Adaptive control of a deployable tensegrity structure

Deployable structures belong to a special class of moveable structures that are capable of form and size change. Controlling movement of deployable structures is important for successful deployment, in-service adaptation and safety. In this paper, measurements and control methodologies contribute to the development of an ecient learning strategy and a damage-compensation algorithm for a deployable tensegrity structure. The general motivation of this work is to develop an ecient bio-inspired control framework through real-time measurement, adaptation, and learning. Building on previous work, an enhanced deployment algorithm involves re-use of control commands in order to reduce computation time for mid-span connection. Simulations are integrated into a stochastic search algorithm and combined with case-reuse as well as real-time measurements. Although data collection requires instrumentation, this methodology performs signicantly better than without real time measurements. This paper presents the procedure and generally applicable methodologies to improve deployment paths, to control the shape of a structure through optimization, and to control the structure to adapt after a damage event.

Design and control of a prototype structure that adapts to loading through large shape changes

Arka Prabhata Reksowardojo, Gennaro Senatore, Ian Smith, Apoorv Srivastava

This paper reports on experimental testing that was carried out on a prototype adaptive structure designed to counteract the effect of loading through controlled large shape changes. The prototype is 6.6 m truss equipped with 12 linear actuators which has been designed through a method that combines geometry optimization and non-linear shape control. The structure is designed to adapt into target shapes that are optimal under each load case. Shape adaptation is achieved through controlled length changes of linear actuators that strategically replace some of the structure elements. The actuator placement is optimized to control the structure into the required target shapes. This way, material utilization is maximized and thus material energy embodied is reduced. Experimental testing is carried out to verify numerical findings and investigate the feasibility of the design method. The applied load is inferred through a classification model based on supervised learning. A control algorithm based on a linear-sequential form of geometry optimization is proposed. Experimental results show that this method successfully allows for real-time shape adaptation to achieve stress homogenization under various loading conditions. Copyright (C) 2020 The Authors.

ELSEVIER2020

Experimental Testing of a Small-Scale Truss Beam That Adapts to Loads Through Large Shape Changes

Arka Prabhata Reksowardojo, Gennaro Senatore, Ian Smith

Adaptive structures have the ability to modify their shape and internal forces through sensing and actuation in order to maintain optimal performance under changing actions. Previous studies have shown that substantial whole-life energy savings with respect to traditional passive designs can be achieved through well-conceived adaptive design strategies. The whole-life energy comprises an embodied part in the material and an operational part for structural adaptation. Structural adaptation through controlled large shape changes allows a significant stress redistribution so that the design is not governed by extreme loads with long return periods. This way, material utilization is maximized and embodied energy is reduced. A design process based on shape optimization has been formulated to obtain shapes that are optimal for each load case. A geometrically non-linear force method is employed to control the structure into required shapes. This paper presents the experimental testing of a small-scale prototype adaptive structure produced by this design process. The structure is a simply supported planar truss. Shape adaptation is achieved through controlled length changes of turnbuckles that strategically replace some of the structural elements. The stress is monitored by strain sensors fitted on some of the truss elements. The nodal coordinates are monitored by an optical tracking system. Numerical predictions and measurements have a minimum Pearson correlation of 0.86 which indicates good accordance. Although scaling effects have to be further investigated, experimental testing on a small-scale prototype has been useful to assess the feasibility of the design and control methods outlined in this work. Results show that stress homogenization through controlled large shape changes is feasible.

2019

Minimum Energy Adaptive Structures – All-In-One Problem Implementation

Jan Friedrich Georg Brütting, Gennaro Senatore, Yafeng Wang

This paper presents new implementations of a design method to obtain minimum energy adaptive structures. The design method was formulated in previous work [1, 2] for reticular structures equipped with sensors (e.g. strain, optical) and linear actuators. The main objective of the design method is the minimization of the structure whole-life energy through reduction of the energy embodied in the material at a small increase in operational energy for structural adaptation. Structural adaptation is employed to counteract the effects of large loading events through control of the shape and of the internal load-path such that the design will not be governed by peak demands that occur rarely. Through balancing material use and operational energy demand, this method explicitly addresses current and future challenges such as material scarcity, energy depletion and reduction of building environmental impacts. This method has been applied to realistic spatial structures of complex layout [3] and it has been successfully tested on a large-scale prototype [4]. The substantial whole-life energy savings obtained through this method (up to 70% for slender structures) and its scalability to structures of complex layout makes it attractive for applications to realistic building scenarios. The All-in-One problem formulation to obtain minimum energy adaptive structures is stated in [1]. However, the method given in [1] to solve the synthesis process is based on a nested optimization scheme which does not guarantee solution optimality. This paper presents a new implementation of the same synthesis process but based on mixed-integer programming (MIP). Branch-and-bound methods are employed to solve the MIP-based formulation to optimality. Topology, cross-section sizing and actuator layout are treated as binary design variables. Instead, member forces, nodal displacements and actuator control commands (i.e. length changes) are treated as continuous state variables. Geometric compatibility between nodal displacements and element deformations is enforced through adaptation, the actuator length changes modify the internal load-path to homogenize the stresses and change the shape of the structure to keep displacements within serviceability limits. Solutions obtained through the MIP-based method are benchmarked against those obtained using the nested optimization approach given in [1].

Minimum Energy Adaptive Structures – All-In-One Problem Implementation

Jan Friedrich Georg Brütting, Gennaro Senatore, Yafeng Wang