Our stick and beam approach to building design is very different from the way nature and other industries develop forms to provide functional needs. As a result, our buildings are generally over designed and difficult to adapt to changing environments.
This research investigates how products and systems in the natural environment, automotive and aeronautical industries respond to changing environmental conditions. We are looking at how these strategies could be applied to building design to investigate how structural systems could adapt to changing forces over time. An adaptive structural system will allow for a better use of materials and a higher standard of energy efficiency, whilst the capacity to adapt to changing conditions will increase the degree of resilience of the system throughout its life cycle.
Adaptive, Responsive Building Structures
Adaptive Structures can be thought of as spatial structures with embedded sensors, actuation and control intelligence. The main components (fig. 1) of an adaptive structural system can be classified in:
- sensing, which is the capacity of the system to reach an awareness of its state (stress, strain, relative positions of its elements);
- actuation; which involves transformation of stored (such as chemical) or supplied (such as electrical, magnetic) energy to the system into mechanical energy. During actuation the system modifies its properties (i.e. stiffness, phase change, chemical composition) or it can restructures the interactions between its elements (i.e. change of its shape);
- control intelligence which processes the information gathered by sensors and provide appropriate input for the actuators in order to keep the system within desired boundaries (closed loop feedback). The importance of a closed feedback loop is paramount to achieve adaptive response since the relation between data gathered by sensors and input to actuators must be determined on-line on the base of the current state of the system. Machine learning techniques can enable adaptive structures to learn from data gathering in order to predict recurrent patterns of load for maximising control efficiency.
- load-bearing capacity in order to withstand static and dynamic loads , external and internal forces (including those from the actuators);
Sensing, actuation and control intelligence can be designed to improve the load bearing capacity of the structure which is enabled to counteract severe loads at occurrence and to monitor continuously its state of stress and deflections. However, since active elements require input energy, adaptive systems must be designed so that the benefits brought by new acquired functionalities outweigh the consumption of energy needed.
Adaptive Structures, what they can do
The scope for adaptive structures can be framed in two main categories:
- response control;
- shape morphing control;
Response control aims at controlling the structural state of the system by varying its stiffness locally and globally on the occurrence of unusual or unexpected loads. This type of control would apply to situations in which forces are generated by external agents such as wind, snow and vibrations caused by earthquakes or aerodynamic forces (i.e. tall buildings and bridges).
The scope of shape morphing control can be divided in two main levels, global and local. The first aims at restructuring the shape of buildings in order to minimize/maximize exposure to external agents such as wind, solar radiation or snow or others. The second aims at controlling smaller regions in order to maintain ambient conditions within desired boundaries. This involves modulation of direct and diffuse light (shading), control and enhancement of buoyancy effects (natural ventilation).
Both of them present a radical change of perspective regarding design criteria:
- The former aims at controlling the structural state of the system by redistribution of internal forces through the action of active elements (change in stiffness or change in length).
- The latter sees as primary design targets a set of optimal deformed shapes in which the structure, or part of it, should be able to morph. These “modal shapes” represent the physical embodiment of desired performances that the building structure will have to provide. In other words, the “modal shapes” are the result of a mapping between external stimuli and desired building performances.
Adaptive Structures for Whole Life Energy Savings
Designing structures with minimal environmental impact is now a common concern in the construction sector. Conventional structural design practice usually involves ensuring that the strength and deformation of the structure meet the required limits to cope with the worst load cases. Most of the time the loads experienced by the structure are much lower than the design load and the structure is effectively overdesigned for most of its working life (fig. 7).
The design methodology formulated in this work takes a substantial shift from conventional methods. By contrast to this conventional passive approach the method presented here replaces passive member strategically with active elements (actuators) which are only activated when the loads reach a certain threshold. The structure can withstand low level of loads passively. Above the threshold, actuation comes in to allow the structure to cope with high but rare loading scenarios. The controlled length change of the actuators allows the pattern of internal forces to be modified, “load path management”.
In so doing stresses can be minimized and homogenized while displacements are kept within desired limits (fig. 8). The design process consists in determining the size of the passive elements as well as the optimal number and position of the actuators to minimise the total energy. This total energy includes the embodied energy of the material used in the structure as well as the operational energy necessary to operate the active elements. We use this dual design to minimize the overall energy required by the structures.
This methodology has been used on a simple truss structure (check the outputs section) and it was showed that it allows significant weight saving compared to conventional passive design. The methodology was extended to cope with the design of more complex 3D configurations (fig. 9) and it is confirmed that an optimum activation threshold exists that leads to design that minimises the total energy of the structure. Compared to an optimised passive design we show that the total energy saving is 6 fold (fig. 10). Comparison with the results from the previous study suggests that the conclusions reached on a simple design are confirmed and strengthened for a more complex structure.
Fig 1 closed loop feedback control for adaptive structures with embedded sensors and actuatros
Fig 2 (a) conventional actuator mechanisms, (b) solid state actuation
Fig 3 the non linear mapping properties of Artificial Neural Networks can be used to derive appropriate laws to control the dynamics of complex structural systems
Fig 4 compliant structures made of continous stripes and joint with active elements. The alternate actuation of the two sets of actuators (blue and red) causes pronounced bending deformation through small actuation forces small forces
Fig 5 the chart plots the region of force vs displacementin which several actuation systems can operate. Those are compared with domains of requirements related to building performances
Fig 6 examples of adaptive morphing structures using the real time physics engine developed for PushMePullMe software. More info in the outputs section (computer program).
Fig 7 embodied/operational energy vs. max expected load
Fig 8 comparison between optimised passive and adaptive structure (actuators represented as magenta lines).More info in the outputs section (publication).
Fig 9 3D adaptive truss structure and optimal actuators topology result of the design/optimization process. More info in the outputs section (publication).
Fig 10 total energy (embodied + operational) comparison for passive and adaptive case study considering 50 years of load history.