Equivalent Lateral Force Procedure for a Building with a Self-Centering

Rocking Story of Cross-Laminated Timber (CLT) Walls

Marco Lo Ricco1, Douglas R. Rammer1, M. Omar Amini1, Al Ghorbanpoor2, Shiling Pei3,

and Reid B. Zimmerman4

ABSTRACT: The Equivalent Lateral Force (ELF) procedure is the most widely used seismic analysis approach, because
of its simplicity and practicality in preliminary and final design phases. This paper applies the ELF procedure to a
hypothetical building that stands 5 stories tall, with a 4-story superstructure supported on a rocking story of elliptically
profiled cross-laminated timber (CLT) walls. First-generation prototypes made from six CLT panels of 5-ply, 175 mm,
thickness—each measuring 2.44 m by 3.66 m in respective width and height—demonstrated that elliptical geometry
controls lateral stiffness, inherent damping, and self-centering of the walls. Full-scale, cyclic, quasi-static, lateral-load-
displacement tests—under simulated gravity loads ranging from 133 to 400 kN—established effective stiffness and
damping inputs for the ELF procedure. The prototypes produced two modes of elliptical pendulum response by changing
steel connections to the floor and ceiling beams. The first connection guides panels through rolling, and the second
connection forces panels into slip-friction for enhanced damping but reduced durability of CLT. Because the base rocking
story of elliptically profiled CLT walls behaves like an inverted pendulum system, the ELF procedure references existing
design provisions for seismically isolated structures.

KEYWORDS: Equivalent Lateral Force (ELF) procedure, seismic isolation, rocking story, pendulum, cross-laminated
timber (CLT)

1 INTRODUCTION 123 1.1 SEISMIC DEVELOPMENT STRATEGY

Over a decade ago, mass timber construction rose to For broad applicability in North America, Pei et al. [3, 4]
international prominence with superstructures primarily mapped a comprehensive strategy to develop CLT
made of cross-laminated timber (CLT) panels. Crews of structures for an array of performance objectives. By
few on-site workers rapidly erected structures up to 9 outlining multiple tiers of seismic performance, the Pei et
stories tall, facilitated by prefabricated CLT panels of al. (2014, 2016) roadmap for structural research opened
robust sizes that functioned versatilely as loadbearing several pathways to building code conformance in the
walls or slabs. 3D seismic shake-table testing of 3- and 7- United States. Amini et al. [5] developed a multistory
story CLT structures built to full-scale, Ceccotti et al. [1, CLT construction system of platform-framed shearwalls
2] showed that such fully panelized archetypes can and determined the seismic performance factors via the
withstand simulated earthquakes. Horizontal rigorous and recently implemented FEMA P695 [6]
accelerations, however, exceeded 3g at a top floor and methodology—which involves prototype testing,
revealed a need to alleviate rigidity of this efficient archetype development, dynamic nonlinear analysis, and
assembly type. Since pioneering these CLT forms, the peer review to quantify the seismic performance factors
producers and archetypes of mass timber buildings have most widely referenced by structural building code
expanded globally. Because many sizable timber provisions.
construction hubs coincide with seismicity, around the Pei et al. [7] led similar efforts to develop prototypes and
world, continued development of seismic mitigation archetypes of mass timber structures, laterally supported
strategies for mass timber remains a priority for growth. by vertically post-tensioned CLT rocking walls, for next-
generation Performance-Based Seismic Design (PBSD)
procedures that aim for earthquake resiliency. Pei et al.

1
Marco Lo Ricco, USDA, Forest Service, Forest Products 2
Al Ghorbanpoor, Department of Civil & Environmental
Laboratory, Madison, WI, United States, Engineering, University of Wisconsin—Milwaukee, United
[email protected] States, [email protected]
Douglas R. Rammer, USDA, Forest Service, Forest Products 3
Shiling Pei, Department of Civil & Environmental
Laboratory, Madison, WI, United States, Engineering, Colorado School of Mines, Golden, CO, United
[email protected] States, [email protected]
M. Omar Amini, USDA, Forest Service, Forest Products 4
Reid B. Zimmerman, KPFF Consulting Engineers, Portland,
Laboratory, Madison, WI, United States, OR, United States, [email protected]
[email protected]
[8] tested a full-scale 2-story mass timber structure
laterally supported by vertically post-tensioned walls of
CLT to demonstrate resiliency of such a rocking system. Stiff superstructure
3D shake-table tests of a 10-story version are planned for
2021 [9].

1.2 CONTROLLED ROCKING

Whether striving for enhanced ductility or resiliency in Rocking story
earthquake mitigation, controlled rocking has proven
Podium
effective for inherently rigid panels. To explain how rigid
structures of seemingly unstable proportions had survived
devastating earthquakes, Housner mathematically Elliptically profiled CLT
demonstrated that dynamic pulses could excite tall and
slender rigid blocks into inverted pendulum motion [10].
Figure 1: Diagram of stiff superstructure over rocking CLT
Despite this breakthrough in mechanics, ways to control story
rocking in the design of earthquake-resistant buildings
developed decades later. Eventually, the precast concrete McVitty and Taylor [17] generally likened the concept of
industry made use of rocking mechanisms with restraints, an engineered soft-story to seismic isolation, because
like vertical and unbonded post-tensioning, and a variety both systems concentrate large displacements at a single
of energy-dissipating connections that engaged when horizontal plane to buffer the superstructure from
gaps opened at panel joints [11]. Based on historical seismically induced forces. Whether establishing the
development and recently proven performance, rocking isolation plane with compact bearings or story-tall
wall adaptations to mass timber prevail in current components, the engineered system must (a) maintain
research [12-14]. Unlike ductile yielding, which produces stability when subject to design displacements in various
permanent inelastic deformation by design, pendulum directions, (b) inherently develop a restoring force that
oscillations settle to an original equilibrium position that increases resistance with increasing lateral displacement,
produces negligible residual displacement. Rocking thus (c) sustain cyclic lateral loading without significant
offers intrinsic seismic resiliency not found in other degradation, and (d) exhibit predictable lateral force-
lateral force-resisting systems. deflection characteristics and damping [15]. This paper
first addresses the theory and prototype testing used to
1.3 SEISMIC ISOLATION AND SOFT-STORY determine engineering properties of the elliptically
OBJECTIVES profiled CLT rocking wall system and then presents an
This project uses ASCE 7-16 [15] Chapter 17 provisions example of the Equivalent Lateral Force (ELF) analysis
of seismic isolation to develop controlled rocking of procedure, to highlight isolation requirements specific to
elliptically profiled CLT wall panels, because the elliptical rocking wall panels. The results show the
prototyping and analysis requirements of the isolation implications of two different mechanisms of story shear
standard provide a framework that achieves resiliency. transfer and geometric proportioning on the isolation,
Table 1, adapted from Chapter C17 of ASCE 7-16, damping, and lateral displacements of the elliptically
compares the performance expectations of fixed-base and profiled systems. Finally, this paper addresses wind
isolated buildings underlying the code provisions and restraints and other resolvable dilemmas that seismic
shows that isolated buildings aspire to preserve structural isolation systems commonly encounter.
and non-structural components, in addition to life safety.
This CLT system enlarges and elaborates the concept of 2 ANALYSIS MODELS
elliptical rolling rod isolation [16] to create a rocking soft Jangid and Londhe [16] proposed elliptical rolling rod
story as Figure 1 illustrates, for example. isolation decades ago, as a compact bearing system
placed beneath a multistory superstructure. Their model
Table 1: Comparison of Earthquake Performance Expectations numerically demonstrated the effectiveness of various
ellipse eccentricities as inverted pendulum isolators.
Earthquake Intensity a Londhe and Jangid [18] later developed expressions to
distinguish non-rolling, rolling, and sliding conditions
Performance Minor Moderate Major
between the elliptical rolling rods and bounding
Life safety F, I F, I F, I horizontal planes of the foundation and superstructure.
Structural F, I F, I I Practical application, however, requires more details of
damage story shear transfer.
insignificant This newly developed CLT version of the elliptical
Nonstructural F, I I I pendulum concept addresses several practical matters.
damage First, realizing only a middle portion of an ellipse
insignificant provides an adequate range of displacement capacity. (i.e.
ellipse ends may be truncated as shown by the dashed
a
F = fixed base; I = isolated. lines of Figures 1 and 2.) Second, the project team
identified two primary mechanisms of shear transfer that derivative of the rocking story velocity; c = half the
lead to either No-Slip Traction Rolling (NSTR) or Slip- moment arm distance between contact points at load-
Friction Rocking (SFR) modes of pendulum behavior, bearing panel edges, and p = half the height between floor
diagrammed in Figure 2. and ceiling.
c c c c Figure 3 plots the lateral force-displacement hysteresis of
W ½xb µrW W xp (µr + µs)W the panel diagrammed in Figure 2a, and idealized by
FL
Fp Equation 1, with 3 cycles of measured test data. The
prototype test apparatus applied lateral displacement

p
p

cycles and 400 kN (90 kips) of vertical load that

represented W to a CLT panel profiled with an ellipse of

d
FP semi-major axis width, a, of 2350 mm (92.5 in.) and semi-
p

p
µr W (µr + µs)W minor axis width, b, of 1829 mm (72 in.), with simulated
FL boundary conditions of Figure 2a. Looking at the figure,
½xb W xp W the idealized NSTR model matched data reasonably well,
(a) (b) with a rolling friction coefficient, µr, assumed as 1
Note: Minor forces, such as panel weight, not shown percent.
Figure 2: Free-Body Diagrams of (a) NSTR and (b) SFR xb (in.)
-20 -10 0 10 20
Subtle differences between the free-body diagrams of 150 34
Figure 2 significantly vary the stiffness, damping, and
connection details of NSTR and SFR. 100
17
50
2.1 NO-SLIP TRACTION ROLLING (NSTR)

FL (kips)
FL (kN) 0 0
The free-body diagram (FBD) of Figure 2a most closely
resembles the conditions of rolling that Jangid and -50
Londhe [16] had assumed. It shows lateral force transfer, -17
FL, applied at the edges of the panel. A counteracting -100
frictional force—the product of rolling friction
-150 -34
coefficient, µr, and superstructure weight, W—adds -500 -300 -100 100 300 500
damping. Rotation, θr, governs the rolling system as the xb (mm)
independent degree of freedom. Panel rotation laterally
displaces the superstructure a total distance, xb, and raises Figure 3: Lateral Force-Displacement Hysteresis of NSTR
or lowers the superstructure strictly vertically (a condition
met with proper layout of walls to prevent superstructure Slip of the wall at the floor or ceiling contact surfaces
rotation). In the plumb panel position, the distance would compromise self-centering of the rolling pendulum
between floor and ceiling beams measures the height of with residual displacements. NSTR therefore requires
the ellipse, 2b, and increases to 2p, as the panel rotates. adequate traction for rolling and a way to preclude slip.
For rigid-body motion (neglecting small displacements in The exploded diagram of Figure 4 illustrates the
the panel and bounding beams), the total vertical prototype assembly designed to serve as a dual-purpose
displacement of the superstructure, therefore, may be displacement restraint that limits slip and the extents of
computed as 2p minus 2b. The moment arm distance, 2c, lateral displacements, using slotted pin connections.
between panel contact points, increases with rotation, so
restoring moment acts proportionally with lateral
displacement.
Major axis width, 2a, and minor axis height, 2b, of the
profiling ellipse factor into the geometric relationships,
based on an assumption fundamental to rolling; the
horizontal length of rolling equates to distance traveled
along the elliptical arc (i.e., no slip occurs). Jangid and
Londhe [16] provide details of the geometric derivations,
which involve an elliptic integral to determine arc length.
Resolving the geometric relationships and forces acting
on the panel in Figure 2a, yields the lateral force estimate
of Equation 1:
𝑐𝑐 (e)
𝐹𝐹𝐿𝐿 = 𝑊𝑊 �𝜇𝜇𝑟𝑟 𝑠𝑠𝑠𝑠𝑠𝑠(𝑥𝑥̇ 𝑏𝑏 ) + � 𝑥𝑥𝑏𝑏 (1)
𝑝𝑝 (a)
where FL = lateral force; W = gravity load of effective (b) (d) (c)
seismic weight; µr = rolling friction coefficient; sgn =
signum function; 𝑥𝑥𝑏𝑏 = lateral displacement and 𝑥𝑥̇ 𝑏𝑏 = the
Figure 4: Exploded assembly of steel beam and CLT for NSTR
In Figure 4, plates (a) confined pin (b) within V-shaped Figure 6 plots the lateral force-displacement hysteresis of
slots. The V-shape traces a path of pin travel that is the panel diagrammed in Figure 2b and idealized by
specific to a rolling ellipse profile, so plates (a) were made Equation 2, with 3 cycles of measured test data.
for interchangeability of CLT panels during prototype xp (in.)
testing. The cutout of plates (c) enveloped the V-shaped -20 -10 0 10 20
slots of 6 elliptical panel profiles illustrated in Table 3, so 150 34
that plates (c) could be fixed via welds to track plates that
bolted to a steel beam. Plates (d) initially locked the pins 100
17
in place with sacrificial notched bolts. Steel pin (b) slid 50
through all plates and pipe bushing (e) embedded in the kd keff

Fp (kips)
Fp (kN)
CLT panel. Because the slots were shaped to engage only 0 0
if the panel slipped or rotated too far, NSTR connections k1
transferred only incidental forces during prototype tests. -50
Eloop -17
-100
2.2 SLIP-FRICTION ROCKING (SFR)
-150 -34
Figure 2b modified the rolling system to actively transfer
-500 -300 -100 100 300 500
lateral forces, FP, through horizontally constrained pins.
xp (mm)
Active force transfer through the pins shortens the
moment arm between FP forces of FBD (b), relative to the Figure 6: Lateral Force-Displacement Hysteresis of SFR
moment arm between FL forces of FBD (a). Horizontal
constraint of the pins adds sliding, denoted by coefficient All aspects of the panel corresponding to the hysteresis of
µs, to the frictional damping forces acting on the edges of Figure 6 match the panel corresponding to Figure 3. In
the panel. In SFR, CLT panels roll and slide fact, prototype testing of both NSTR and SFR
simultaneously, because the vertically slotted connection configurations used the same CLT panel and swapped the
plates of Figure 5 constrain horizontal movement of the vertically slotted SFR connection plates of Figure 5 for
pins. the V-slotted NSTR restraints of Figure 4 in respective
tests. Like Figure 3, Figure 6 presents lateral load-
displacement under an applied vertical load, W, of 400 kN
(90 kips). Both figures, furthermore, use the same scale to
facilitate visual contrast. SFR clearly generates more area
within the hysteresis loops, labeled Eloop in Figure 6 for
energy dissipated per cycle. Equation 2 accounts for this
energy increase by estimating a sliding friction
coefficient, µs, equal to 9 percent, so that sliding and
rolling coefficients sum to 10 percent.
For practical engineering analysis, Figures 3 and 6 are
used to calculate effective stiffness and damping
respective to NSTR and SFR. Though based on nonlinear
equations, both lateral force-displacement plots show that
the idealized hysteresis resembles the bilinear force-
Figure 5: Exploded assembly of steel connection for SFR deflection model referenced in ASCE 7-16 [15] Chapter
17 provisions for determining isolator characteristics. The
Vertical slots forced shear transfer to occur primarily three slopes in Figure 6 illustrate various stiffness
through bearing among the pin and plates and CLT computations derived from test data. Linear fitting along
bushing pipe. Though SFR and NSTR exhibit the same the loading and unloading branches of the hysteresis data
rotational capacity, θr, the total lateral translation, 2xp, of determines slope kd, which corresponds with the stiffness
SFR must always be less than the total lateral translation, that Equation 2 predicted using mechanics. A linear slope
xb, of NSTR, because of connection constraint. Lo Ricco fitting the nearly vertical legs of the hysteresis loop
et al. [19] derived Equation 2 to quantify effects of a determines k1, stiffness in the reversal from loading to
shorter moment arm between story shears through the unloading at extreme displacements of the hysteresis
pins and frictional sliding resistance, in addition to cycle. Slope keff determines effective stiffness, for
rolling. practical assessment of system performance. Changes to
keff, as seen over the 3 cycles plotted in Figure 6, warrants
𝑐𝑐 2𝑝𝑝 review of the system for damage, like Figure 9 shows.
𝐹𝐹𝑝𝑝 = 𝑊𝑊 �(𝜇𝜇𝑟𝑟 + 𝜇𝜇𝑠𝑠 )𝑠𝑠𝑠𝑠𝑠𝑠(𝑥𝑥̇ 𝑝𝑝 ) � − 1�� 𝑥𝑥𝑝𝑝 (2)
𝑑𝑑 cos 𝜃𝜃𝑟𝑟 𝑑𝑑 cos 𝜃𝜃𝑟𝑟
Effective stiffness, keff, depends on the magnitude of
where Fp = lateral force traveling through pins; µs = displacements and corresponding forces at the maximum
sliding friction coefficient; 𝑥𝑥𝑝𝑝 = lateral displacement and positive and negative excursions of the hysteresis cycle,
𝑥𝑥𝑝𝑝̇ = the derivative of the rocking story velocity with as computed by Equation 3:
added horizontal pin constraint; d = distance between wall |𝐹𝐹+ | + |𝐹𝐹− |
panel centroid to pin center, and θr = rotation of the 𝑘𝑘𝑒𝑒𝑒𝑒𝑒𝑒 = (3)
|∆+ | + |∆− |
elliptical panel.
where lateral forces F+ and F- correspond with the rigid floor masses atop elliptical rolling rods, excited with
positive ∆+ and negative ∆- maximum-magnitude historical records of earthquakes. They showed that
displacements of the hysteresis cycle [15]. ellipse eccentricity, e, expressed by Equation 5 in terms
of semi-major and -minor axis lengths a and b,
Determining effective stiffness, in turn, enables respectively, controlled stiffness of the system.
computation of effective damping, Beff, according to
Equation 4: 𝑎𝑎2 − 𝑏𝑏2
𝑒𝑒 = � (5)
2 𝐸𝐸𝑙𝑙𝑙𝑙𝑙𝑙𝑝𝑝 𝑎𝑎2
𝐵𝐵𝑒𝑒𝑒𝑒𝑒𝑒 = (4)
𝜋𝜋 𝑘𝑘𝑒𝑒𝑒𝑒𝑒𝑒 (|∆+ | + |∆− |)2
Ellipses with low eccentricity reduced base shear of the
where Eloop, sums the area enclosed by each lateral force- superstructure, Vb, but with greater peak lateral
deflection hysteresis cycle to determine the amount of displacements in the rocking story. More oblate ellipses
energy dissipation [15]. produced lesser peak lateral displacements but increased
base shear, Vb. When e = 0, the ellipse is a circle with no
2.3 EQUIVALENT LATERAL FORCE (ELF) resistance to lateral load. When e = 1, the wall panel
ANALYSIS essentially becomes rectangular. To test effects of
ELF is the most practical analysis approach, because the geometric proportioning at the scale of a one-story-tall
procedure simplifies the dynamic effects of earthquakes wall, this study prototyped the panels and connections of
into statically applied forces. ELF analysis has long Table 3, for cyclic lateral displacement-controlled testing.
served the design of fixed-base buildings, and ASCE 7-
16 [15] Chapter 17 provisions for isolated buildings Table 3: Prototype test matrix
include considerations specific to isolated buildings. 1 constant rectangular CLT 5-ply a panel size
McVitty and Taylor [17] summarize a design procedure
for isolated structures that references ELF in 5 of the 7 2.44 m × 3.66 m (8 ft × 12 ft)
steps. The following example carries out the first few × 6 ellipse profiles of varying eccentricity, e
steps of the procedure to compare NSTR and SFR e = 0.94 e = 0.82
performance based on preliminary layouts and estimates.

2.3.1 Objectives
The Pacific Northwest region of the United States is home
to a timber construction industry that has erected several e = 0.91 e = 0.73
notable mass timber buildings in a region of significant
seismicity. Through ELF analysis, this example of a Risk
Category II building compares NSTR and SFR rocking
wall configurations for a Site Class C location in Portland,
Oregon. Table 2 summarizes the seismic hazard values e = 0.88 e = 0.63
[20]. Building code requirements for isolated structures
focus on the Risk-targeted Maximum Considered
Earthquake, MCER, event. This example will demonstrate
a range of workable rocking wall options, for comparison
with a fixed-base building. Illustrated at maximum rotations
Table 2: Spectral response acceleration parameters (SRAPs) × 2 connection configurations
Parameter Value Description
Ss 0.884 mapped MCER, 5% damped,
NSTR SFR
SRAP at short periods (0.2 s)
S1 0.396 mapped MCER, 5% damped,
SRAP at a period of 1.0 s
SMS 1.061 the MCER, 5% damped, SRAP Figures 2a & 4 Figures 2b & 5
at short periods (0.2 s) adjusted
for site class effects × 3 weights, W c
SM1 0.594 the MCER, 5% damped, SRAP 133 kN 267 kN 400 kN
at a period of 1.0 s adjusted for
site class effects
SDS 0.707 design, 5% damped, SRAP at
short periods (0.2 s) = 36 cyclic displacement protocols planned b
SD1 0.396 design, 5% damped, SRAP at a a
Total panel thickness = 170 or 175 mm due to change in
period of 1.0 s planing face laminations between CLT production runs
b SFR could not complete extreme cycles under medium
2.3.2 Scope and Limitations of Prototypes and high W, because of CLT damage near pins
Jangid and Londhe [16, 18] numerically analyzed c
Gravity simulated by synchronized vertical actuators
dynamic responses of a moment-frame structure with
ASCE 7-16 [15] Chapter 17 requirements for prototype displacements, the apparatus was reconfigured to displace
testing are written for direct application to building from the plumb position only in one direction.
projects. The standard protocols express seismic
displacement steps for each test cycle as fractions of the
maximum lateral translational displacement, DM, of the
center of rigidity for an isolator layout. At 0.75DM the W
standard requires site-specific parameters SM1 and SMS to
FL
determine if more than 10 cycles are required to (b)
demonstrate durability. The site parameters listed in
Table 2, for example, prescribe 17 cycles at the 0.75DM
load step. The standard, furthermore, requires tests under
3 load combinations to determine whether uplift or (c)
additional weight on isolators has an appreciable effect on (a) FL (d)
stiffness, damping, or stability. Therefore, each of these
displacement, durability, and stability requirements needs (b)
a known building design and site location. W

For a generalized approach, we substituted the weights of Figure 8: Prototype test setup
Table 3 for design load combinations and maximum
lateral displacement capacity of an individual wall Throughout tests, NSTR protected panels from
component, δM, for DM. Figure 7 shows the cyclic significant damage. In contrast, SFR damaged CLT
protocol in terms of δM, which varies for each of the 6 around the pin bushings and wore out the loadbearing
elliptical profiles of Table 3 with the safe rotational edges, as shown in Figure 9. Therefore, many SFR
capacity (determined by contact of the corners) illustrated protocols under medium and high weight could not be
by the test matrix. For a given ellipse eccentricity, NSTR completed [19, 21]. Nonetheless, this example assumes
and SFR panels have the same rotational capacity, but the that SFR behaves as idealized in Figure 6, with broader
vertically slotted connections of SFR imposed a distribution of forces at pins and edges. The first-
horizontal pin constraint that reduces δM by the generation prototypes reserved many practical details,
magnitudes plotted in Figure 7. like wind restraints, to focus on seismic behaviors of the
rolling wall pendulum. The prototypes, moreover, only
δ (in.) tested bearing contact surfaces of softwood CLT and
0 5 10 15 20 25 30 35 unpainted hot-rolled steel plates in a conditioned lab
environment. The ASCE 7-16 [15] Chapter 17 isolation
SFR standard requires dynamic testing of such practical details
0.91 0.94

NSTR
to envelope displacements and forces transferred by the
SFR
NSTR system in analysis. Therefore, first-generation prototypes
Ellipse eccentricity, e

provide a substantial but limited basis for the following

SFR ELF analysis example. Particularly for SFR, more
0.88

NSTR
practical development is required prior to construction of
SFR elliptically rocking soft stories in actual buildings.
0.82

NSTR

SFR
0.73

NSTR

SFR
0.63

NSTR

0 500 1000
Step: 0.25 0.50 0.67 0.75 1.00 δM (mm)
# cycles: 3 3 3 18 6
Figure 7: Cyclic lateral displacement of quasi-static protocols (a) Hole elongation around steel bushing
Figure 8 pictures the test setup. Horizontally oriented
actuators (a), stacked in series for cyclic displacements up
to ±432 mm (17 in.), pushed or pulled to roll bottom beam
(b) along a linear track mounted to the floor. Vertically
oriented actuators (c) stroked synchronously to level and
roll the top beam (d) along columns, while the actuators
maintained a constant sum of load as the contact point
traveled along the top beam span. For lateral displacement
steps up to 432 mm (17.0 in.), the apparatus could apply (b) Edge wear of 3 panels
fully reversed cycles, displacing the panel nearly equally
left and right of the plumb position. For greater lateral Figure 9: Damage observed after SFR tests
2.3.3 Conditions of ELF Applicability Equations 8, 9 and 10 respectively calculate DM, DTM, and
ASCE 7-16 [15] Chapter 17 places seven main conditions kM of the rocking story.
on ELF for use as a final analysis procedure in the design 𝑠𝑠𝑆𝑆𝑀𝑀1 𝑇𝑇𝑀𝑀
of isolated structures. This example complies with the 𝐷𝐷𝑀𝑀 = (8)
following limitations. 4𝜋𝜋 2 𝐵𝐵𝑀𝑀
1. Site Class must fall within the range of A to D. where BM = numerical coefficient for the effective
2. The effective period of the of the isolated structure, damping of the isolation system βM at the
TM, must fall below 5.0 seconds. displacement DM, as provided in ASCE 7-16 [15]
3. The supported structure must be limited to 19.81 m Table 17.5-1.
(65 ft) in height, or 4 stories, though the code permits
exceptions if overturning effects on the 𝑦𝑦 12𝑒𝑒𝑟𝑟
superstructure avoid placing any of the isolators into 𝐷𝐷𝑇𝑇𝑀𝑀 = 𝐷𝐷𝑀𝑀 �1 + � 2 � 2 � (9)
net tension. Figure 1, therefore, presents a 4-story 𝑃𝑃𝑇𝑇 𝑏𝑏 + 𝑑𝑑2
superstructure over a rocking story, for a building
that rises a total of 5 stories. where y = the distance between centers of rigidity of the
4. Effective damping of the isolation system must not isolation system and the element of interest measured
exceed 30 percent. perpendicular to the direction of seismic loading;
5. The effective period at maximum displacement, TM er = the actual eccentricity measured in plan between the
calculated by Equation 6, must exceed the elastic center of mass of the superstructure and center of rigidity
fixed-base period of the structure by at least 3 times. of the soft-story, plus accidental eccentricity taken as 5%
of the longest plan dimension of the structure
𝑊𝑊 perpendicular to the direction of force under
𝑇𝑇𝑀𝑀 = 2𝜋𝜋� (6)
𝑘𝑘𝑀𝑀 𝑠𝑠 consideration;

where kM = effective stiffness at maximum lateral b = the shortest plan dimension of the structure measured
displacement and g = gravitational acceleration. In perpendicular to d;
other words, the rocking story must support a
relatively rigid superstructure. In the absence of data 1 2 3 4 5
specific to a CLT buildings, Equation 7, from ASCE
7-16 [15] Chapter 12, approximates the fundamental
A
period.

DTM
𝑇𝑇𝑎𝑎 = 𝐶𝐶𝑡𝑡 ℎ𝑛𝑛𝑥𝑥 (7)
B
DM
Center
where Ct = 0.0488 (metric), 0.02 (U.S. units) and x = of

b
0.75 for all structural systems, other than the steel Rigidity
and concrete systems specified in ASCE 7-16 [15] C
Table 12.8-2, and hn = structure height. Providing a
relatively rigid superstructure to meet this relative
period requirement enhances effectiveness of the
D
isolation.
6. The superstructure must be regular. d
7. The isolation system has:
d = 4 Equal spaces @ 7.315 m (24 ft) = 29.261 m (96 ft)
a. Effective stiffness, kM, at maximum
b = 3 Equal spaces @ 7.315 m (24 ft) = 25.946 m (72 ft)
displacement, DM, greater than one-third the er = 0.05d + distance between centers of rigidity and mass
effective stiffness at 0.20DM, to prevent excessive
softening. Figure 10: Plan view of laterally displaced rocking story
b. Lateral restoring force at DM that exceeds the
lateral force at 0.5DM by at least 0.025W, for PT = ratio of effective translational period of the isolation
reliable self-centering. Ellipse eccentricity, as system to effective torsional period of the isolation
discussed previously, must be great enough to act system, which need not be less than 1.0. ASCE 7-16 [15]
as a pendulum, rather than a circular roller. Chapter 17 provides an approximate method of
c. Freedom to displace through the total maximum determining this period ratio, for use prior to dynamic
displacement, DTM equal to the sum of analysis.
translational and torsional displacements, before
engaging restraints. ∑|𝐹𝐹𝑀𝑀+ | + ∑|𝐹𝐹𝑀𝑀− |
𝑘𝑘𝑀𝑀 = (10)
Figure 10 shows a regular scheme of rolling pendulum 2𝐷𝐷𝑀𝑀
walls, with panels forming a core and perimeter. The
diagram superposes translational DM and total DTM effects where
that story shear, Vb, and eccentricity, er, generate. ∑𝐹𝐹𝑀𝑀+ = sum for all isolator units of the absolute value of
force at a positive displacement equal to DM.
∑𝐹𝐹𝑀𝑀− = sum for all isolator units of the absolute value of motion. Because frictional sliding was the predominant
force at a negative displacement equal to DM. damping mechanism, less sliding produced less damping.
The V-shaped slots of NSTR connections grew less
In the building layout of Figure 1, the 4-story pronounced as ellipse eccentricity increased, so more
superstructure height, hn from Level 2 through roof, oblate profiles generally survived SFR longest in
measures 15.85 m (52 ft) tall, and the overall 5-story prototype tests, before succumbing to the damage
building rises to (65 ft) with 3.96 m (13 ft) of height photographed in Figure 9.
allotted equally to each story. Table 3 provides an
estimate of building weights at each level and the
60.0
summation. ASCE 7-16 [15] Chapter 4 specifies live load
intensities and reduction factors, R1, for various
occupancies and structural components. 45.0

keff (kN/mm)
The effective seismic weight of the superstructure SFR
structure includes the entire dead load and half of the 30.0
Equation (2)
reduced live load to yield W equal to 12,113 kN (2723
kips). Sizing thickness of individual CLT wall panels and NSTR
15.0
other detailed design checks must include the load
combinations specified in ASCE 7-16 [15], Chapter 17. Equation (1)
Effective seismic weight W factors into the Equation 6 0.0
effective period calculation, TM, which in turn influences 0.6 0.7 0.8 0.9 1
e
maximum displacement, DM, and effective stiffness, kM,
of the system. Figure 11: Effective stiffness (W = 1.0 kN) at 0.5δM of panels
profiled to various ellipse eccentricities
Table 4: Building weights
D L 25
Level (kN) (kips) (kN) (kips) R1
20
Roof 1117 251 646 145 0.611 SFR
βeff (% of critical)

5 1911 430 2541 571 0.562 15

4 1911 430 2541 571 0.562 10

3 1911 430 2541 571 0.562 NSTR
5
2 2210 497 2541 571 0.562

∑ 9060 2037 10,810 2429 0

0.6 0.7 0.8 0.9 1
e
3 RESULTS
Figure 12: Effective damping at 0.5δM of panels profiled to
Figures 11 and 12 summarize the effective engineering various ellipse eccentricities
properties, at the 0.50δM step plotted in Figure 7, of 6 CLT
Table 5 summarizes the results of one iteration of
panels prototyped in the Table 3 test matrix. Figure 11
analysis, using the effective properties of Figures 11 and
plots the mean effective stiffness, with each keff data point
12 to determine effective stiffness, kM, and period, TM, at
representing 3 load trials of the Table 3 test matrix, by maximum displacement, DM. Among 12 computations,
normalizing results to a unit of vertical load W. This only the panel profiled to 0.82 ellipse eccentricity
normalization typically produced the same lateral
matched DM and 0.5δM on the first iteration. For the other
stiffness values within a few percent error. Generally,
11 options, kM must be recomputed using effective
lateral stiffness increased nonlinearly with ellipse
properties near the DM tabulated here, until iterations
eccentricity, with SFR further increasing stiffness relative
converge. Prototype data at 0.25δM, for example, may
to NSTR.
lead to rapid convergence of the 0.73e SFR wall, because
Figure 12 plots the effective damping results. Except for the first iteration of DM equates to approximately 0.25δM.
one outlier, the panel with least elliptical eccentricity, For displacements that fall between displacement steps of
effective damping essentially equated to 5 percent for the the testing protocols, idealized bilinear hysteresis models,
NSTR configuration. As expected, SFR produced like the generic fit or mechanical models of Equations 1
considerably more damping. In contrast to the relatively and 2, may estimate engineering properties.
constant damping of NSTR, the damping of SFR varies
nonlinearly and diminishes with increasing ellipse Because approximate period, Ta, of the superstructure
eccentricity. The sliding distance that panels must travel equals 0.4 seconds, the rocking system should aim for a
to rotate in SFR offers a plausible explanation. The period at least 1.2 seconds long. The 0.82e walls
vertically slotted plates of Figure 5, force panels of lesser lengthened the period to 2.5 seconds with a lateral
eccentricity to slide greater distances through the rolling maximum displacement of 376 mm (14.8 in.). In
configurations like Figure 10, torsional effects typically 1.2
amplify displacement demands up to 15 percent [15]. This Ta of fixed-base
1.0
places DTM within ample lateral displacement reach, δM,
of the wall components. In addition to in-plane wall 0.8
movement, walls of the rocking story must undergo out-
0.6
of-plane displacements compatible with the magnitudes

Sa (g)
of DM and DTM. Flexible steel connections and profiling 0.4 TM of 0.82e NSTR
load bearing surfaces of CLT panels to eliminate abrupt
corners are critical to stability during these large 0.2
displacement excursions in both lateral directions. Biaxial 0.0
testing is therefore needed to further validate this rocking 0 1 2 3 4 5
story system of elliptically profiled walls. Period T (s)
Table 5: Effective engineering properties at DM Figure 13: MCER response spectrum
kM TM DM 0.5δM
Table 6: Base shears determined by ELF for MCER
e BM a (kN/mm) (s) (mm) (mm)
0.63 1.0 2.91 4.1 604 430 Fixed-Based (R = 3) Rocking Story Superstructure
0.73 1.0 4.72 3.2 474 434 Va Vb Vst
(kN) (kips) (kN) (kips) (kN) (kips)
NSTR

0.82 1.0 7.51 2.5 376 370

0.88 1.0 12.23 2.0 295 267 4285 963 2824 635 2218 499
0.91 1.0 17.44 1.7 247 198 (12.8-1) (17.5-5) (17.5-7)
0.94 1.0 28.95 1.3 192 124 Bottom row references ASCE 7-16 [15] equation numbers.
a
0.63 1.5 7.15 2.6 257 408 To calculate V for MCER, SMS and SM1 substitute SDS and SD1 in
0.73 1.5 13.69 1.9 186 358 standard design earthquake equations.
0.82 1.4 20.59 1.5 165 283 R = Response Modification Coefficient
SFR

0.88 1.3 24.95 1.4 159 207

Equation 11 for shear applied to the rocking story
0.91 1.3 37.67 1.1 133 154
captures the essence of isolation systems with a classic
0.94 1.1 56.20 0.9 125 99
model of a spring. Therein lies the resiliency of the rolling
See Equations (10) (6) (8)
a
Linearly interpolated from ASCE 7-16 [15] Table 17.5-1 elliptical system and power of simplicity in the ELF
method of analysis.
4 CONCLUSIONS 𝑉𝑉𝑏𝑏 = 𝑘𝑘𝑀𝑀 𝐷𝐷𝑀𝑀 (11)
Figure 13 plots the MCER response spectrum of Table 2 The rocking story apportions Vb to walls in the form of FL
parameters and superposes the solution highlighted in or FP respective to NSTR and SFR systems diagrammed
Table 5 and approximate period of a fixed-base, 5-story in Figure 2. One connection constraint changed the role
building. The rocking story reduced spectral acceleration of friction, from the primary mechanism of shear transfer
nearly 80% according to the ELF analysis. Though rolling to a counteracting force that added damping. Because the
isolation can achieve effective periods of 3 or 4 seconds, magnitude of friction between the bearing contact
according to Table 5, the response spectrum shows less surfaces was unknown, full-scale prototype testing was
dramatic Sa reductions for longer periods beyond 2.5 necessary for better estimation of forces and the
seconds. hysteresis characteristics.
The ELF procedure readily determines the base shears of
Table 6 for comparison of fixed-based and isolated 5 FURTHER DEVELOPMENT
schemes. For the isolated building, Vb acts at floor and Because physical construction profoundly influences
ceiling levels of the rocking story, and Vst accumulates performance, ASCE 7-16 [15] Chapter 17 seismic
shear from the remaining upper floors and roof. Relative isolation provisions require prototype testing. Future
to the fixed-base building, the rocking story provides a prototypes of this elliptical system may place sacrificial
35% reduction in base shear. Table 1, however, reveals a wind restraints at corners of the panels and U-shaped
major mismatch in this base shear comparison. For the flexural plates spanning panel joints to assess practical
fixed-base building to remain elastic at the MCER event, details. Future prototypes may characterize additional
design must target roughly 3 times the tabulated base contact bearing surfaces in dynamic tests and model long-
shear, V, to eliminate ductility. In other words, the term performance of wood. To address questions
reduction in base shear grows to approximately 75% for common to isolation systems, the performance-based
the rocking story building if the fixed-base structure provisions of ASCE 7-16 [15] provide the general
targets a similar level of insignificant damage during the framework to analytically and physically develop
MCER event. complete isolation systems.
ACKNOWLEDGEMENT Wood Project. Available from:
http://nheritallwood.mines.edu/index.html 2018
The U.S. Forest Service, in cooperation with the [cited 2020 September 8].
University of Wisconsin—Milwaukee, funded this [10] Housner G.W., The behavior of inverted pendulum
research through the Wood Innovations Grant Number: structures during earthquakes. Bulletin of the
2016-DG-11420004-170. seismological society of America, 53(2): 403-417,
1963.
REFERENCES [11] Restrepo J. and Rahman A., Seismic Performance of
Self-Centering Structural Walls Incorporating
[1] Ceccotti A. and Follesa M., Seismic Behavior of
Energy Dissipators. Journal of Structural
Multi-Storey XLam Buildings, International
Engineering, 133(11): 1560-1570, 2007.
Workshop on Earthquake Engineering on Timber
[12] Ganey R., Berman J., Akbas T., Loftus S., Daniel
Structures, pages 81-95, COST E29- European
Dolan J., Sause R., Ricles J., Pei S., van de Lindt
Cooperation in the Field of Scientific and Technical
J.W., and Blomgren H.-E., Experimental
Research: Coimbra, Portugal, 2006.
Investigation of Self-Centering Cross-Laminated
[2] Ceccotti A., Sandhaas C., Okabe M., Yasumura M.,
Timber Walls. Journal of Structural Engineering,
Minowa C., and Kawai N., SOFIE project – 3D
143(10): 04017135, 2017.
shaking table test on a seven-storey full-scale cross-
[13] Akbas T., Sause R., Ricles J.M., Ganey R., Berman
laminated timber building. Earthquake Engineering
J., Loftus S., Dolan J.D., Pei S., Lindt J.W.v.d., and
& Structural Dynamics, 42(13): 2003-2021, 2013.
Blomgren H.-E., Analytical and Experimental
[3] Pei S., Berman J., Dolan J.D., van de Lindt J.W.,
Lateral-Load Response of Self-Centering
Ricles J., Sause R., Blomgren H.-E., Popovski M.,
Posttensioned CLT Walls. Journal of Structural
and Rammer D.R. Progress on the development of
Engineering, 143(6): 04017019, 2017.
seismic resilient Tall CLT Buildings in the Pacific
[14] Kovacs M. and Wiebe L. Controlled Rocking
Northwest. In: Proceedings of the 2014 World
Cross-Laminated Timber Walls for Regions of
Conference on Timber Engineering, Quebec City,
Low-to-Moderate Seismicity. In: World Conference
Canada, 2014.
on Timber Engineering (WCTE 2016), Vienna,
[4] Pei S., van de Lindt J.W., Popovski M., Berman
Austria, 2016.
J.W., Dolan J.D., Ricles J., Sause R., Blomgren H.-
[15] American Society of Civil Engineers (ASCE) and
E., and Rammer D.R., Cross-Laminated Timber for
Structural Engineering Institute (SEI), Minimum
Seismic Regions: Progress and Challenges for
Design Loads and associated Criteria for Buildings
Research and Implementation. Journal of Structural
and Other Structures (ASCE/SEI 7-16), American
Engineering, 142(4): E2514001, 2016.
Society of Civil Engineers: Reston, VA, USA,
[5] Amini M.O., van de Lindt J.W., Rammer D.R., Pei
2017.
S., Line P., and Popovski M., Systematic
[16] Jangid R.S. and Londhe Y.B., Effectiveness of
experimental investigation to support the
Elliptical Rolling Rods for Base Isolation. Journal
development of seismic performance factors for
of Structural Engineering, 124(4): 469-472, 1998.
cross laminated timber shear wall systems.
[17] McVitty W. and Taylor A., Chapter 15: Seismically
Engineering Structures, 172: 392-404, 2018.
Isolated Structures, In R. Pekelnicky, editor, 2015
[6] Federal Emergency Management Agency (FEMA):
NEHRP Recommended Seismic Provisions: Design
Quantification of Building Seismic Performance
examples. FEMA P-1051, pages (15)1-58. Federal
Factors, Federal Emergency Management Agency
Emergency Management Agency (FEMA),
(FEMA), Washington D.C., 2009.
Washington D.C., 2016.
[7] Pei S., van de Lindt J.W., Ricles. J.M., Sause R.,
[18] Londhe Y.B. and Jangid R.S., Dynamic response of
Berman J., Ryan K., Dolan J.D., Buchanan A.H.,
structures supported on elliptical rolling rods.
Robinson T., McDonnel E., Blomgren H.-E.,
Doboku Gakkai Ronbunshu, 612: 11-20, 1999.
Popovski M., and Rammer D.R., Development and
[19] Lo Ricco M., Ghorbanpoor A., Pei S., Rammer
Full-Scale Validation of Resilience-Based Seismic
D.R., Begel M., Bridwell J., and Zimmerman R.B.,
Design of Tall Wood Buildings: The NHERI
Prototyping a Passively Self-Centering Cross-
Tallwood Project, Journal Name: Proceedings of
Laminated Timber Rocking Wall System:
the New Zealand Society for Earthquake
Analytical and Experimental Investigation. Wood
Engineering Annual Conference, April 27-29,
Design Focus, 28(2): 23-39, 2018.
Wellington, New Zealand, 2017.
[20] Structural Engineers Association of California
[8] Pei S., van de Lindt J.W., Barbosa A., Berman J.,
(SEAOC). Seismic Design Map Tool. 2020 [cited
McDonnell E., Dolan J.D., Zimmerman R.B., Sause
2020 August 12].
R., Ricles J., and Ryan K. Full-Scale Shake Table
[21] Lo Ricco M.: Prototyping Elliptically Profiled
Test of Mass-Timber Building With Resilient Post-
Inverted Pendulum Walls in Cross-Laminated
Tensioned Rocking Walls. In: World Conference on
Timber (CLT) for Passive Sellf-Centering and
Timber Engineering, Seoul, Republic of Korea,
Seismic Resiliency. Civil & Environmental
2018.
Engineering, Doctor of Philosophy in Engineering.
[9] Pei S., Berman J., Ryan K., Ricles J., Sause R.,
University of Wisconsin-Milwaukee, Milwaukee,
Dolan J.D., and van de Lindt J.W. Natural Hazards
Wisconsin, 2019.
Engineering Research infrastructure (NHERI) Tall
In: Proceedings of the World Conference on Timber Engineering, August 9-12, 2021, Santiago, Chile.
Paper No: TE0321.