Control System Design via Constraint Satisfaction using Convolutional Neural Networks and Black Hole Optimization

Yaghoobi, Saber; Fadali, M. Sami

doi:10.5220/0010618902320239

Cited by 1 publication

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The structure is a model of a 20-story building designed for the SAC Phase II Project (Gupta and Krawinkler, 1999). To design the controllers, constraints and evaluation criteria are formulated and the control design problem is considered as a Constraint Satisfaction Problem (CSP) (Zakian and Al-Naib, 1973; Yaghoobi and Fadali, 2021). The evaluation criteria are defined to satisfy constraints on various response quantities, including maximum drift, base shear, ductility, residual story drift, and control force.…”

Section: Introductionmentioning

confidence: 99%

Performance-based active controller design for nonlinear structures using modified black hole optimization

Yaghoobi

Fadali

Pekcan

2023

Journal of Vibration and Control

View full text Add to dashboard Cite

This paper presents a novel approach that facilitates the design of active controllers to mitigate seismically induced damage in structural systems. The proposed method is based on stochastic Modified Black Hole optimization algorithm. Two traditional controllers, namely Proportional-Integral-Derivative (PID) and Linear–Quadratic Gaussian (LQG) controllers were designed, and their performance was demonstrated on a benchmark 20-story steel-framed building. Evaluation criteria were defined to satisfy constraints on various response quantities, including drift, base shear, ductility, residual story drift, and control force. The constraint limits were defined in view of performance-based design requirements for the benchmark structure. The performance of the controllers was contrasted with that of traditional LQG, and significant reductions of all response quantities were achieved for design-level earthquakes.

show abstract

Section: Introductionmentioning

confidence: 99%