Eliseo Pablo Vergara González

Outlier Detection and Data Cleaning in Multivariate Non-Normal Samples: The PAELLA Algorithm

A new method of outlier detection and data cleaning for both normal and non-normal multivariate data sets is proposed. It is based on an iterated local fit without a priori metric assumptions. We propose a new approach supported by finite mixture clustering which provides good results with large data sets. A multi-step structure, consisting of three phases, is developed. The importance of outlier detection in industrial modeling for open-loop control prediction is also described. The described algorithm gives good results both in simulations runs with artificial data sets and with experimental data sets recorded in a rubber factory. Finally, some discussion about this methodology is exposed.

A Soft Computing Approach to Optimize the Production of Biodiesel

There is an increasing global concern for environmental protection for the conservation of non-renewal natural resources. It needs to be obtain an alternative, renewable and biodegradable combustible like biodiesel. Waste cooking oil is a potential replacement for vegetable oils in the production of biodiesel. Biodiesel is synthesized by direct transesterification of vegetable oils, which is controlled by several inputs or process variables, including the dosage of catalyst, process temperature, mixing speed, mixing time, humidity and impurities of waste cooking oil. This study proposes a methodology to improve the production of biodiesel based on the use of soft computing techniques to predict several features of biodiesel production. The method selected a group of regression models based on Support Vector Machines (SVM) techniques to perform a prediction of several properties of a biodiesel sample taking into account a configuration of 7 test inputs. This test inputs were: molar ratio, dosage of catalyst, temperature, mixing speed, mixing time, humidity and impurities. Then and based on these inputs, the features to predict were: yield, turbidity, density, viscosity and high heating to obtain a better understanding of the process. Finally, considering the samples of the design of experiments studied, it has been observed that SVM models, based on a radial basic function kernel, record accurate results, with the best performance in four of the five features, improving in all the cases the accuracy obtained using linear regression.

Adjust the Thermo-Mechanical Properties of Finite Element Models Welded Joints Based on Soft Computing Techniques

An appropriate characterization of the thermo-mechanical behavior of elastic-plastic Finite Element (FE) models is essential to ensure realistic results when welded joints are studied. The welded joints are subject to severe angular distortion produced by an intense heat concentration on a very small area when they are manufactured. For this reason, the angular distortion and the temperature field, which the joints are subjected, is very difficult to model with the Finite Element Method (FEM) when nonlinear effects such as plasticity of the material, radiation and thermal contacts are considered. This paper sets out a methodology to determine the most appropriate parameters needed for modelling the thermo-mechanical behavior in welded joints FE models. The work is based on experimental data (temperature field and angular distortion) and the combined use of Support Vector Machines (SVM) and Genetic Algorithms (GA) with multi-objective functions. The proposed methodology is applied for modelling Butt joint with single V-groove weld manufactured by Gas Metal Arc Welding (GMAW) process when the parameters of speed, current and voltage are, respectively, 6 mm/sec 140 amps and 26 V.

Cyclone Performance Prediction Using Linear Regression Techniques

A wide range of industrial fields utilize cyclone separators and so, evaluating their performance according to different materials and varying operating conditions could contribute useful information and could also save these industries significant amounts of capital. This study models cyclone performance using linear regression techniques and low errors were obtained in comparison with the values obtained from real experiments. Linear regression and generalized linear regression techniques, simple and enhanced with Gradient Boosting techniques, were used to create linear models with low errors of approximately 0.83 % in cyclone performance.

A Soft Computing Approach to Optimize the Clarification Process in Wastewater Treatment

The coagulation process allows for the removal of colloidal particles suspended in wastewater. Estimating the amount of coagulant required to effectively remove these colloidal particles is usually determined experimentally by the jar test. The configuration of this test is often performed in an iterative manner which has the disadvantage of requiring a significant period of experimentation and an excessive amount of coagulant consumption. This study proposes a methodology to determine the optimum natural coagulant dose while at the same time eliminating the maximum amount of colloidal particles suspended in the wastewater. An estimation of the amount of colloidal particles removed from the wastewater is determined by the turbidity in a standardized jar test, which is applied to the wastewater at the wastewater treatment plant in Logrono (Spain). The methodology proposed is based on the combined use of soft computing techniques and evolutionary techniques based on Genetic Algorithms (GA). Firstly, a group of regression models based on neural networks techniques was performed to predict the final turbidity of a wastewater sample taking into consideration a configuration of jar test inputs. The jar test inputs are: initial turbidity, natural coagulant dosage, temperature, mix speed and mix time. Finally, the best combination of jar test inputs to obtain the optimum natural coagulant dose, while also eliminating the maximum amount of colloidal particles, was achieved by applying evolutionary optimization techniques to the most accurate regression models obtained beforehand.

Design of a Device to Eliminate Isocyanuric Acid from Water

This paper shows the design of a device for partial eliminating of isocyanuric acid (ICN) from swimming pool water using melamine additives. The renewal process of swimming pool water through its own purification makes absolutely necessary the elimination of isocyanuric acid that has been accumulated in the water over time. An excess of isocyanuric acid in water will then prevent chlorine effectiveness in the pool water and as a result, becomes harmful to human health. Therefore, the disinfection stage is considered as well as Isocyanuric acid (ICN) stabilization and as doing this is achieved through melamine-photometry filtering of insoluble complex ICN-M. The overall objective of these stages of purification is to eventually eliminate ICN from swimming pool. The overall objective of this device is to eventually eliminate ICN from swimming pool and then make it safe for human uses, a case that has been considered viable technologically and economically in the system treatment.