Facial reconstruction

Search LJMU Research Online

Browse Repository | Browse E-Theses

Active Structural Control of Single and Multi-Span Beam Structures Subjected to Transient Loads

Sievert, L (2022) Active Structural Control of Single and Multi-Span Beam Structures Subjected to Transient Loads. Doctoral thesis, Liverpool John Moores University.

[img]
Preview
Text
2022sievertPhD(1).pdf - Published Version
Available under License Creative Commons Attribution Non-commercial.

Download (6MB) | Preview

Abstract

This study directly addresses the problem of active control of beam structures under the action of moving masses. In this regard, experimental implementations of the particular active control solutions are still rarely seen in the literature. The main objective is to experimentally implement and validate active control solutions for two small-scale test stands with the aim to reduce the structural deflection.
The first supporting structure is modelled as an Euler–Bernoulli simply supported beam, acted upon by moving masses of different weights and velocities. The experimental implementation of the proposed optimal controller poses a particular set of challenges as compared with numerical solutions. Specifically, it can include errors due to discretization and the states cannot be directly measured. The resulting limitations of classical optimal observer techniques are stated and consequently the states are estimated by a method utilizing the mode shapes. It is shown both numerically and experimentally that using electromagnetic actuation, a reduced order controller designed using a time-varying algorithm, provides a reduction of the maximum deflection of up to 38% as compared with the uncontrolled structure. Herein an augmented system model is utilised, which includes the moving mass in the system equation. The controller performance and robustness were tested against a representative set of possible moving load parameters. In consequence of the variations in moving mass weight and speed, the controller gain requires a supplementary adaptation. A simple algorithm that schedules the gain as a function of the weight and speed of the moving mass can achieve both a good performance and an adjustment of the control effort to the specific design requirements.
In the second part of this study cubic and linear displacement feedback control approaches are studied experimentally for a simply supported beam as well as for the two-span continuous beam. The two-span beam structure is modelled by approximating the support by spring damper elements of high stiffness and damping coefficient. Piezoelectric macro fibre composites serve as actuators. The control methods are, compared to the previous approach, more straightforward to implement and can handle a stream of moving masses. However, optimality and stability cannot be guaranteed and have to be validated experimentally. The linear displacement feedback shows better performance for low weights of the moving masses whereas the cubic displacement feedback achieves higher deflection reduction for higher weights.

In the last part, constrained model predictive control is studied numerically for both of the structures. This is currently the only control approach which can take into account saturation limits explicitly by quadratic programming. In this way, better performance is achieved for both test structures as compared to the displacement feedback control approaches.

Item Type: Thesis (Doctoral)
Uncontrolled Keywords: Active vibration control; Experimental; Feedback control; Model predictive control; Moving mass; Multi-span beam structure; Piezoelectric actuators; State estimation; Time-varying optimal control
Subjects: T Technology > TA Engineering (General). Civil engineering (General)
Divisions: Engineering
SWORD Depositor: A Symplectic
Date Deposited: 28 Jun 2022 11:31
Last Modified: 30 Aug 2022 09:52
DOI or ID number: 10.24377/LJMU.t.00017076
Supervisors: Stancioiu, D, Matthews, C and Rothwell, G
URI: https://researchonline.ljmu.ac.uk/id/eprint/17076
View Item View Item