Beatriz Souza Leite Pires de Lima

A comparison of models for uncertainty analysis by the finite element method

Abstract Uncertainty in structural engineering analysis exists in the architecture of a structural system, its basic parameters, the information resulting from the abstracted aspects of the system, and the non-abstracted or unknown aspects of the system. Also, uncertainty is present as a result of prediction models, analysis and design of structures, and general lack of knowledge about the behavior of real structures. One of the important factors that lead to errors in numerical predictions is the degree of precision in obtaining the relevant parameters. In this paper we discuss two different methodologies: 1. Classical probabilistic approach, in which the properties are treated as random variables. Stochastic Finite Element Methods are examined using both Monte Carlo Simulation and Perturbation Methods. 2. Possibilistic approach, by a model based on the theory of fuzzy sets. Some results are presented to point out the main characteristics of the two methodologies (Lima, D.Sc. Thesis, COPPE/Federal University Rio de Janeiro, 1996).

Design of interpretable fuzzy rule-based classifiers using spectral analysis with structure and parameters optimization

This paper presents a design method for fuzzy rule-based systems that performs data modeling consistently according to the symbolic relations expressed by the rules. The focus of the model is the interpretability of the rules and the models accuracy, such that it can be used as tool for data understanding. The number of rules is defined by the eigenstructure analysis of the similarity matrix, which is computed from data. The rule induction algorithm runs a clustering algorithm on the dataset and associates one rule to each cluster. Each rule is selected among all possible combinations of one-dimensional fuzzy sets, as the one nearest to a clusters center. The rules are weighted in order to improve the classifier performance and the weights are computed by a bounded quadratic optimization problem. The model complexity is minimized in a structure selection search, performed by a genetic algorithm that selects simultaneously the most representative subset of variables and also the number of fuzzy sets in the fuzzy partition of the selected variables. The resulting model is evaluated on a set of benchmark datasets for classification problems. The results show that the proposed approach produces accurate and yet compact fuzzy classifiers. The resulting model is also evaluated from an interpretability point of view, showing how the rule weights provide additional information to help data understanding and model exploitation.

Developing a fuzzy decision support system to determine the location of a landfill site

This paper presents two case studies of municipal solid waste site location using a decision-support system based on fuzzy logic. This problem is very complex, as it requires the evaluation of different criteria, which involve environmental, social and economic data. Such data deal with a wide range of information that presents not only quantitative, but also qualitative knowledge. In order to deal with this characteristic, the developed system employs fuzzy rules due to its ability to treat linguistic variables and the human way of thinking. Conventional approaches tend to be less effective in dealing with the imprecise or vague nature of the linguistic assessment. A case study for selecting the location of a new municipal solid waste landfill for the city of Petropolis in Rio de Janeiro is presented. Testing of the proposed method was carried out using data from the municipal solid waste location for another municipality in Rio de Janeiro, Brazil.

Optimal design of submarine pipeline routes by genetic algorithm with different constraint handling techniques

Abstract This work deals with optimization methods for the selection of submarine pipeline routes, employed to carry the oil & gas from offshore platforms. The main motives are related to the assessment of constraint-handling techniques, an important issue in the application of genetic algorithms and other nature-inspired algorithms to such complex, real-world engineering problems. Several methods associated to the modeling and solution of the optimization problem are addressed, including: the geometrical parameterization of candidate routes; their encoding in the context of the genetic algorithm; and, especially, the incorporation into the objective function of the several design criteria involved in the route evaluation. Initially, we propose grouping the design criteria as either “soft” or “hard”, according to the practical consequences of their violation. Then, the latter criteria are associated to different constraint-handling techniques: the classical static penalty function method, and more advanced techniques such as the Adaptive Penalty Method, the e-Constrained method, and the Ho-Shimizu technique. Case studies are presented to compare the performance of these methods, applied to actual offshore scenarios. The results indicate the importance of clearly characterizing feasible and infeasible solutions, according to the classification of design criteria as “soft” or “hard” respectively. They also indicate that the static penalty approach is not adequate, while the other techniques performed better, especially the e-Constrained and the Ho-Shimizu methods. Finally, it is seen that the optimization tool may reduce the design time to assess an optimal route, providing accurate results, and minimizing the costs of installation and operation of submarine pipelines.

Optimization of Steel Catenary Risers for Offshore Oil Production Using Artificial Immune System

This work presents an application of Artificial Immune System (AIS) using Clonalg to the synthesis and optimization procedure of a Steel Catenary Riser (SCR) for floating oil production systems at deep and ultra-deep waters. The evaluation of the behavior of riser configurations, needed for the calculation of the fitness function in the optimization procedure by an evolutionary algorithm, requires a large number of time-consuming Finite Element analyses. Therefore, it is important to reduce the number of analyses; in this paper, the effectiveness of AIS for this purpose is assessed in this real-world industrial application. The results indicate that the AIS approach is more effective than Genetic Algorithms (GA), generating better solutions with smaller number of evaluations.

Balanced ranking method for constrained optimization problems using evolutionary algorithms

This work presents a new technique to handle constraints in the solution of optimization problems by evolutionary algorithms - the Balanced Ranking Method (BRM). In this method the fitness function is based on two rankings, for feasible and infeasible solutions respectively. The rankings are merged according to deterministic criteria that consider the status of the search process and specific properties of the population. The focus of the BRM method is to comprise a constraint-handling technique (CHT) that is not coupled to the optimization algorithm, and thus can be implemented into different algorithms. The method is compared with other well-known CHTs that follow this same uncoupled approach, all implemented into a canonical Genetic Algorithm. Two well-known suites of benchmark functions and five engineering problems are used as case studies. The performance of the different CHTs is assessed by nonparametric statistical tests, including the Sign test and the Wilcoxon Signed-Ranks test. The results indicate that the BRM presents a good performance, being reliable and efficient, while maintaining its uncoupled characteristic leading to an easy implementation and hybridization with any search algorithm.

A multi-model approach for long-term runoff modeling using rainfall forecasts

This work presents the development of a rainfall-runoff model for the Iguacu River basin in the south of Brazil. The model was developed to support the operational planning of hydroelectric power plants and is intended to compute natural flow predictions based on meteorological rain forecasts. A recurrent fuzzy system model was employed, with parameters estimated by a genetic algorithm using observed rainfall as input. This work presents the recurrent fuzzy model within a multi-model approach, where the input data are furnished as an envelope, resulting in a prediction envelope that has demonstrated the ability to produce robust results.

Application of Genetic Algorithms to the Synthesis of Riser Configurations

The purpose of this work is to describe the application of Genetic Algorithms in the search of the best configuration of catenary riser systems in deep waters. Particularly, an optimization methodology based on genetic algorithms is implemented on a computer program, in order to seek an optimum geometric configuration for a steel catenary riser in a lazy-wave configuration. This problem is characterized by a very large space of possible solutions; the use of traditional methods is an exhaustive work, since there is a large number of variables and parameters that define this type of system. Genetic algorithms are more robust than the more commonly used optimization techniques. They use random choice as a tool to guide a search toward regions of the search space with likely improvements. Some differences such as the coding of the parameter set, the search from a population of points, the use of objective functions and randomized operators are factors that contribute to the robustness of a genetic algorithm and result in advantages over traditional techniques. The implemented methodology has as baseline one or more criteria established by the experience of the offshore engineer. The implementation of an intelligent methodology oriented specifically to the optimization and synthesis of riser configurations will not only facilitate the work of manipulating a huge mass of data, but also assure the best alternative between all the possible ones, searching in a much larger space of possible solutions than classical methods.© 2003 ASME

Synthesis and Optimization of Submarine Pipeline Routes Considering On-Bottom Stability Criteria

Researchers from Petrobras and LAMCSO/COPPE/UFRJ are currently involved in the development and implementation of a computational tool, based in Evolutionary Algorithms, for the synthesis and optimization of submarine pipeline routes. In this tool, randomly generated candidate routes are evaluated in terms of several criteria, incorporated in an objective (or fitness) function to take into account the relevant aspects that should be considered in the design of a route. A previous work [1] described the initial steps taken towards the development of such tool. In that work, attention was dedicated to the geometrical representation of a route, and to some of the terms of the objective function associated with a preliminary, global step of the optimization process (such as total pipeline length, and geographical-topographical issues associated with the route geometry and to the seabottom bathymetry and obstacles). Now, this work focuses in other aspects related to the structural behavior of the pipe, under hydrostatic and environmental loadings; more specifically, special attention is dedicated to the implementation of On-Bottom Stability (OBS) criteria such as the proposed in the RP-F109 code [2]. Case studies are presented to illustrate the use of the optimization tool and to assess the influence of the OBS criteria.Copyright

Potential collaboration discovery using document clustering and community structure detection

Complex network analysis is a growing research area in a wide variety of domains and has recently become closely associated with data, text and web mining. One of the most active areas in the study of complex networks is the detection of community structure, which can be related to the clustering problem in data mining. This paper employs a community structure detection algorithm for document clustering in order to discover potential relationships in a social network. The proposed approach is explored in a case study of potential collaboration discovery among the research staff in the Graduate Civil Engineering Department of the Federal University of Rio de Janeiro, Brazil. The results show that the combined use of both techniques provides useful insights on the relationships, both existent and potential, among individuals in the social network.