Computational & Technology Resources
an online resource for computational,
engineering & technology publications
PROCEEDINGS OF THE EIGHTH INTERNATIONAL CONFERENCE ON COMPUTATIONAL STRUCTURES TECHNOLOGY
Edited by: B.H.V. Topping, G. Montero and R. Montenegro
Evolutionary Optimization of Strategies for the Demolition of Buildings with Explosive Charges Using Multibody Dynamics
M. Baitsch, M. Breidt, M. Ilikkan and D. Hartmann
Ruhr-University of Bochum, Germany
M. Baitsch, M. Breidt, M. Ilikkan, D. Hartmann, "Evolutionary Optimization of Strategies for the Demolition of Buildings with Explosive Charges Using Multibody Dynamics", in B.H.V. Topping, G. Montero, R. Montenegro, (Editors), "Proceedings of the Eighth International Conference on Computational Structures Technology", Civil-Comp Press, Stirlingshire, UK, Paper 230, 2006. doi:10.4203/ccp.83.230
Keywords: demolition, controlled explosives, multibody dynamics, optimization, evolution strategies, distributed software components.
The controlled destruction (demolition) of buildings by means of explosive charges represents an efficient technology and is successfully applied to complex high-rise buildings in urban areas since years. However, the collapse of the building caused by the ignition of explosive loads is associated with a high level of uncertainty. Therefore, the demolition using explosive charges requires a suitable strategy in terms of efficiency and safety. Today, the development of this strategy is done rather intuitively, mainly based upon experience. But many examples of "non optimal" demolitions show that this conventional approach is not sufficient. The present contribution introduces a new approach to the design of demolition strategies that employs an efficient simulation model using multibody dynamics and optimization methodology.
The simulation model developed is based on a multi-level model of the blasting process by applying modern computing methods and computational mechanics. Three main levels are introduced: On the first level (local level) the effects of the exploding charges are modeled such that the volitional damages can be captured and described. On the second level (near field level) the effects of the local damages on the adjacent structure are analyzed. Based on the first and second level, finally the collapse of the entire structure is modeled on the third level (global level) including fracture processes and relevant contact mechanisms. The core physical model of the simulation model of the global level is a special multibody model that is created adaptively during the simulation process. The effects of the local and the near field failure or fracture processes of the reinforced concrete parts respectively, are represented in terms of multibody subsystems that employ special kinematics as well as force elements with nonlinear characteristics in form of force-displacement relations (resistance characteristic curves). The multibody approach is favoured over a nonlinear transient finite element analysis because of computing time constraints imposed by the optimization. The realistic and efficient simulation of the building collapse provides a sound foundation for the use of optimization methods to support the planning of the demolition. In the planning stage, the engineer has to decide about the number and placement of the explosive charges and the time flow of the ignition. The decision about the parameters of the demolition strategy is driven by several criteria that have to be taken into account. The most important criterion is the so called debris area that is potentially affected by falling parts of the building. In the vicinity of other buildings or subterranean constructions also vibrations caused by the impact of the falling parts on the soil may be critical. In order to use optimization methods, the design problem has to be expressed as an optimization problem as realistically as possible. The formulation of the optimization problem is based on the specification of the principal strategy (e.g. vertical collapse, directed collapse or folded collapse) and the related parameters. On this basis, the placement and the ignition time sequence are chosen as design variables for the optimization problem. The debris area is considered as the most important design criterion and therefore directly employed as the objective function. The resulting optimization problem does not have constraints which is desirable because unconstrained problems are generally easier to solve. Unfortunately, the objective function turns out to be noisy and lack C1-continuity which is probably caused by the multibody dynamics simulation on which the prediction of the debris area is based. Therefore, evolution strategies (ES) are employed for the actual optimization because they are known to be robust and reliable even for problems to which gradient based methods can not be applied.
The application examples stemming from the demolition of civil and industrial structures and buildings presented in the full paper show that optimization methods can substantially improve the process of designing demolition strategies. Using ES, it is possible to find solutions that can hardly be identified by an engineer because of the complex and non-intuitive behavior of collapsing buildings.
purchase the full-text of this paper (price £20)