Implementation of bi-directional evolutionary structural optimization (BESO) in the underground excavations and slope stability

Abstract

Stability of underground excavation is a major concern because any failure may lead to costly damages and possible loss of life. Since the underground material does not have predefined properties, engineers are exposed to a complex situation. In order to avoid adverse effects of failure, researchers and engineers have tried to fix the problem by taking higher factor of safety for the design of external supporting system, reinforcement and optimizing the shape of underground caverns.

During the last 40 years, advances in technology and application of computer made the job for civil and mining engineers much easier. Therefore, the application of finite element method (FEM) gives the engineers the power to simulate any shape and analyse any real case problems. Finally, as a gap was felt in finding the best shapes of design, engineers searched through different fields to solve it. As a result, the development of optimization methods gave researchers a chance to use computer to design the optimum shape.

Although it was primarily proposed to optimize the shape of structural members, topology optimization has been used by few researchers in geotechnical applications. Despite its great potential, in this field, very few and simple studies have been carried out which were limited to two dimensional designs and the presence of joints, discontinuities, materials with various mechanical properties and junctions as well as material non-linearity were not implemented and investigated thoroughly.

In practical projects, most underground excavations are firstly designed according to the experience of engineers and then numerically studied by simulating it by computer programs thereby predicting the weakness and robustness in the design. But here in this research, computer is used to decide how to excavate and cast reinforcement based on different objective functions.

To do so, a computer program is developed to perform the optimization that includes firstly setting up finite element method and then applying the Bi-directional Evolutionary Structural Optimization (ESO). While there are many methods to perform optimization, due to its robustness in producing reasonable results in designing geotechnical problems and ease of linking with other computer programs (finite elements packages), BESO has been chosen.

Secondly two-dimensional patterns are simulated and used as the verification of correctness of the procedure by comparing with other researchers’ results. Lastly 3D models are considered, analysed and the outcomes are presented. Finally, plasticity is considered and methods to implement measures related to stability of the design domain are discussed.