Advances in Engineering Software 198 (2024) 103765

Contents lists available at ScienceDirect

Advances in Engineering Software

journal homepage: www.elsevier.com/locate/advengsoft

Comparative analysis of seismic response reduction in multi-storey

buildings equipped with base isolation and passive/active friction-tuned
mass dampers
Morteza Akbari a,*, Mohammad Seifi b , Tomasz Falborski a , Robert Jankowski a
a
Faculty of Civil and Environmental Engineering, Gdańsk University of Technology, Gdańsk, Poland
b
Department of Civil and Industrial Engineering, Sapienza University of Roma, Roma, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: This study presents an innovative approach to mitigating seismic responses in multi-storey buildings equipped
Comparative analysis with a base-isolation (BI) system and passive friction-tuned mass dampers (PFTMDs). The key innovation lies in
Proportional integral derivative controller the combined use of a BI system and a PFTMD system, as well as the activation of this mechanical system by
Base isolation
controllers. Additionally, the research design optimizes the parameters of these devices specifically for each
Passive friction-tuned mass damper
Earthquake
earthquake scenario and compares the results to the average of the optimal parameters, which has not been
Active friction-tuned mass damper investigated in previous studies. In this study, a 10-storey structure is modeled, featuring a BI system beneath the
Multi-objective particle swarm optimization first floor and a PFTMD system on the roof. The parameters for the BI, PFTMD, BI-PFTMD, and BI-active FTMD
(BI-AFTMD) systems are independently optimized using a multi-objective particle swarm optimization (MOPSO)
algorithm. To enhance the passive BI-PFTMD system, a proportional-integral-derivative (PID) controller is
incorporated into the friction-tuned mass damper system, resulting in the BI-AFTMD hybrid control system that
adjusts the final control force transmitted to the structure. The seismic performance of these systems is assessed
for the 10-storey building under both far-field and near-field earthquakes. The findings reveal that these control
systems significantly decrease average peak displacement, acceleration, and inter-storey drift as compared to an
uncontrolled structure, especially when system parameters are optimized for the same earthquake scenario.
Using average optimal parameters, the BI-AFTMD system achieves the most substantial reduction in average
peak displacement, while the BI system offers the greatest reduction in average peak acceleration and inter-
storey drift.

1. Introduction buildings to earthquakes, protecting them from potential damage [1–8].

By severing the structure from the horizontal component of ground
Earthquakes, one of nature’s most devastating forces, have the po- motion, a process known as seismic isolation can reduce the acceleration
tential to cause immense destruction and loss of life. With their unpre- of the structure and produce a system with a dominant frequency that is
dictable nature and widespread impact, researchers and engineers have far lower than the prevailing seismic frequency [9–12].
been striving to develop effective strategies to mitigate their destructive Passive control systems are designed to enhance the stability and
effects. One such approach gaining significant attention is the use of safety of structures without requiring external energy input. They use
structural control tools. Structural control refers to the integration of inherent properties of materials and devices, such as dampers and iso-
various techniques and devices within building and infrastructure lators, to absorb and dissipate energy from environmental forces like
design to actively modify the dynamic response of structures subjected earthquakes or wind. These systems are cost-effective and reliable, of-
to earthquake forces. These devices aim to enhance the overall resilience fering a low-maintenance solution for reducing vibrations and mini-
and safety of buildings, reduce structural damage, and minimize the risk mizing damage. The seismic response of a single-storey asymmetric
to occupants during seismic events. Different kinds of dampers and structure supported on several base isolation systems during impact with
seismic isolators can be used to lessen the structural response of neighboring structures was investigated by Matsagar and Jangid [13].

* Corresponding author.
E-mail address: [email protected] (M. Akbari).

https://doi.org/10.1016/j.advengsoft.2024.103765
Received 18 June 2024; Received in revised form 1 August 2024; Accepted 28 August 2024
Available online 10 September 2024
0965-9978/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

They discovered that when a base-isolated structure collides with nearby and Jangid [28] conducted a parametric study to examine how key
structures, its lateral-torsional reaction is worse. Base-raft displacement parameters such as mass ratio, tuning frequency ratio, and slip force
decreases and superstructure acceleration increases as a result of this affect the performance of a TMFD. The results indicated that for a given
contact. However, the isolation system still performs better than a excitation level, there are optimal values for the mass ratio, tuning
non-isolated structure in mitigating seismic effects [13]. Falborski and frequency ratio, and damper slip force that minimize the peak
Jankowski [14] examined the dynamic response of a one-storey steel displacement of the primary structure [28]. Additionally, it was
structural model and found that seismic excitations could pose a serious observed that, with an appropriately selected slip force, the TMFD can
threat to the structure’s safety by causing intense vibrations. They also effectively and significantly control undesirable system responses [28].
conducted an experimental study to assess the effectiveness of elasto- Jiang et al. [29] investigated the potential of using damping generated
meric polymer bearings (EPBs) in reducing structural vibrations during by the friction between movable flange-mounted ball bearings and a
dynamic excitations [15]. The study revealed that the use of EPB stationary shaft. Their experimental and simulation results showed that
significantly enhanced the dynamic characteristics by reducing struc- the proposed FTMD device significantly enhanced the damping ratio of
tural vibrations [15]. Ocak et al. [16] examined a seismic isolator the primary structure from 0.35 % to 5.326 % during free vibration and
positioned at the foundation of a structure, optimized through an reduced uncontrolled structural response by approximately 90 % at a
adaptive harmony search algorithm (AHS) across different earthquake tuned frequency [29]. Khatibinia et al. [30] studied the seismic
scenarios. Their findings revealed that isolators with lower damping vulnerability control of 10-storey steel moment–resisting frames
ratios necessitate greater ductility while increasing the damping ratio (SMRFs) equipped with an optimized FTMD system. Their seismic
enhances control efficiency by further limiting the isolator’s movement fragility assessment revealed that the optimized FTMD significantly
[16]. Using a specially formulated flexible polymer material with enhances the seismic performance of the controlled SMRF as compared
improved damping properties, Falborski and Jankowski [17,18] carried to the uncontrolled SMRF across various damage states [30]. Beshara-
out an extensive study to confirm the effectiveness of an advanced tian et al. [31] employed a particle swarm optimization (PSO) algorithm
mathematical model in simulating the complex mechanical behavior of to simultaneously optimize the frequency ratio and friction ratio pa-
a seismic isolation system composed of polymer bearings (PBs). Another rameters of FTMDs installed on multi-degree of freedom (MDOF) sys-
effective method for controlling vibration involves the use of a tuned tems. Their findings indicated that FTMDs with optimized parameters
mass damper (TMD). Dadkhah et al. [19] examined the optimization of significantly reduced the average roof displacements of the examined
TMD parameters in a six-storey steel frame. They found that optimizing MDOF structures by up to 39 %, although the extent of the reduction
the TMD parameters to minimize the drift ratio reduced the displace- varied depending on the structure’s period [31]. Labbafi et al. [32]
ment of the structure. The unresolved issue of passive TMD reliance on examined the effectiveness of a new type of TMD, known as the
seismic signals has not been fully explored [19]. friction-tuned mass damper (FTMD), in reducing the seismic response of
It has been proposed that by dissipating friction energy, passive a four-span integral bridge. Their numerical results demonstrated that
friction-tuned mass dampers (PFTMDs) could improve TMD perfor- the FTMD outperforms the traditional TMD in decreasing the deck
mance by reducing input energy. A nonlinear damping system displacement [32].
comprising a TMD and nonlinear dry (pure) frictional damping at the Hybrid control systems are produced by fusing two or more active,
interface between the TMD and the primary structure was examined by semi-active, or passive control systems. Apart from their advantages,
Gewei and Basu [20]. Etedali et al. [21] examined the optimal design for TMDs and base-isolated (BI) systems could have certain disadvantages.
TMD and friction-tuned mass damper (FTMD) parameters through Researchers are therefore intrigued by the application of a hybrid system
multi-objective computational optimization [21]. The findings indicated that blends the BI structure with the TMD. The hybrid vibration miti-
that FTMDs were more effective than TMDs in reducing seismic re- gation strategy proposed by Yang et al. [33] includes passive TMD and
sponses. A low-rise building with multiple friction-tuned mass dampers rubber bearings. Tsai [34] showed that the frequency of earthquakes has
(MFTMDs) was compared to one with a single friction-tuned mass a substantial effect on the performance of BI structures fitted with TMDs.
damper (SFTMD) in Pisal’s study [22]. The findings demonstrated that Stanikzai et al. [35] investigated the seismic response of base-isolated
MFTMDs were superior to SFTMDs of equivalent mass in terms of their buildings equipped with single-tuned mass dampers (STMDs),
ability to decrease the structural response [22]. To determine the ideal multiple-tuned mass dampers (MTMDs), and distributed multiple-tuned
tuning of TMD and FTMD parameters for seismic applications, Etedali mass dampers (d-MTMDs) under real earthquake ground motions. They
et al. [23] carried out a numerical analysis utilizing a multi-objective concluded that installing a tuned mass damper on each floor level of a
particle swarm optimization technique, taking soil-structure interac- base-isolated building effectively reduces the structural response, spe-
tion (SSI) effects into consideration. Chung et al. [24] optimized the cifically in terms of top-floor acceleration and bearing displacement
design of a friction pendulum-tuned mass damper (FPTMD) using [35]. To enhance control performance, Gao et al. [36] suggested a tuned
varying friction coefficients for wind-excited tall buildings. The study negative stiffness inerter damper (TNSID) for BI structures. To lessen
revealed that the effectiveness of FPTMDs with two patterns of friction vibrations in high-rise buildings, Naderpour et al. [37] examined the
coefficients was largely influenced by the frequency tuning ratio rather efficacy of a hybrid control technique that combines dashpots that are
than by the friction parameters. Considering the effects of SSI, Etedali directly connected to the ground with nontraditional, BI-tuned mass
et al. [25,26] investigated the probabilistic reliability assessment of dampers. Akbari et al. [38] studied the seismic performance of a 5-storey
high-rise buildings with FTMDs and TMDs under near-field earthquakes. shear building equipped with Magneto-Rheological (MR) dampers
This review shows that, particularly for large earthquakes close to faults, installed at the base isolation level and two active tendons located on the
soft soil frequently reduces the efficiency of FTMD and TMD in reducing first and second floors of the structure. Kontoni and Farghaly [39]
the likelihood of failure. Pisal and Jangid [27] explored the performance examined BI and TMD systems, considering the SSI. Wang et al. [40]
of a tuned mass friction damper (TMFD) in mitigating the resonant investigated the use of a semi-active tuned mass damper (STMD) with
response of bridges under multi-axle vehicular loads. They examined variable stiffness and damping to enhance the seismic performance of
three configurations: a TMFD at the bridge’s mid-span, a multiple-tuned passive base-isolated structures. Their numerical results showed that the
mass friction dampers (MTMFD) system concentrated at the mid-span STMD generally provides the best control effect in both linear and
and an MTMFD system with TMFD units distributed along the bridge’s nonlinear models. In certain engineering and industrial applications,
length. The study concluded that an optimized MTMFD system proportional-integral-derivative controllers are utilized. The
concentrated at the mid-span is more effective than both an optimized cost-effectiveness and practicality of PID controllers have been demon-
TMFD at the same location and an optimized MTMFD system with strated (see [41]). A proportional integral derivative controller was
distributed TMFD units, despite having the same total mass [27]. Pisal presented by Palizvan Zand et al. [42], and it was shown that the

2
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 1. Model of a 10-storey building with various systems installed.

controller could considerably lower the maximum displacement, accel- placement algorithm to enhance the structures’ response to seismic ac-
eration, and velocity of structures that are susceptible to far-field tivities. Numerical results indicated that this control algorithm effec-
earthquakes. Shahi et al. [43] examined a mathematical model devel- tively reduces the response of building structures while requiring only a
oped for the time-domain analysis of tall buildings with ATMDs minimal amount of control force [50].
considering SSI effects. They discovered that the effectiveness of PID and
linear quadratic regulator (LQR) controllers decreased as the soil soft-
1.1. Research gaps
ness increased. Some scientists have used these controllers to lessen the
structural reactions to earthquakes (see [44–46] for example). Etedali
A literature review of passive, active, semi-active, and hybrid control
et al. [47] examined a new framework of fractional order
systems reveals several research gaps that the proposed method aims to
proportional-derivative-integral (FOPID) controller series with a fuzzy
address. These gaps are outlined as follows:
proportional-derivative (FPD) controller, called the OFPD-FOPID
controller. The results showed that the OFPD-FOPID controller per-
• (a) Past research has focused on designing passive, active, and
formed better than the FOPID controller in reducing peak floor
combined control systems. However, due to the unpredictable nature
displacement, acceleration, and inter-storey drift. Wang et al. [48]
of earthquakes, the optimal design of these systems’ parameters
investigated a semi-active tuned mass damper (STMD) with variable
using optimization algorithms for specific earthquake scenarios is
mass and damping (VMD-STMD). Their results revealed significant
not considered a practical approach.
variation in structural dynamic responses and TMD performance due to
• (b) Additionally, although base isolation systems and friction-tuned
the randomness of crowd-structure interaction (CSI). Overall, the
mass dampers have been utilized as components of active and pas-
VMD-STMD demonstrated the best control effect and the smallest
sive control systems, there has been limited research on the inte-
standard deviation, indicating superior control robustness. Pnevmatikos
gration of passive base isolation systems with passive and active
and Gantes [49] studied the impact of two practical issues time delay
friction-tuned mass dampers.
and saturation effect on the performance of controlled structures. They
obtained response reduction surfaces that illustrate the combined effects
of time delay and force saturation on the buildings’ behavior. Based on 1.2. Objectives of this study
their findings, they concluded the optimal selection of control systems
and the desired properties of control devices [49]. Pnevmatikos [50] The developed optimal systems aim to calculate passive control ef-
investigated the use of a control algorithm for civil structures subjected forts and precise hybrid control in civil engineering, with a focus on
to earthquake excitation. The research focused on employing the pole intelligence and practical application. The main goals of this research
are summarized as follows:

3
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Table 1 The BI system and the equation of motion (1) are connected. The BI-
Formulation of PFTMD system [20–23]. AFTMD system is connected to the equation of motion (3), whereas the
Friction Force Fd = fs sgn(ẋd (t) − ẋN (t)) PFTMD and BI-PFTMD systems are tied to the equation of motion (2).
The equations of motion of the entire system studied by the Lagrangian
Friction Coefficient fs
μd =
md .g method are as follows:

PFTMD velocity ẋd (t) Top floor velocity ẋN (t) [M]{ẍ(t)} + [C]{ẋ(t)} + [K]{x(t)} = − [M]{1}äg (1)
Friction coefficient of the PFTMD μd Slip force fs
acceleration [M]{ẍ(t)} + [C]{ẋ(t)} + [K]{x(t)} = − [M]{1}äg + {z}Fd (t) (2)
PFTMD mass md Gravitational g
acceleration
[M]{ẍ(t)} + [C]{ẋ(t)} + [K]{x(t)} = − [M]{1}äg + {z}Fd (t)
+ {z}Factive− FTMD (t) (3)
• Addressing Research Gap (a): To tackle the unpredictability of
earthquakes and the impracticality of designing control systems The relative displacement {x(t)}, velocity {ẋ(t)}, and acceleration
optimized for specific seismic events, this study proposes using the {ẍ(t)} of each storey are used to formulate the equation of motion for a
average of optimal parameters for each system. This approach ap- building with both passive and hybrid control systems. The building’s
plies to both far-field and near-field seismic excitations, providing a structural properties are represented by the mass matrix [M], damping
more rational design for structural control systems by considering matrix [C], and stiffness matrix [K]. The friction force of the PFTMD
the average optimal parameters across various earthquakes. system is denoted by Fd(t), and the active control force of the PFTMD
• Addressing Research Gap (b): To fill the gap related to the limited system is indicated by Factive − FTMD(t). The displacement vector {x(t)}, as
research on combining passive base isolation systems with friction- reported in Eq. (4), contains the relative displacement of floors xj(t),
tuned mass dampers, this study optimizes the parameters of these (j = 1, …, N), the displacement of the PFTMD system xd(t), and the
systems in a 10-storey building. This optimization is conducted displacement of the BI system xb(t), as follows, taking N to be the
through two scenarios: specific optimal design for each earthquake number of degrees of freedom characterizing the kinematics of the
and an average optimal design based on all earthquakes. Subse- superstructure:
quently, the study compares the seismic efficiency of various sys- ⎧ ⎫
tems, including a passive base isolation system (BI), a passive ⎪ xb (t) ⎪
⎪
⎪ ⎪
friction-tuned mass damper system (PFTMD), a dual approach ⎪ x1 (t) ⎪
⎨ ⎪
⎬
combining a BI system with a PFTMD system (BI-PFTMD), and a
{x(t)}12×1 = ⋮ (4)
hybrid control approach combining a BI system with an active ⎪ ⎪
friction-tuned mass damper system (BI-AFTMD). The MOPSO algo-
⎪
⎪ xN (t) ⎪
⎪
⎪
⎩ ⎪
⎭
rithm is used to optimize the system parameters, and a PID controller xd (t)
is implemented to adjust the active control power of the BI-AFTMD
hybrid system. where {1} and äg are the unit vector and ground acceleration, respec-
tively. While {z} is the location vector of PFTMD, BI-PFTMD and BI-
2. Equation of motion for an N-degree-of-freedom system AFTMD systems:
equipped with a BI-AFTMD ⎧ ⎫
⎪ 0 ⎪
Fig. 1 displays different control vibration devices applied to a 10-sto-
⎪
⎪ ⎪
⎪
⎪
⎨ 0 ⎪
⎬
rey 2D building. Namely, a BI building is equipped with the proposed
active FTMD system. The masses of the building’s storeys are lumped at {z}12×1 = ⋮ (5)
⎪
⎪ ⎪
slab levels. The structural system is based on the following key ⎪
⎪ − 1⎪
⎪
⎪
assumptions: ⎩ ⎭
1
- the PFTMD is attached to the primary structure, The following is how the mass, damping, and stiffness matrices can
- the PFTMD has a friction coefficient denoted as μd, be obtained:
- the PFTMD has specific properties, including a mass (md), damping ⎡ ⎤
(cd), and stiffness (kd), mb 0 0 0 0 0 0
- the BI system is also incorporated into the structural system,
⎢
⎢ m1 0 0 0 0 0 ⎥⎥
⎢ m2 0 0 0 0 ⎥
- the BI system has a mass (mb), damping (cb), and stiffness (kb), ⎢ ⎥
[M]12×12 = ⎢
⎢ ⋱ {0} {0} {0} ⎥
⎥ (6)
⎢ m9 0 0 ⎥
These assumptions provide the foundation for analyzing and ⎢
⎣ m10 0 ⎦
⎥
designing the structural control system, considering the properties and Sym. md
behavior of both the PFTMD and BI systems in conjunction with the
⎡ ⎤
primary structure. The inclusion of the PFTMD and BI systems in- cb + c1 − c1 0 0 0 0 0
troduces additional damping and isolation mechanisms that can effec- ⎢ c1 + c2 − c2 0 0 0 0 ⎥
⎢ ⎥
tively mitigate the overall vibration and displacement of the building. ⎢
⎢ c2 + c3 ⋱ ⋮ 0 0 ⎥⎥
These control systems are designed to target the dominant vibration [C]12×12 =⎢
⎢ ⋱ − c9 ⋮ 0 ⎥⎥
modes and dissipate or isolate the energy input, potentially reducing the ⎢
⎢ c9 + c10 − c10 ⋮ ⎥⎥
importance of local effects like torsion and deviations.
⎣ c10 + cd − cd ⎦
As shown in Fig. 1, two more degrees of freedom are added to the Sym. cd
system: one to describe the PFTMD control system and another to (7)
describe the BI control system.

4
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

⎡ ⎤
kb + k1 − k1 0 0 0 0 0 • Analytical approximation methods: Methods like perturbation
⎢
⎢ k1 + k2 − k2 0 0 0 0 ⎥⎥ techniques and harmonic balance can provide approximate analyt-
⎢
⎢ k2 + k3 ⋱ ⋮ 0 0 ⎥⎥ ical solutions to the non-linear equations.
[K]12×12 =⎢
⎢ ⋱ − k9 ⋮ 0 ⎥⎥ • Hybrid methods: Combining numerical and analytical approaches
⎢
⎢ k9 + k10 − k10 ⋮ ⎥⎥ can offer a balance between accuracy and computational efficiency.
⎣ ck10 + kd − kd ⎦
Sym. kd
The choice of solution technique depends on the complexity of the
(8) non-linear friction force model, the desired accuracy, and the available
Table 1 illustrates the method for calculating the friction force of the computational resources. Numerical methods are generally favored for
PFTMD system. their robustness and flexibility in handling complex non-linearities,
Additionally, the damper force can be expressed by the following while analytical and hybrid methods can be useful for gaining insights
equation [22,23]: into the system’s behavior and reducing computational effort [20,28,52,
54].
Fs = fs Z (9) In this research, the Runge-Kutta numerical method with a step size
of 0.001 was utilized in MATLAB software and the Simulink toolbox to
where fs denotes the limiting friction force or slip force of the damper,
solve the linear and nonlinear equations of motion. This was done to
and Z is the non-dimensional hysteretic component, which satisfies the
determine the response of structures equipped with the studied systems
following first-order non-linear differential equation:
under seismic excitation. The non-linear pure dry friction coefficient in
dz the PFTMD system is highly sensitive; Therefore, if the friction coeffi-
q = A(ẊPFTMD − ẊN ) − β|ẊPFTMD − ẊN |Z|Z|n− 1 − τ(ẊPFTMD − ẊN )|Z|n
dt cient is too high during low-acceleration earthquakes, the PFTMD sys-
(10) tem may not slip and the friction prevents the structural responses from
decreasing. Conversely, if the friction coefficient is too low during high-
where ẊPFTMD refers to the velocity of the PFTMD and ẊN represents the acceleration earthquakes, the system may fail to dissipate enough en-
velocity of the top floor. q is the yield displacement of the frictional force ergy, also resulting in inadequate reduction of the structural responses.
loop, and A, β, τ, and n are non-dimensional parameters of the hysteretic To address this, the friction force coefficient of the PFTMD system is
loop that control its shape. The parameters are chosen to ensure typical optimized using the MOPSO algorithm.
Coulomb-friction damping [22,23].
3. The multi-objective particle swarm optimization
2.1. Non-linear friction force computational algorithm

The friction force in a passive friction-tuned mass damper is inher- The optimization field has seen significant progress in recent years,
ently non-linear, implying that it does not vary linearly with the relative with one notable technique being the multi-objective particle swarm
velocity between the main structure and the mass. This non-linearity can optimization (MOPSO) algorithm. MOPSO combines swarm intelligence
arise from various factors, including the type of friction (e.g., Coulomb, and multi-objective optimization to address complex real-world prob-
viscous, or a combination) and the specific design and implementation lems with conflicting objectives. Unlike single-objective optimization,
of the friction mechanism [20,51–53]. MOPSO aims to find a set of solutions that represent a trade-off between
these conflicting objectives. Inspired by bird flocking and fish schooling
2.1.1. Modeling the non-linear friction force behavior, MOPSO has proven to be an effective approach [55–57].
To accurately capture the system’s response to seismic excitation, it
is essential to model the non-linear friction force effectively. Several 3.1. Fundamentals of particle swarm optimization (PSO)
approaches can be employed for this purpose:
Before delving into MOPSO, understanding the basic principles of
• Piecewise-linear approximation: This method approximates the particle swarm optimization (PSO) is essential. PSO simulates the social
non-linear friction force using a series of linear segments, each valid behavior of particles in a search space and has been successful in single-
for a specific range of relative velocities. objective optimization problems. In PSO, a swarm of particles moves
• Analytical models: Closed-form expressions for the non-linear through the search space, where each particle represents a potential
friction force are derived based on the specific friction characteris- solution. Particles adjust their positions based on their own experience
tics of the system. and the experience of neighboring particles to find the optimal solution
• Numerical models: The non-linear friction force is represented [55–57].
using numerical methods, such as lookup tables or interpolation and
optimization techniques. 3.2. Extension to multi-objective optimization

2.1.2. Incorporating non-linear friction into the equations of motion MOPSO extends this computational algorithm to handle multi-
Once the non-linear friction force is modeled, it needs to be incor- objective optimization problems and aims to maintain a diverse set of
porated into the equations of motion governing the PFTMD system. This nondominated solutions, known as the Pareto front. The goal is to find a
integration typically results in a set of non-linear differential equations set of solutions where no single solution is superior in all objectives, but
that describe the system’s response to seismic excitation. rather each solution is optimal in the context of trade-offs among ob-
jectives [55–57].
2.1.3. Solution techniques
Solving the resulting non-linear differential equations to determine 3.3. Mathematical formulation of MOPSO
the system’s response can be challenging. Various techniques are
available, including: The MOPSO algorithm can be described by the following equations:

• Numerical integration methods: Techniques such as the Runge- 1 Velocity Update Equation:
Kutta method and the Newmark-beta method are commonly used ( ) ( )
vt+1 = wvti + c1 r1 pti − xti + c2 r2 git − xti (11)
to numerically integrate the non-linear differential equations. i

5
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 2. The MOPSO flowchart.

Algorithm Setup and Parameter Considerations

Setting up the MOPSO algorithm involves several key steps and
where: considerations:

1. Initialization:
• Initialize a population (swarm) of particles with random positions
• vt+1
i is the velocity of particle i at iteration t + 1.
and velocities within the search space.
• w is the inertia weight.
• Evaluate the objective functions for each particle.
• c1 and c2 are cognitive and social coefficients, respectively.
• Initialize the personal best position (pi) for each particle and the
• r1 and r2 are random numbers uniformly distributed in [0,1].
global best position (gi) based on Pareto dominance.
• pti is the best position found by particle i up to iteration t.
2. Parameter Settings:
• git is the best position found by any particle in the swarm,
• Inertia Weight (w): Controls the exploration and exploitation bal-
considering the multi-objective context.
ance. A larger w facilitates exploration, while a smaller w enhances
• xti is the current position of particle i at iteration t.
exploitation.
• Cognitive Coefficient (c1): Reflects the tendency of particles to
2. Position Update Equation: return to their personal best positions.
xt+1 = xti + vt+1 (12) • Social Coefficient (c2): Indicates the tendency of particles to move
i i
towards the global best position.
• Random Factors (r1,r2): Introduce stochasticity to the velocity
update, aiding in escaping local optima.
where xt+1
i is the new position of particle i at iteration t + 1. 3. Velocity and Position Update:

6
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 3. A block diagram illustrating the suggested control scheme.

• Update the velocity and position of each particle using the velocity front. The crowding distance measures the density of solutions sur-
and position update equations. rounding a particular solution, ensuring a well-distributed Pareto front
• Apply boundary constraints to ensure particles remain within the [55–57]. Fig. 2 shows a simplified flowchart of the basic steps of the
search space. MOPSO algorithm.
4. Update Personal and Global Bests:
• Evaluate the objective functions for the updated positions. 4. Proportional integral derivative (PID) controller
• Update the personal best positions (pi) and potentially the global
best positions (gi) based on Pareto dominance. The proportional-integral-derivative (PID) controller is a widely
5. Archive Maintenance: used feedback control mechanism employed in various engineering
• Maintain an external archive of non-dominated solutions. systems. Feedback control systems are essential in numerous industries
• Use mechanisms like crowding distance to ensure diversity in the to achieve the desired system behavior and performance. The PID
archive. controller is a popular choice due to its simplicity and effectiveness in
6. Termination Criteria: regulating dynamic systems. The PID controller operates by continu-
• The algorithm can terminate based on a predefined number of it- ously measuring the error between the desired set point and the actual
erations, convergence criteria, or computational budget. system output. It then adjusts the control signal based on three terms:
proportional (P), integral (I), and derivative (D). This section explains
3.4. Pareto dominance and crowding distance the principles behind each term and their contributions to the overall
control action [58].
In MOPSO, Pareto dominance is used to compare solutions. A solu-
tion A is said to dominate solution B if A is no worse than B in all ob- 4.1. Proportional (P) term
jectives and is strictly better in at least one objective. The concept of
crowding distance is often employed to maintain diversity in the Pareto The proportional term is directly proportional to the error between

7
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 4. Pareto curve resulting from the MOPSO algorithm for the optimal setting of the BI-AFTMD system.

the set point and the process output. It provides an immediate response Output(t) is the control signal output at time t.
to the current error and determines the control signal proportionally. Error (t) is the current error (set point - process output) at time t.
The P term helps reduce steady-state errors but may result in overshoot KP is the proportional gain, which determines the strength of the P
or oscillations. term.
KI is the integral time constant, defining the responsiveness of the I
4.2. Integral (I) term term.
KD is the derivative time constant, regulating the influence of the D
The integral term considers the accumulated error over time. It in- term.
tegrates the error to eliminate steady-state errors, ensuring that the Tuning a PID controller is crucial for realizing optimal system per-
system reaches the desired set point. The term provides long-term sta- formance. Various methods exist, including manual tuning, the Zie-
bility but can introduce overshoot and a sluggish response if not gler–Nichols method, and optimization algorithms. Fig. 3 represents the
appropriately tuned. implementation of the PID controller for the seismic control of a
building equipped with a passive FTMD.
4.3. Derivative (D) term For the optimal design of the PID strategy, the following objective
functions are considered:
The derivative term calculates the rate of change of the error. It ⎧
anticipates future error trends and counteracts rapid changes in the ⎪
⎪
⎪ Find : x = φPID
d
⎪
⎨ ( )
system. The D term improves system response time, reduces overshoot,
minimize : f(x) = max(max ‖xi (t)‖), and max max ‖ẍi (t)‖ i = 1, ⋯, N
and enhances stability. However, excessive D gain can amplify noise and ⎪
⎪
⎪
⎪
introduce instability. ⎩ Subjcted to : L ≤ φPID ≤ U
d
The control signal output of a PID controller is calculated as the sum
(14)
of the individual contributions from the P, I, and D terms. The formula is
expressed as follows [58–60]: The vector of design parameters for the proposed PID controller is
T 8 9
d = [KP , KI , KD ] . The intervals 0 ≤ KP ≤ 10 , 0 ≤ KI ≤ 10 , and 0 ≤ KD
φPID
[ ( ) ∫
1 7
Output (t) =KP × Error (t) + × [0, t] Error (dt) + KD ≤ 10 were taken into consideration for the lower bound, L, and upper
K1
] (13) bound, U, of the PID parameters. The parameters ‖xi(t)‖ and ‖ ẍi (t) ‖
d (Error (t))
× represent the displacement and acceleration, respectively, of the struc-
dt ture. The primary objective of the study is to minimize the simultaneous
maximum values of these two parameters. This is achieved by
where:

8
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Table 2 optimal PID controller parameters for the studied earthquakes can be
Optimal parameters of the PID controller for the studied earthquakes. found in Table 2. Additionally, Table 3 shows the average of the optimal
Earthquakes KP KI KD PID parameters for the studied earthquakes, which are considered
design criteria.
El Centro 4.2090 e + 05 2.0327 e + 08 4.6412 e + 06
Coalinga 6.0995 e + 06 8.4622 e + 08 9.6279 e + 06
Loma Prieta 3.5954 e + 06 4.4860 e + 08 4.0858 e + 06 5. Details of example building and optimized system parameters
Northridge 3.4152 e + 07 3.1124 e + 08 9.5734 e + 06 BI and PFTMD

For the numerical evaluation of the BI, PFTMD, BI-PFTMD, and BI-
Table 3 AFTMD systems, as well as the effectiveness of the MOPSO algorithms,
Average optimal parameters of the PID controller for the studied earthquakes. a realistic 10-storey building located in Mashhad, Iran, was selected as
Earthquakes KP KI KD the case study. The mass and stiffness matrices of the building were
El Centro 11066950 452325000 6982075
calculated using the matrix analysis procedure. The damping matrix was
Coalinga assumed to be proportional to the mass and stiffness matrices, using the
Loma Prieta Rayleigh method. To determine the proportionality coefficients, the
Northridge modal damping ratios of the first two modes were set to approximately 2
% of the critical value. The building is a steel structure with a braced
frame system [61]. The BI system is strategically positioned beneath the
Table 4 building, while the PFTMD system is installed on the highest floor. This
Natural frequencies of the example building for the first eight vibrational modes.
Table 5
Building 10-storey Building 10-storey
Structural parameters of the 10-storey building [61].
1st frequency (rad/s) 9.1893 5thfrequency (rad/s) 68.1363
Storey (hi , i = 1, …, (mi, i=1, (ci ,i = 1,…,10) (ki ,i = 1,…,10)
2nd frequency (rad/s) 24.2486 6th frequency (rad/s) 79.2101
10) …,10)
3rd frequency (rad/s) 39.3724 7th frequency (rad/s) 89.3738
4th frequency (rad/s) 54.5978 8th frequency (rad/s) 98.8097 1 3 381200 2.6280 e + 06 1.1627 e+ 09
2 3 379000 2.7451 e + 06 1.2161 e+ 09
3 3 380600 2.5362 e + 06 1.1210 e+ 09
optimizing the parameters of the PID controller, which are specifically 4 3 274300 2.4447 e + 06 1.0883 e+ 09
tailored for the analyzed earthquakes. Fig. 4 shows the Pareto curve of 5 3 280300 2.4458 e + 06 1.0883 e+ 09
6 3 278300 2.0919 e + 06 0.9276 e+ 09
the BI-AFTMD system in the studied earthquakes, a graphical repre- 7 3 276300 1.9518 e + 06 0.8641 e+ 09
sentation that elucidates the trade-off relationships among various 8 3 276300 1.7817 e + 06 0.7868 e+ 09
objective functions attained through the utilization of the multi- 9 3 275300 1.5063 e + 06 0.6617 e+ 09
objective particle swarm optimization algorithm. The purpose of this 10 3 274500 1.5062 e + 06 0.6617 e+ 09
algorithm is to determine the optimal settings for the BI-AFTMD system. hi: Storey height (m).
The Pareto curve demonstrates the collection of optimal solutions that mi: Storey mass (kg).
strike an optimal balance between conflicting objectives, highlighting ci: Storey damping (N.s/m).
the most favorable trade-offs. Detailed information regarding the ki: Storey stiffness (N/m).

Fig. 5. Mode shapes of the building for the first eight vibrational modes.

9
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 6. Time histories of earthquakes used in the study.

Table 6
List of earthquake events used in this study.
Earthquakes Statian Year Fault type Magnitude PGA (g) PGV (cm/s) Components Depth (km)

El Centro Imperial valley 1940 Far-field 6.95 0.348 43 NS 10

Coalinga Parkfield - Cholame 1983 Far-field 6.70 0.098 15 NS 13
Loma Prieta Gilroy Array #3 1989 Near-field 6.93 0.642 115 EW 19
Northridge Beverly Hills - 14145 1994 Near-field 6.69 0.433 75 NS 20

where PGA is the peak ground acceleration and PGV is the peak ground velocity.

Table 7
Optimum parameters of the utilized devices.
Systems Optimum parameters Earthquakes

El Centro Coalinga Loma Prieta Northridge

BI Topt
b
1.2914 1.9154 1.5168 1.9213
ξopt
b
0.1972 0.1933 0.1532 0.0171
PFTMD λ 0.0288 0.0447 0.0458 0.0215
fdopt 1.1501 0.8907 0.9941 0.9119
ξopt
d
0.0418 0.1751 0.1287 0.1809
μopt
d
0.1522 0.0271 0.2171 0.0100
BI-PFTMD Topt
b
1.2734 0.0100 1.4207 0.1412
ξopt
b
0.1709 0.1729 0.1616 0.2000
λ 0.0322 0.0500 0.0466 0.0500
opt
fd 0.7900 0.8760 1.0001 0.7047
ξopt
d
0.0791 0.1701 0.0881 0.2000
μopt
d
0.2170 0.0380 0.2712 0.0100

Table 8
Average optimum parameters of the utilized devices.
Earthquakes BI PFTMD BI-PFTMD

Tbopt ξopt
d
λ fdopt ξopt
d μopt
d Topt
b ξopt
d
λ fdopt ξopt
d μopt
d

El Centro
Coalinga
Loma Prieta 1.6619 0.1402 0.0352 0.9867 0.1316 0.1016 0.7113 0.1764 0.0447 0.8427 0.1343 0.1341
Northridge

10
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 7. Pareto curves resulting from the MOPSO algorithm for the optimal setting of the BI, PFTMD and, BI-PFTMD systems.

analysis yielded the natural frequencies of the structure and their cor- systems.
responding mode shapes. The first eight natural frequencies of the When solving a multi-objective optimization problem, it is expected
building are presented in Table 4, and the corresponding mode shapes that the obtained solutions will effectively represent the Pareto front,
are visually depicted in Fig. 5. For detailed information regarding the which demonstrates the trade-off between conflicting objectives [55].
specific characteristics of the 10-storey building, refer to Table 5. However, in practical problem-solving scenarios, it is often observed
The selection of this particular building as the case study allows for a that the solutions tend to be concentrated in a limited region of the
realistic and comprehensive evaluation of the proposed systems and Pareto front. To address this issue, it is necessary to strike a balance
optimization techniques. The detailed analysis of the building’s struc- between the convergence (closeness to the true Pareto front) and dis-
tural properties, including the mass, stiffness, and damping matrices, tribution (spread along the Pareto front) of the Pareto optimal solutions
provides a solid foundation for understanding the dynamic behavior of [56,57]. In the present study, the MOPSO algorithm is employed to
the structure and the effectiveness of the proposed control systems. By identify the optimal parameters for the PFTMD, BI, and BI-PFTMD sys-
presenting the natural frequencies and mode shapes, the research en- tems. The main objective is to address a two-objective optimization
ables a deeper understanding of the building’s vibrational characteris- problem, which involves the simultaneous minimization of both the
tics, which is crucial for the design and implementation of the control peak displacement and acceleration of the structure. By utilizing the

11
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Table 9
Peak seismic responses and the corresponding percentage reductions.
Earthquakes Uncontrolled BI PFTMD BI-PFTMD BI-AFTMD BI PFTMD BI-PFTMD BI-
AFTMD

Far-field El Centro Peak displacement (cm) Reduction (%)

21.05 8.50 9.16 8.07 6.85 60 56 62 67
Peak acceleration (g)
1.92 0.49 1.05 0.47 0.41 74 45 76 79
Peak inter-storey drift (%)
0.94 0.14 0.43 0.14 0.12 85 54 85 87
Coalinga Peak displacement (cm) Reduction (%)
7.05 2.53 2.26 2.00 0.95 64 68 72 87
Peak acceleration (g)
0.65 0.13 0.28 0.28 0.17 80 57 57 74
Peak inter-storey drift (%)
0.32 0.02 0.10 0.10 0.06 94 69 69 81
Near-field Loma Prieta Peak displacement (cm) Reduction (%)
25.24 9.79 10.54 9.47 7.22 61 58 62 71
Peak acceleration (g)
2.90 1.00 2.05 0.96 0.82 66 29 67 72
Peak inter-storey drift (%)
1.19 0.15 0.53 0.20 0.18 87 55 83 85
Northridge Peak displacement (cm) Reduction (%)
18.63 18.05 12.30 12.34 6.11 3 34 34 67
Peak acceleration (g)
1.47 0.63 1.30 1.18 0.62 57 12 20 58
Peak inter-storey drift (%)
0.87 0.15 0.65 0.61 0.36 83 26 30 59

Table 10
Peak seismic responses and percentage reductions based on average optimal parameters.
Earthquakes Uncontrolled BI PFTMD BI-PFTMD BI-AFTMD BI PFTMD BI-PFTMD BI-
AFTMD

Far-field El Centro Peak displacement (cm) Reduction (%)

21.05 11.74 10.35 12.48 4.55 44 51 41 78
Peak acceleration (g)
1.92 0.43 1.29 0.84 0.66 78 33 56 66
Peak inter-storey drift (%)
0.94 0.13 0.43 0.35 0.19 86 54 63 80
Coalinga Peak displacement (cm) Reduction (%)
7.05 3.46 3.96 4.07 1.45 51 44 42 79
Peak acceleration (g)
0.65 0.15 0.41 0.28 0.23 77 37 57 65
Peak inter-storey drift (%)
0.32 0.04 0.15 0.13 0.07 88 53 60 78
Near-field Loma Prieta Peak displacement (cm) Reduction (%)
25.24 12.16 14.52 12.85 7.51 52 42 49 70
Peak acceleration (g)
2.90 0.95 2.46 1.39 1.09 67 15 52 62
Peak inter-storey drift (%)
1.19 0.13 0.70 0.41 0.32 89 41 66 73
Northridge Peak displacement (cm) Reduction (%)
18.63 26.00 18.28 25.23 10.19 -40 2 -35 45
Peak acceleration (g)
1.47 0.65 1.46 1.03 0.57 56 1 30 61
Peak inter-storey drift (%)
0.87 0.28 0.86 0.72 0.30 68 1 17 66

MOPSO algorithm, this study aims to find the optimal solutions that The optimization problem is defined for the optimum design of the
strike a balance between minimizing peak displacement and accelera- BI, PFTMD, and BI-PFTMD systems parameters, as follows:
tion, thereby enhancing the overall structural performance. The MOPSO ⎧
algorithm is well-suited for this task as it is designed to effectively opt opt opt opt opt
⎪ Find : x = λ, fd , ξd , μd , ξb and Tb
⎪
⎪
⎪ ( )
explore the trade-offs between conflicting objectives. By maintaining a ⎪
⎪
⎪
⎨ minimize : f(x) = max(max ‖x (t)‖), and max max ‖ẍ (t)‖ i = 1, ⋯, N
diverse set of Pareto optimal solutions, the MOPSO approach can help i i

overcome the tendency for solutions to concentrate in a limited region of ⎪

⎪
⎪
⎪ Subjcted to : 0.01 ≤ γ ≤ 0.05, 0.6 ≤ fd ≤ 1.4, 0.01 ≤ ξd ≤ 0.2,
⎪
the Pareto front, which is a common challenge in multi-objective opti- ⎪
⎪
⎩ and 0.01 ≤ μd ≤ 0.5, 0.01 ≤ ξb ≤ 0.2, 0.01 ≤ Tb ≤ 2
mization problems. Through the application of the MOPSO algorithm,
the study aims to identify a set of optimal design parameters for the (15)
PFTMD, BI, and BI-PFTMD systems that provide the best compromise
The PFTMD optimal parameters are the mass ratio (λ), frequency coef-
between minimizing peak displacement and acceleration. This balanced
ficient (fdopt ), damping coefficient (ξopt opt
d ), and friction coefficient (μd ). The
approach is crucial for improving the structural performance and
ensuring the resilience of the building under various loading conditions. optimal parameters of the BI system are the damping coefficient (ξopt b ) and

12
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 8. Responses of peak displacement, acceleration, and inter-storey drift of the building during different seismic events.

natural period (Tbopt ). where max ‖xi(t)‖ and max ‖ ẍi (t) ‖ represent the ⎧
( )2
maximum displacement and acceleration of the i-th floor, respectively. ⎪
⎪
⎪ 2π
The PFTMD system parameters are mass (md), damping (cd), stiff-
⎪
⎪
⎪ kb = m t ×
⎪
⎪
⎪ Tb
ness (kd), and slip force ( fs). Moreover, the BI system parameters are ⎪
⎪
⎪
⎪ ( )
stiffness (kb), damping (cb), and mass (mb), and g is the gravitational
⎪
⎪ 2π
⎨ cb = 2mt
⎪ ξ
acceleration. (BI) → Tb b (17)
⎪
In Eq. (17), mt is the mass of all storeys of the building together with ⎪
⎪
⎪ mb = m1
⎪
the mass of the BI slab, and ωs is the fundamental frequency of the pri- ⎪
⎪
⎪
⎪
⎪ N
mary structure. ⎪
⎪ ∑
⎪ mt =
⎪
⎪ mi + mb
⎧ ⎪
⎩ i=1
⎪
⎪
⎪
⎪
⎪ md = λmt
⎪
⎨ cd = 2md ξopt
⎪ opt
d fd ωs
Similar to other heuristic optimization techniques, it is important to
(PFTMD) → ( opt )2 (16) tune the parameters of the MOPSO algorithm to achieve sensible and
⎪
⎪ cd = md fd ωs
⎪
⎪
⎪ opt
effective results. The key parameters that need to be set include the
⎩ fs = μd md g
⎪
⎪ population size (n) and the exploration-exploitation balance, which can
be controlled through the inertia weight (w) and the cognitive (c1) and
social (c2) acceleration coefficients. A larger population size (n) allows
for a more comprehensive exploration of the search space, which can be

13
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 9. Time responses of peak displacement, acceleration, inter-storey drift of the structure and diagram of active friction force during the El Centro earthquake.

beneficial for finding the Pareto optimal solutions that effectively This study focuses on selecting a design vector that minimizes the
represent the trade-off between the conflicting objectives. However, peak displacement of the floor while ensuring that the peak acceleration
increasing the population size also leads to higher computational efforts remains within acceptable limits. Based on these criteria, the optimal
and longer running times. Based on the experience of the authors and the design of the BI, PFTMD, and combined BI-PFTMD systems is deter-
complexity of the optimization problems at hand, a population size of n mined. By employing the MOPSO optimization process, the optimal
= 35 is selected for the PFTMD, BI, and BI-PFTMD systems. The inertia parameters for the BI, PFTMD, and BI-PFTMD systems in response to the
weight (w) controls the balance between the exploration and exploita- El Centro, Coalinga, Loma Prieta, and Northridge earthquakes are
tion capabilities of the MOPSO algorithm. A larger inertia weight pro- identified. Table 7 provides an overview of the optimal parameters
motes global exploration, while a smaller inertia weight encourages specifically tailored for the BI, PFTMD, and BI-PFTMD systems. These
local exploitation. In this study, the inertia weight is set to w = 0.8, parameters were meticulously adjusted to enhance the overall perfor-
which provides a good balance between exploration and exploitation. mance and functionality of the systems. To establish a rational design
The cognitive acceleration coefficient (c1) and the social acceleration criterion, the average values of the optimal parameters are computed
coefficient (c2) determine the relative influence of the particle’s own and presented in Table 8.
best position and the swarm’s best position, respectively, on the parti- Next, the proposed optimal PID controller, discussed in detail in the
cle’s movement. A higher c1 value encourages individual exploration, previous section, is implemented. By integrating the PID controller into
while a higher c2 value promotes social learning and convergence to- the control system, the aim is to significantly amplify the performance
wards the global best solution. For the current optimization problems, and effectiveness of the structure’s response, leading to improved
the values of c1 and c2 are set to 2.0 and 2.0, respectively, which is a overall system behaviour. Fig. 7 visually presents the Pareto curves
common and effective combination for MOPSO algorithms. By setting derived from the MOPSO algorithm, which was specifically applied to
these parameter values, the MOPSO algorithm is expected to effectively the BI, PFTMD, and BI-PFTMD systems.
explore the search space, maintain a diverse set of Pareto optimal so-
lutions, and strike a balance between minimizing peak displacement and 6. Simulation results
acceleration for the PFTMD, BI, and BI-PFTMD systems.
In this study, the El Centro, Coalinga, Loma Prieta, and Northridge Tables 9 and 10 provide a comprehensive comparison of the seismic
earthquakes were specifically analyzed. The MOPSO computational al- structural responses between the uncontrolled structure and different
gorithm was utilized to determine the optimal settings for the BI system, systems, namely BI, PFTMD, BI-PFTMD, and BI-AFTMD. These responses
PFTMD system, and combined BI-PFTMD systems. Fig. 6 displays the encompass peak displacement, acceleration, and inter-storey drift
time histories of the earthquakes used in this study, providing a visual values. Table 9 presents the seismic responses obtained when optimizing
representation of the ground motion records. Additionally, Table 6 the parameters of each system separately, taking into account the spe-
presents the characteristics and attributes of the earthquakes investi- cific characteristics of each earthquake considered in the study. Table 10
gated in the study. shows the responses obtained by utilizing the average values of the

14
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 10. Time responses of peak displacement, acceleration, inter-storey drift of the structure and diagram of active friction force during the Coalinga earthquake.

optimized parameters. These average values are used as design co- those obtained when each system is individually optimized for a specific
efficients in this particular study. The analysis conducted in the study is earthquake scenario.
based on the data presented in Table 10, which provides a more Based on the analysis of Tables 9 and 10, it is evident that structures
generalized approach by incorporating the averaged optimized param- equipped with BI, PFTMD, BI-PFTMD, and BI-AFTMD systems exhibit a
eters. By utilizing these design coefficients, this study aims to assess the significant reduction in peak displacement, acceleration, and inter-
seismic performance of the structures under consideration. storey drift during the Loma Prieta earthquake as compared to the un-
Based on the findings presented in Tables 9 and 10, the structures controlled conditions. By comparing the results presented in Tables 9
equipped with BI (Base Isolated), PFTMD (Passive Friction TMD), BI- and 10, it can be concluded that when the systems are individually
PFTMD (Combined Base Isolated and Passive Friction TMD), and BI- optimized for each earthquake scenario, the BI-AFTMD system demon-
AFTMD (Combined Base Isolated and Active Friction TMD) systems strates superior performance in reducing peak displacement and peak
demonstrates a significant decrease in displacement, acceleration, and acceleration. In contrast, the BI system shows better results in mini-
inter-storey drift when subjected to the El Centro earthquake as mizing inter-storey drift as compared to the other systems. However, if
compared to the uncontrolled state. Upon comparing the results of Ta- the average of the optimal values is used, the findings differ slightly. In
bles 9 and 10, it is evident that when the systems are optimized indi- this case, the BI-AFTMD system performs better than the other systems
vidually for each earthquake scenario, the BI-AFTMD system exhibited in reducing peak displacement, while the BI system outperforms the
superior performance in reducing peak displacement, acceleration, and other systems in reducing both peak acceleration and inter-storey drift.
inter-storey drift as compared to the other systems. However, if the The analysis of Table 9 reveals that structures equipped with BI,
optimal values are derived using the average, the BI-AFTMD system PFTMD, BI-PFTMD, and BI-AFTMD systems exhibit a significant reduc-
outperforms only the other systems in reducing the peak displacement of tion in displacement, acceleration, and inter-storey drift during the
the structure. In contrast, the BI system performs better than the other Northridge earthquake as compared to those under uncontrolled con-
systems in reducing peak acceleration and inter-storey drift. ditions. However, according to the results presented in Table 10, during
The analysis of Tables 9 and 10 highlights that structures with BI, the Northridge earthquake, the peak displacement of the structure
PFTMD, BI-PFTMD, and BI-AFTMD systems show significant reductions equipped with both BI systems and a combination of BI and BI-PFTMD
in displacement, acceleration, and inter-storey drift during the Coalinga systems increased as compared to that under uncontrolled conditions.
earthquake as compared to the uncontrolled conditions. When the sys- In contrast, the remaining systems show better performance for different
tems are individually optimized for each earthquake scenario, the BI- response parameters as compared to the uncontrolled state. After indi-
AFTMD system is superior to the other systems in reducing the peak vidual optimization for each specific earthquake scenario, the BI-
displacement, while the BI system is better than the other systems in AFTMD system performs better than the other systems in reducing
minimizing the peak acceleration and inter-storey drift. However, if the displacement and peak acceleration. In contrast, the BI system shows
average of the optimal values is used, the overall findings align with superior performance in minimizing interclass drift as compared to

15
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 11. Time responses of peak displacement, acceleration, inter-storey drift of the structure and diagram of active friction force during the Loma Prieta earthquake.

other systems. However, when considering the average of the optimal most effective in reducing peak acceleration and inter-storey drift, while
values, the overall findings remain consistent with the results obtained the BI-AFTMD system is best at reducing peak displacement. For the
from the optimization of each system for each earthquake scenario. Loma Prieta earthquake, the bar charts once more highlight the superior
Based on the findings derived from Tables 9 and 10, it is evident that the performance of the controlled structures over the uncontrolled struc-
integration of the BI and PFTMD systems resulted in a well-balanced ture. The BI system achieves the best reduction in peak acceleration and
performance across all the conducted analyses. The combination of inter-storey drift, while the BI-AFTMD system is the most effective at
these systems, known as the BI-PFTMD system, capitalizes on the ad- reducing peak displacement. Lastly, the Northridge earthquake results
vantages offered by both base isolation and passive friction-tuned mass depicted in Fig. 8 show mixed outcomes. Although the PFTMD and BI-
dampers, effectively mitigating various response parameters. This AFTMD systems reduce the measured response parameters as
achievement underscores the efficacy of the system in enhancing compared to the uncontrolled structure, the BI and combined BI-PFTMD
structural performance and resilience. In Fig. 8, the seismic structural systems actually increase peak displacement under the Northridge
responses of an uncontrolled structure are depicted alongside structures earthquake. Overall, the bar charts provide a thorough visual repre-
equipped with various systems, namely, BI, PFTMD, BI-PFTMD, and BI- sentation of the seismic performance of both the uncontrolled structure
AFTMD. These results are based on the design parameters outlined in and the various control systems, illustrating the benefits of implement-
Table 10. The measured responses encompass the peak displacement, ing BI, PFTMD, BI-PFTMD, and BI-AFTMD systems to enhance structural
acceleration, and inter-storey drift. resilience.
To derive comprehensive results, the structural responses to the The findings depicted in Figures 9–11, and 12 showcase the time
analyzed earthquakes are depicted in Fig. 8, focusing on peak history responses of a structure subjected to four different earthquakes
displacement, acceleration, and inter-storey drift. These results include (El Centro, Coalinga, Loma Prieta, and Northridge). These responses are
five scenarios: the uncontrolled structure and four different control based on the use of average optimal parameters for various control
systems, each optimized based on average parameters. The bar charts systems. The figures visually present the peak displacement, accelera-
visually compare the seismic performance of the uncontrolled structure tion, and inter-storey drift for both the uncontrolled structure and the
with that of the structures equipped with the various control systems. structure equipped with different control systems, including the BI,
For the El Centro earthquake, the charts reveal that the BI, PFTMD, BI- PFTMD, BI-PFTMD, and BI-AFTMD systems. Analysis of these results
PFTMD, and BI-AFTMD systems all result in substantial reductions in reveals a significant reduction in the average peak displacement, ac-
peak displacement, acceleration, and inter-storey drift as compared to celeration, and inter-storey drift for structures equipped with these
the uncontrolled structure. The BI system demonstrates the best per- control systems as compared to the uncontrolled structure. Specifically,
formance in reducing peak acceleration and inter-storey drift, while the the average peak displacement reductions are as follows: BI system: 27
BI-AFTMD system excels in minimizing peak displacement. The results %, PFTMD system: 35 %, BI-PFTMD system: 24 %, and BI-AFTMD sys-
for the Coalinga earthquake show similar patterns, with controlled tem: 68 %. The average peak acceleration reductions are BI system: 70
structures significantly lowering the measured response parameters as %, PFTMD system: 22 %, BI-PFTMD system: 49 %, and BI-AFTMD sys-
compared to the uncontrolled structure. Here again, the BI system is tem: 64 %. The average peak inter-storey drift reductions are BI system:

16
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 12. Time responses of peak displacement, acceleration, inter-storey drift of the structure and diagram of active friction force during the Northridge earthquake.

83 %, PFTMD system: 37 %, BI-PFTMD system: 52 %, and BI-AFTMD 7. Conclusions

system: 74 %. Furthermore, the control force diagram for the BI-
AFTMD system is presented, illustrating the amount of control force This study introduced an innovative seismic control strategy known
generated by the system, including the sum of active control force and as the BI-AFTMD system, which integrates a BI system, a passive friction-
friction force during the earthquake. These results underscore the tuned mass damper (PFTMD), and a PID controller. Additionally,
effective mitigation of structural response during seismic events through optimal passive designs for the PFTMD, BI, and BI-PFTMD systems were
the implementation of the BI, PFTMD, BI-PFTMD, and BI-AFTMD con- developed using design procedures based on the MOPSO algorithm. The
trol systems, leading to significant reductions in displacements, accel- seismic performance of the BI, PFTMD, BI-PFTMD, and BI-AFTMD sys-
erations, and inter-storey drifts as compared to the uncontrolled tems was evaluated on a 10-storey structure subjected to both far-field
structure. and near-field earthquakes. These investigations included a compara-
Fig. 13 illustrates the active friction control force of the BI-AFTMD tive analysis of the proposed systems for two optimal design strategy
system in terms of peak displacement and acceleration, based on the modes tailored to each earthquake type and the use of average optimal
earthquakes analyzed in this study. The depicted forces encompass the parameters. The results of these evaluations led to the following
combined effects of friction force and active control force. These dia- conclusions:
grams offer valuable insights into the relationship between peak
displacement, peak acceleration, and the necessary control forces to a) The unpredictable nature of future earthquakes, in terms of both
mitigate the structure’s displacement and acceleration. Design pro- intensity and frequency content, makes it challenging to design
fessionals can leverage these diagrams to determine the required control structural control systems that can effectively mitigate the structural
forces for dampers and isolators, based on the anticipated displacements responses across a wide range of seismic events. Given that the
and accelerations of the structure. By analyzing these graphs, one can optimal design parameters for each control system are specific to the
assess the effectiveness of the BI-AFTMD system in controlling characteristics of a particular earthquake, it is not feasible to expect
displacement and acceleration and determine if performance improve- these systems to perform well in reducing structural responses during
ments are needed. Additionally, a comparative analysis of different other earthquakes. Therefore, the use of average optimal parameters
earthquakes reveals that the BI-AFTMD system requires a more sub- for each control system can be considered a more practical and
stantial active friction control force to mitigate the structural response logical design approach. Additionally, the design of a single control
during the Northridge earthquake. Conversely, it demands the least system that demonstrates good performance when exposed to
active friction control force for the Coalinga earthquake. This variation various earthquake scenarios can be viewed as an ideal optimal
is attributed to the observed peak displacement and peak acceleration design. Furthermore, a comparative seismic evaluation of passive
relative to other earthquakes. and hybrid control systems can provide valuable guidance to engi-
neers and designers in selecting the most appropriate system to
effectively reduce seismic responses.

17
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

Fig. 13. Hysteresis loop diagram of the BI-AFTMD system according to peak displacement and acceleration of the structure under the studied earthquakes.

b) The study examined the effectiveness of implementing various displacements, accelerations, and inter-storey drifts during seismic
structural control systems - BI, PFTMD, BI-PFTMD, and BI-AFTMD in events. This suggests that the BI, PFTMD, BI-PFTMD, and BI-AFTMD
reducing the structural responses of a building subjected to earth- systems have the potential to enhance the seismic performance of
quakes. Considering both the optimal parameter design scenarios for structures and improve their resilience against earthquakes.
each earthquake and the average of the optimal designs in different d) While the results presented in this study are valuable for the analyzed
earthquakes, the results demonstrated that these control systems structural system, there are certain limitations. This study focused on
significantly reduce the average peak displacement, acceleration, a structure with one lateral degree of freedom at each storey level
and inter-storey drift compared to the uncontrolled structure. Spe- and considered only four ground motions. Future research should
cifically, the BI-AFTMD system achieved the greatest reduction in include a broader range of earthquake records and different struc-
average peak displacement, with a 71 % decrease, and in average tures in terms of materials and building heights to further analyze
peak acceleration, with a 67 % decrease. For average peak inter- and validate the findings. Nevertheless, the evaluated seismic per-
storey drift, the BI system achieved the greatest reduction, with an formances of the various control systems provide valuable insights
85 % decrease. These findings underscore the efficacy of structural for designing effective and efficient strategies to mitigate the
control systems, particularly the BI-AFTMD system, in mitigating the damaging effects of earthquakes on structures.
adverse effects of earthquakes on building structures. The study
contributes to the advancement of earthquake-resistant design and These conclusions highlight the effectiveness and potential of the
the development of effective strategies for enhancing the safety and proposed BI-AFTMD control strategy, as well as the importance of
resilience of buildings in seismic-prone regions. optimizing the design of passive control systems for enhancing seismic
c) These findings indicated that the implementation of these systems performance.
effectively mitigates the structural response, resulting in reduced

18
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

8. Future research directions References

To advance and broaden the findings of this study, future research [1] Li Z, Ma R, Xu K, Han Q. Closed-form solutions for the optimal design of lever-arm
tuned mass damper inerter (LTMDI). Mech Syst Signal Process 2024;206:110889.
should focus on several key areas: https://doi.org/10.1016/j.ymssp.2023.110889.
[2] Wang L, Zhou Y, Shi W. Seismic multi-objective stochastic parameters optimization
1. Advanced Optimization Algorithms and Robust Controller of multiple tuned mass damper system for a large podium twin towers structure.
Soil Dyn Earthq Eng 2024;177:108428. https://doi.org/10.1016/j.
Systems: soildyn.2023.108428.
• Investigate sophisticated optimization algorithms and robust [3] Kandemir EC, Mortazavi A. Optimization of seismic base isolation system using a
controller systems to enhance the design process of seismic control fuzzy reinforced swarm intelligence. Adv Eng Softw 2022;174:103323. https://doi.
org/10.1016/j.advengsoft.2022.103323.
systems. [4] Abdulhadi M, Xun’an Z, Atroshchenko E, Rungamornrat J. Vibration control of
• Develop and test algorithms that can potentially improve the inerter-enhanced mega sub-controlled structure system (MSCSS) and the reliability
performance and efficiency of the control strategies examined in analysis of the structure under seismic action. Eng Struct 2024;304:117508.
https://doi.org/10.1016/j.engstruct.2024.117508.
this study.
[5] Smarsly K, Law KH. Decentralized fault detection and isolation in wireless
2. Diverse Structural Configurations: structural health monitoring systems using analytical redundancy. Adv Eng Softw
• Consider a wider range of structural configurations to generalize 2014;73. https://doi.org/10.1016/j.advengsoft.2014.02.005. 1-0.
the study’s findings. [6] Kim HS, Roschke PN. GA-fuzzy control of smart base isolated benchmark building
using supervisory control technique. Adv Eng Softw 2007;38(7):453–65. https://
• Include various structural types, materials, and building heights to doi.org/10.1016/j.advengsoft.2006.10.004.
ensure the applicability of control systems to different real-world [7] Ghodke S, Jangid RS. Equivalent linear elastic-viscous model of shape memory
scenarios. alloy for isolated structures. Adv Eng Softw 2016;99:1–8. https://doi.org/10.1016/
j.advengsoft.2016.04.005.
• Analyze how these configurations affect the effectiveness and [8] Zhang H, Wang L, Shi W. Seismic control of adaptive variable stiffness intelligent
adaptability of seismic control systems. structures using fuzzy control strategy combined with LSTM. J Build Eng 2023;78:
3. Practical Implementation: 107549. https://doi.org/10.1016/j.jobe.2023.107549.
[9] Kelly JM. Base isolation: linear theory and design. Earthq Spectra 1990;6(2):
• Conduct research on the practical aspects of implementing pro- 223–44. https://doi.org/10.1193/1.158556.
posed control systems, such as BI-AFTMD, into existing structures. [10] Mahmoud S, Jankowski R. Pounding-involved response of isolated and non-
• Examine the challenges and feasibility of integrating these sys- isolated buildings under earthquake excitation. Earthq Struct 2010;1(3):231–52.
https://doi.org/10.12989/eas.2010.1.3.23.
tems, focusing on structural compatibility, retrofitting re- [11] Zhang C, Duan C, Sun L. Inter-storey isolation versus base isolation using friction
quirements, and installation procedures. pendulum systems. Int J Struct Stabil Dyn 2024;24(2). https://doi.org/10.1142/
• Identify potential obstacles and develop solutions to facilitate the S0219455424500226.
[12] Ocak A, Işıkdağ Ü, Bekdaş G, Nigdeli SM, Kim S, Geem ZW. Prediction of damping
adoption of these control systems in engineering practice.
capacity demand in seismic base isolators via machine learning. CMES-Comput
4. Soil-Structure Interaction Effects: Model Eng Sci 2024;138(3). https://doi.org/10.32604/cmes.2023.030418.
• Study the effects of soil-structure interactions on the performance [13] Matsagar VA, Jangid RS. Impact response of torsionally coupled base-isolated
of control systems. structures. J Vibrat Control 2010;16(11):1623–49. https://doi.org/10.1177/
10775463091032.
• Develop models and simulations to understand how varying soil [14] Falborski T, Jankowski R. Numerical evaluation of dynamic response of a steel
conditions impact the effectiveness of seismic control measures. structure model under various seismic excitations. Procedia Eng 2017;172:277–83.
5. Cost-Effectiveness: https://doi.org/10.1016/j.proeng.2017.02.114.
[15] Falborski T, Jankowski R. Shaking table experimental study on the base isolation
• Analyze the cost-effectiveness of implementing advanced control system made of polymer bearings. In: InProceedings of the 15th World Conference
systems in real-world applications. on Earthquake Engineering, Paper; 2012. p. 24–8.
• Compare the costs and benefits to determine the economic [16] Ocak A, Nigdeli SM, Bekdaş G, Kim S, Geem ZW. Optimization of seismic base
isolation system using adaptive harmony search algorithm. Sustainability 2022;14
viability and encourage wider adoption in engineering projects. (12):7456. https://doi.org/10.3390/su14127456.
[17] Falborski T, Jankowski R. Advanced hysteretic model of a prototype seismic
By addressing these areas, future research can refine and generalize isolation system made of polymeric bearings. Appl Sci 2018;8(3):400. https://doi.
org/10.3390/app8030400.
the findings of this study, contributing to the development of more [18] Falborski T, Jankowski R. Experimental study on effectiveness of a prototype
efficient and adaptable seismic control systems for diverse engineering seismic isolation system made of polymeric bearings. Appl Sci 2017;7(8):808.
applications. https://doi.org/10.3390/app7080808.
[19] Dadkhah M, Kamgar R, Heidarzadeh H, Jakubczyk-Gałczyńska A, Jankowski R.
Improvement of performance level of steel moment-resisting frames using tuned
CRediT authorship contribution statement mass damper system. Appl Sci 2020;10(10):3403. https://doi.org/10.3390/
app10103403.
[20] Gewei Z, Basu B. A study on friction-tuned mass damper: harmonic solution and
Morteza Akbari: Writing – review & editing, Writing – original statistical linearization. J Vibrat Control 2011;17(5):721–31. https://doi.org/
draft, Visualization, Validation, Software, Resources, Methodology, 10.1177/1077546309354967.
Investigation, Formal analysis, Data curation, Conceptualization. [21] Etedali S, Akbari M, Seifi M. MOCS-based optimum design of TMD and FTMD for
tall buildings under near-field earthquakes including SSI effects. Soil Dyn Earthq
Mohammad Seifi: Writing – review & editing, Validation, Resources, Eng 2019;119:36–50. https://doi.org/10.1016/j.soildyn.2018.12.027.
Methodology. Tomasz Falborski: Writing – review & editing, Super- [22] Pisal AY. Seismic response of multi-story structure with multiple tuned mass
vision, Investigation, Formal analysis. Robert Jankowski: Writing – friction dampers. Int J Adv Struct Eng (IJASE) 2015;7:81–92. https://doi.org/
10.1007/s40091-014-0079-9.
review & editing, Supervision, Resources, Project administration, [23] Etedali S, Seifi M, Akbari M. A numerical study on optimal FTMD parameters
Investigation. considering soil-structure interaction effects. Geomech Eng 2018;16(5):527–38.
https://doi.org/10.12989/gae.2018.16.5.527.
[24] Chung LL, Wu LY, Lien KH, Chen HH, Huang HH. Optimal design of friction
pendulum tuned mass damper with varying friction coefficient. Struct Control
Declaration of competing interest Health Monit 2013;20(4):544–59. https://doi.org/10.1002/stc.514.
[25] Etedali S, Akbari M, Seifi M. Friction tuned mass dampers in seismic-excited high-
The authors declare that they have no known competing financial rise buildings with SSI effects: A reliability assessment. J Earthq Tsunami 2023;17
(02):2250022. https://doi.org/10.1142/S1793431122500221.
interests or personal relationships that could have appeared to influence [26] Etedali S, Seifi M, Akbari M. Failure probability of tall buildings with TMD in the
the work reported in this paper. presence of structural, seismic, and soil uncertainties. Struct Eng Mech 2023;85(3):
381–91. https://doi.org/10.12989/sem.2023.85.3.381.
[27] Pisal AY, Jangid RS. Vibration control of bridge subjected to multi-axle vehicle
Data availability
using multiple tuned mass friction dampers. Int J Adv Struct Eng (IJASE) 2016;8:
213–27. https://doi.org/10.1007/s40091-016-0124-y.
Data will be made available on request.

19
M. Akbari et al. Advances in Engineering Software 198 (2024) 103765

[28] Pisal AY, Jangid RS. Dynamic response of structure with tuned mass friction [44] Sohrabi MR, Etedali S, Zamni AA. Simultaneous optimal design of PID and MATMD
damper. Int J Adv Struct Eng 2016;8:363–77. https://doi.org/10.1007/s40091- Considering their Proposed Elevation Placement for Seismic Control of Tall
016-0136-7. Buildings on Soft Soil. Int J Indust Electron Control Optimiz 2023;6(4). https://doi.
[29] Jiang J, Ho SC, Markle NJ, Wang N, Song G. Design and control performance of a org/10.22111/ieco.2023.46162.1494.
frictional tuned mass damper with bearing–shaft assemblies. J Vibrat Control [45] Wang L, Zhao Y, Liu J. A Kriging-based decoupled non-probability reliability-based
2019;25(12):1812–22. https://doi.org/10.1177/1077546319832429. design optimization scheme for piezoelectric PID control systems. Mech Syst Signal
[30] Khatibinia M, Bijari S, Nezhad FS, Gharehbaghi S. Seismic vulnerability control of Process 2023;203:110714. https://doi.org/10.1016/j.ymssp.2023.110714.
steel moment–resisting frames using optimum friction tuned mass damper. Iran J [46] Borase RP, Maghade DK, Sondkar SY, Pawar SN. A review of PID control, tuning
Sci Technol Trans Civil Eng 2024:1–8. https://doi.org/10.1007/s40996-024- methods and applications. Int J Dyn Control 2021;9:818–27. https://doi.org/
01453-2. 10.1007/s40435-020-00665-4.
[31] Besharatian B, Riahi HT, Garcia R, Hajirasouliha I. A practical optimisation method [47] Etedali S, Zamani AA, Akbari M, Seifi M. A new seismic control framework of
for friction tuned mass dampers in multi-storey buildings subjected to earthquake optimal PIλDµ controller series with fuzzy PD controller including soil-structure
excitations. Soil Dyn Earthq Eng 2024;184:108857. https://doi.org/10.1016/j. interaction. J Franklin Inst 2023;360(14):10536–63. https://doi.org/10.1016/j.
soildyn.2024.108857. jfranklin.2023.08.003.
[32] Labbafi SF, Shooshtari A, Mohtashami E. Optimal design of friction tuned mass [48] Wang L, Zhou Y, Shi W. Random crowd-induced vibration in footbridge and
damper for seismic control of an integral bridge. Structures 2023;58:105200. adaptive control using semi-active TMD including crowd-structure interaction. Eng
https://doi.org/10.1016/j.istruc.2023.105200. Elsevier. Struct 2024;306:117839. https://doi.org/10.1016/j.engstruct.2024.117839.
[33] Yang JN, Danielians A, Liu SC. Aseismic hybrid control systems for building [49] Nikos GP, Charis JG. Influence of time delay and saturation capacity in control of
structures. J Eng Mech 1991;117(4):836–53. https://doi.org/10.1061/(ASCE) structures under seismic excitations. Smart Struct Syst. Int J 2011;8(5):479–90.
0733-9399(1991)117:4(836). https://doi.org/10.12989/sss.2011.8.5.449.
[34] Tsai HC. The effect of tuned-mass dampers on the seismic response of base-isolated [50] Pnevmatikos N. Pole placement algorithm for control of civil structures subjected
structures. Int J Solids Struct 1995;32(8-9):1195–210. https://doi.org/10.1016/ to earthquake excitation. J Appl Comput Mech 2017;3(1):25–36. https://doi.org/
0020-7683(94)00150-U. 10.22055/jacm.2017.12603.
[35] Stanikzai MH, Elias S, Matsagar VA, Jain AK. Seismic response control of base- [51] Xiang Y, Tan P, He H, Shang J, Zhang Y. Seismic optimization of variable friction
isolated buildings using multiple tuned mass dampers. Struct Des Tall Spec Build pendulum tuned mass damper with hysteretic damping characteristic. Soil Dyn
2019;28(3):e1576. https://doi.org/10.1002/tal.1576. Earthq Eng 2022;160:107381. https://doi.org/10.1016/j.soildyn.2022.107381.
[36] Gao H, Xing C, Wang H, Li J, Zhang Y. Performance improvement and demand- [52] McMillan AJ. A non-linear friction model for self-excited vibrations. J Sound Vib
oriented optimum design of the tuned negative stiffness inerter damper for base- 1997;205(3):323–35. https://doi.org/10.1006/jsvi.1997.1053.
isolated structures. J. Build. Eng. 2023;63:105488. https://doi.org/10.1016/j. [53] Misovec KM. Friction compensation using adaptive non-linear control with
jobe.2022.105488. persistent excitation. Int J Control 1999;72(5):457–79. https://doi.org/10.1080/
[37] Naderpour H, Naji N, Burkacki D, Jankowski R. Seismic response of high-rise 002071799221091.
buildings equipped with base isolation and non-traditional tuned mass dampers. [54] Matta E, Greco R. Modeling and design of tuned mass dampers using sliding
Appl. Sci. 2019;9(6):1201. https://doi.org/10.3390/app9061201. variable friction pendulum bearings. Acta Mech 2020;231:5021–46. https://doi.
[38] Akbari M, Zand JP, Falborski T, Jankowski R. Advanced seismic control strategies org/10.1007/s00707-020-02801-9.
for smart base isolation buildings utilizing active tendon and MR dampers. Eng [55] Mirjalili S, Dong JS, Mirjalili S, Dong JS. Multi-objective particle swarm
Struct 2024;318:118756. https://doi.org/10.1016/j.engstruct.2024.118756. optimization. Multi-Object Optimiz Artif Intell Tech 2020:21–36. https://doi.org/
[39] Kontoni DP, Farghaly AA. The effect of base isolation and tuned mass dampers on 10.1007/978-3-030-24835-2.
the seismic response of RC high-rise buildings considering soil-structure [56] Kumar V, Minz S. Multi-objective particle swarm optimization: an introduction.
interaction. Earthq Struct 2019;17(4):425–34. https://doi.org/10.12989/ SmartCR 2014;4(5):335–53. https://doi.org/10.6029/smartcr.2014.05.001.
eas.2019.17.4.425. [57] Li M, Chen H, Wang X, Zhong N, Lu S. An improved particle swarm optimization
[40] Wang L, Nagarajaiah S, Shi W, Zhou Y. Seismic performance improvement of base- algorithm with adaptive inertia weights. Int J Inf Technol Decis Mak 2019;18(03):
isolated structures using a semi-active tuned mass damper. Eng Struct 2022;271: 833–66. https://doi.org/10.1142/S0219622019500147.
114963. https://doi.org/10.1016/j.engstruct.2022.114963. [58] Ogata K. Modern control engineering. India: Prentice Hall; 2009.
[41] Etedali S, Sohrabi MR, Tavakoli S. Optimal PD/PID control of smart base isolated [59] Bishop R.C. Modern control systems. 2011.
buildings equipped with piezoelectric friction dampers. Earthq Eng Eng Vibration [60] Åström KJ, Murray R. Feedback systems: an introduction for scientists and
2013;12:39–54. https://doi.org/10.1007/s11803-013-0150-8. engineers. Princeton University Press; 2021.
[42] Zand JP, Sabouri J, Katebi J, Nouri M. A new time-domain robust anti-windup PID [61] Pourzeynali S, Zarif M. Multi-objective optimization of seismically isolated high-
control scheme for vibration suppression of building structure. Eng Struct 2021; rise building structures using genetic algorithms. J Sound Vib 2008;311(3-5):
244:112819. https://doi.org/10.1016/j.engstruct.2021.112819. 1141–60. https://doi.org/10.1016/j.jsv.2007.10.008.
[43] Shahi M, Sohrabi MR, Etedali S. Seismic control of high-rise buildings equipped
with ATMD including soil-structure interaction effects. J Earthq Tsunami 2018;12
(03):1850010. https://doi.org/10.1142/S1793431118500100.

20