Multi-criteria optimisation of a barrel vault structure for emergency relief

The escalating amount of disasters presents huge challenges for humanitarian and develop- ment organisations. Those organisations have a vital task by providing shelter assistance for the affected population once they are hit by a hazard. Two main categories of shelters can be distinguished. On the one hand the family-shelters, which are provide for temporary housing of the affected population. On the other hand, humanitarian organisations also provide larger shelters, which are called collective service tents and are necessary for field hospital, commu- nity centre, storage rooms, ... . From experts on the field, it has been noted that the current family shelters are adequate enough for the emergency use, but there is a demand for better service collective tents because i) the current tents are structurally complex which slows the building process and causes errors in the construction, ii) the current tents are not adaptable. Every disaster is different, but the current tents are sold as a 'one size fits all' product with a very limited range of choice, iii) after the initial emergency phase, these tents are not reused. In this paper, we investigate the potential of using deployable structures as alternative to the current kit-of-part collective service tents. Beside the facts that deployable structures are compact structures which can be quickly and easily deployed on-site, we are exploring the possibility of designing the component that compose them as such that they can be used in a second phase - once the emergency phase is over - as kit-of-part structures for the local
population.
First of all, the new concept is discussed in the paper. What are the different advantages of using deployable shelters? What is the opinion of experts on the field about this new concept? Which kind of scissor structures can be applied for humanitarian relief? Which are the limitations of using scissor structures?
Second, a size and shape optimisation of one type of scissor structure, the barrel vault, is presented. The geometry of this structure is fully parameterised and implemented in the finite element program Abaqus. Then, a multi-criteria evolutionary optimisation algorithm, NSGA-II, is used in order to make a optimisation of the barrel vault model to two different objectives: the weight and the compactness. Two important advantages for using evolutionary algorithms for this work are i) that they support different types of design variables (continuous and discrete) and this is necessary for scissor structures because for example, the length of the structure is a continuous variable, but the number of scissor units is discrete and ii) the chance to 'stuck' in a local minimum is smaller because these algorithms work on populations rather than on individual items.
The results of the optimisation of the barrel vault model are very promising. NSGA-II is a efficient algorithm to optimise deployable structure based on scissor elements. In the next phase of this research, we will look at the way the design can have variable spans and sizes, which make them adaptable. The aim is to work with modules which can be combined in such a way that adaptable construction can be obtained. By designing such a construction, humanitarian organisations have the opportunity to adapt their construction in function of the needed help.

Jaar van publicatie:2014

Trefwoorden:multi-criteria optimisation, shelter, humanitarian aid, numerical simulation, scissor structure