The definition of Statistical Energy Analysis (SEA) models for large complex structures is highly conditioned by the classification of the structure elements into a set of coupled subsystems and the subsequent determination of the loss factors representing both the internal damping and the coupling between subsystems. The accurate definition of the complete system can lead to excessively large models as the size and complexity increases. This fact can also rise practical issues for the experimental determination of the loss factors. This work presents a formulation of reduced SEA models for incomplete systems defined by a set of effective loss factors. This reduced SEA model provides a feasible number of subsystems for the application of the Power Injection Method (PIM). For structures of high complexity, their components accessibility can be restricted, for instance internal equipments or panels. For these cases the use of PIM to carry out an experimental SEA analysis is not possible. New methods are presented for this case in combination with the reduced SEA models. These methods allow defining some of the model loss factors that could not be obtained through PIM. The methods are validated with a numerical analysis case and they are also applied to an actual spacecraft structure with accessibility restrictions: a solar wing in folded configuration.

No abstract

This work presents an improved version of the Modified Simulated Annealing Algorithm (I-MSAA). Modified Simulated Annealing Algorithm (MSAA) was recently introduced for solving global optimization problems and is a newly improved version of the simulated annealing (SA) with three modifications: preliminary exploration, search step and probability of acceptance. The I-MSAA proposed here does not perform a preliminary exploration and reduces the probability of accepting worse solutions. I-MSAA was evaluated in benchmark functions (constrained and unconstrained) reported in the literature. The results indicated that I-MSAA is a good tool to optimize functions of high complexity.

No abstract

This paper presents the size optimization of trusses structures using the recently developed Improved Modified Simulated Annealing Algorithm (I-MSAA). I-MSAA was recently introduced for solving global optimization problems and is a newly improved version of the Modified Simulated Annealing Algorithm (MSAA) with two modifications: i) reduction of probability of accepting worse solutions; ii) the starting point is chosen randomly. I-MSAA was evaluated in five benchmark problems of truss size optimization. The results were compared by those reported by other metaheuristic algorithms and indicated that I-MSAA is stable and efficient to optimize this type of problems.