal R. P., Yeow T. Z. "Quantifying building engineering demand parameters in seismic events", New Zealand Society of Earthquake Engineering Conference, Christchurch, 13-15 April 2012. Paper 50.
  5. Dantanarayana H.N., MacRae G.A., Dhakal R. P., Yeow T.Z. & S.R. Uma, "Quantifying Building Engineering Demand Parameters in Seismic Events", 14 World Conference on Earthquake Engineering, Lisbon, Portugal, August 2012. Paper number 3424. Poster presentation.
  6. English R., "Hysteretic Influence on Earthquake Induced Sliding Damage of Contents", Undergraduate Project, Supervised by MacRae G.A, Yeow T. Z, and Dhakal R., University of Canterbury, Christchurch, 2011.
  7. English, R., MacRae, G.A. and Dhakal, R.P. (2012) Hysteretic Influence on Earthquake Induced Sliding Damage of Contents. Christchurch, New Zealand: New Zealand Society for Earthquake Engineering Annual Conference (NZSEE2012), 13-15 Apr 2012. Paper 56.
  8. Iihoshi C., MacRae G.A., Rodgers G.W., Chase J.G., Viability of Slab Sliding System for Single Story Structure, International Conference on Earthquake Engineering (ICEE 2014), Barcelona, Spain, February, 27-28, 2014. 14SP020066.
  9. Iihoshi C., MacRae G.A., Rodgers G.W., Chase J.G., Viability of Slab Sliding System for Single Story Structure, World Academy of Science, Engineering and Technology, International Journal of Mechanical, Industrial Science and Engineering, 8(2), 2014. Paper #129.
  10. Iihoshi C., MacRae G.A., Rodgers G.W. and Chase J.G., Dynamic Behaviour of a Slab Sliding System, 14th International Conference on Structural and Geotechnical Engineering, 2015, ICSGE 14, Cairo, Egypt, 2015.
  11. Lin S-L, MacRae G.A, English R., Yeow T. Z, Dhakal R. P. "Contents sliding response spectra", New Zealand Society of Earthquake Engineering Conference, Christchurch, 13-15 April 2012. Paper 63.
  12. Lin S.L., MacRae G.A., Dhakal R.P., & Yeow T. Z., "Contents Sliding Response Spectra", 14 World Conference on Earthquake Engineering, Lisbon, Portugal, August 2012. Paper number 3601. Poster presentation.
  13. Lin S.L., MacRae G.A., Dhakal R.P., Yeow T.Z., "Building Contents Sliding during Earthquakes", New Zealand Society of Earthquake Engineering Conference, Wellington, 26-28 April 2013. Paper P25.
  14. Lin S-L, MacRae G.A, English R., Yeow T. Z, Dhakal R. P., 2015, "Building Contents Sliding demands in Elastically Responding Structures", Engineering Structures, 86, pp. 182-191 .
  15. MacRae, G. A.; Au, E.; Pettinga, D.; Deam, B.; Sadashiva, V., (2008) A Simple Design Method for Yielding Structures Subject to Torsion, 14th World Conference on Earthquake Engineering, Beijing, China, 12-17 October, Paper ID: 05-01-0514.
  16. MacRae G. A., 2010. "Some Steel Seismic Research Issues", in Proceedings of the Steel Structures Workshop 2010, Research Directions for Steel Structures, compiled by MacRae G. A. and Clifton G. C., University of Canterbury, 13-14 April.
  17. Riley-Smith H., Cain E.S., Yeow T. Z., MacRae G. A. and Dhakal R., "Building Content Sliding Demand - Analytical Studies Of Contents In Elastic, MDOF Structures", New Zealand Society of Earthquake Engineering Conference, Aotea Centre, Auckland, 21-23 March 2014. Paper P36.
  18. Sadashiva, V.K., MacRae, G.A. and Deam, B.L. (2009) Determination of structural irregularity limits - mass irregularity example. Bulletin of the New Zealand Society for Earthquake Engineering, 42, 4, 288-301.
  19. Yeow T.Z., MacRae G.A., Dhakal R.P., Bradley B.A., "Probabilistic seismic indoor injury estimation", New Zealand Society of Earthquake Engineering Conference, Wellington, 26-28 April 2013. Paper 6.
  20. Yeow T, MacRae G, Dhakal R. & Bradley B. "Preliminary Experimental Verification Of Current Content Sliding Modelling Techniques", New Zealand Society of Earthquake Engineering Conference, Aotea Centre, Auckland, 21-23 March 2014. Paper O67.
  21. Yeow, T.Z., MacRae, G.A., Dhakal, R.P. and Bradley, B.A. (2014) Experimental studies on the sliding behavior of building contents. Anchorage, AK, USA: 10th US National Conference on Earthquake Engineering (NCEE10), 21-25 Jul 2014. 10pp.
  22. Yeow T.Z., MacRae G.A. & Dhakal R.P. (2015). "Incorporating content sliding into seismic building performance assessments", New Zealand Society of Earthquake Engineering Annual Conference Proceedings, Rotorua, April. PAPER and POSTER
  23. Yeow T. Z., MacRae G. A., and Dhakal R, (2015). Effect of Strength and Stiffness on Single-Storey Buildings on Content Sliding Response in Earthquakes, 8th International Conference on Behavior of Steel Structures in Seismic Areas, STESSA, Shanghai, China, July 1-3, p1047-1054.
  24. Yeow T. Z.; MacRae G. A; Dhakal R. P; Lin S-L, Predicting The Maximum Total Sliding Displacement Of Contents In Earthquakes, Journal of Architectural Engineering, ASCE, 2016, 22(1): 04015013. DOI: 10.1061/(ASCE)AE.1943-5568.0000193.
  25. Yeow T. Z.; Doctoral Dissertation, Supervised by MacRae GA, Dhakal R and Bradley B., University of Canterbury, 2016.

Gregory MacRae, Trevor Yeow, and Rajesh DhakalPage 1 of 6