Neurofuzzy Systems

In situations where maximum load demand exceeds power generation capacity, frequency would drop during cases of overload; and if appropriate under-frequency load curtailment scheme is not put in place, it could cause permanent turbine damage or affect the expected working operation of electric equipment. Various existing works on load curtailment in power systems have been done to great effect using meta-heuristic and classical control techniques. This paper seeks to comparatively analyze the Fuzzy Logic, Fuzzy-PID and Neuro-Fuzzy (ANFIS) controllers and their contribution to achieving load curtailment in a Single Area Power System via continuous load control. The transfer function model of the power system and the design of the controllers are done using Simulink, MATLAB software. The controllers are designed to output power imbalance based on frequency deviation during various system overload conditions. They also shed loads continuously until frequency is restored within the safe...

Resumo: A engenharia geotécnica tem sido de extrema importância para o desenvolvimento territorial, sendo aplicada desde a exploração de minérios, portos até em estruturas profundas como túneis e obras rodoviárias. Sua influência e análise vêm ajudando significantemente nos aprimoramentos de estruturas mais seguras. Dentro deste contexto, uma das obras mais importantes no sul catarinense é a rodovia SC-446, popularmente conhecida como Via Rápida, projetada para receber 22.000 veículos por dia e com extensão de 12,7 km. Sua construção permitiu um novo eixo de desenvolvimento, não só pela mobilidade urbana, mas também no crescimento da região. Entretanto, esta via apresenta condições geotécnicas sujeitas a movimentos de massa em alguns trechos, podendo colocar em risco os usuários que por ela transitam. Um destes pontos é o talude rodoviário do km 06 que apresenta processos deflagradores de movimentos de massa. Neste sentido, buscou-se com este estudo realizar uma análise de estabilidade deste talude, utilizando-se os métodos usuais de estado de equilíbrio limite por meio de uma rotina de modelagem matemática no software GEO5 2020 ® , obtendo-se o fator de segurança (F.S) e comparando-se o mesmo com a norma ABNT NBR 11.682:2009. Um modelo da geometria inicial do talude foi inserido na ferramenta computacional e juntamente com seus parâmetros característicos, determinou-se seu coeficiente de segurança, por meio do método de Bishop (1955), resultando-se em um valor de 1,06 para o talude em estado seco e 1,02 para o talude em estado parcialmente saturado. Por intermédio desta análise, constatou-se que o talude encontra-se instável segundo as recomendações da ABNT NBR 11.682:2009. Para o talude em questão, indica-se o emprego da técnica de solo grampeado com paramento em concreto projetado, utilizando-se uma malha regular com espaçamento entre grampos de 1,45 metros, comprimento de grampos variando de 7 a 11 metros e diâmetro de 32 mm, com diâmetro dos furos de 100 mm e resistência a tração de 500 Mpa, para assim obter-se um fator de segurança igual a 1,5, recomendado pela norma ABNT NBR 11.682:2009

Planning of demand forecasting for perishable products is important for any type of industry that manufactures or distributes, especially if it has a seasonal behavior and a difficult to predict variability. This paper proposes a metaheuristic based on Ant Colony Optimization (ACO) for the combination of forecasts of multiple products, based on three models: Mixed Linear Model (MLM), Bayesian Regression Model with Innovation (BRM) and Dynamic Linear Bayesian Model (BDLM), which are part of the proposed combination whose process is based on minimizing the Mean of Absolute percentage Error (SMAPE) indicator. It is found that the BDLM and BRM methodologies obtain good results on an individual basis, being better BRM, however, the ACO algorithm designed yields a better result, facilitating an adequate prediction of the demand of several products of a company in the meat buffer sector.

Lake Fúquene is one of the most important lentic water bodies in Colombia,
since it supplies water for thousands of people and multiple productive activities.
However, the strong environmental degradation that it has suffered due to anthropogenic
intervention, through discharges from point and non-point sources, have generated
nutrient accumulation processes and hypoxic conditions that have completely altered its
dynamics. In order to analyze these processes, two water quality monitoring were carried
out in April and October 2019 in different geographical points of the lake evaluating the
spatial variation of the concentrations of the parameters analyzed. The space-time concentration of pollutants such as ammonia, evidence its high state of degradation.
The continuous contribution of pollutants has exceeded the capacity of the lake and the improvement processes implemented by agencies to achieve acceptable water quality continue to be seriously affected by their habitat depletion despite the efforts made to decontaminate this important water body

This study is aimed at obtaining a relationship between the values defining bead geometry and the welding parameters and also to select optimum welding parameters. For this reason, an experimental study has been realized. The welding parameters such as the arc current, arc voltage, and welding speed which have the most effect on bead geometry are considered, and the other parameters are held as constant. Four, three, and five different values for the arc current, the arc voltage, and welding speed are used, respectively. So, sixty samples made of St 52-3 material were prepared. The bead geometries of the samples are analyzed, and the thickness and penetration values of the weld bead are measured. Then, the relationship between the welding parameters is modeled by using artificial neural network (ANN) and neurofuzzy system approach. Each model is checked for its adequacy by using test data which are selected from experimental results. Then, the models developed are compared with rega...

The poor management of hotspots in electrical systems often leads to very devastating consequences; an extremity is fire outbreak which could result in loss of lives and/or properties. Hence, the necessity to effectively manage or handle electrical hotspots cannot be overstated. This paper proposes a model for arresting the occurrence of the destructive activities of electrical hotspots in industries by controlling the temperature, humidity and dust density within electrical components/ equipment. To this end, a thermal imaging camera is employed for the detection of various locations and magnitudes of hotspots within electrical systems. Based on the globally approved industrial standards for the prevention of thermally induced electrical systems failure, each of the electrical components and equipment [whose thermal excursion is beyond the allowable temperature rise under measured load values (i.e. ΔT corr)] is identified and treated by adopting the recommended actions. Additionally, a fuzzy logic control (FLC) system is designed. This is further developed into an adaptive neuro-fuzzy inference system (ANFIS) for the control of the operation of the air handling unit (AHU) and the aspirator suction speed. This arrangement, thus, leads to heat reduction and dust elimination within the electrical components/equipment in industrial space, thus preventing the destructive effects of the occurrence of hotspots. However, for the sake of graphical representations of this scheme, the MATLAB environment is created for the generation of the optimum temperatures at various locations within the electrical systems. From this development, it is established that the framework has very high potential to eliminate the catastrophic effects of hotspots in electrical systems.

In situations where maximum load demand exceeds power generation capacity, frequency would drop during cases of overload; and if appropriate under-frequency load curtailment scheme is not put in place, it could cause permanent turbine damage or affect the expected working operation of electric equipment. Various existing works on load curtailment in power systems have been done to great effect using meta-heuristic and classical control techniques. This paper seeks to comparatively analyze the Fuzzy Logic, Fuzzy-PID and Neuro-Fuzzy (ANFIS) controllers and their contribution to achieving load curtailment in a Single Area Power System via continuous load control. The transfer function model of the power system and the design of the controllers are done using Simulink, MATLAB software. The controllers are designed to output power imbalance based on frequency deviation during various system overload conditions. They also shed loads continuously until frequency is restored within the safe...

A metodologia FMEA – Failure Mode and Effects Analysis é citada em normas internacionais de gestão de riscos, e com aplicações já conhecidas em barragens e indústrias em geral, que por meio da análise de modos de falha e seus efeitos busca analisar riscos antes que os eventos danosos se concretizem. O objetivo principal desta pesquisa é propor e adaptar essa metodologia para a gestão de riscos de escorregamentos de solo no âmbito da Proteção Civil do Porto, de forma a possibilitar menores discrepâncias entre avaliadores e fomentar a discussão dos riscos em equipe. Para tal, as seguintes etapas são propostas: i) definição dos taludes a serem avaliados; ii) desenvolvimento de uma ficha de campo para análise de perigosidade e consequências atribuindo pontos para os índices de acordo com sua criticidade; iii) análise do risco com definição de cenários, e combinação dos resultados de perigosidade e consequências em uma matriz e planilha. Foram identificados oito taludes de solo com graus...

Present work was aimed to design Mamdani- Fuzzy Inference System (FIS), Sugeno -FIS and Sugeno-Adaptive Neuro- Fuzzy Inference System (ANFIS) model for the prediction of CPUE of fish. The system was implemented using MATLAB fuzzy toolbox. A prediction of CPUE was made using the models trained. The accuracy of fuzzy inference system models was compared using mean square error (MSE) and average error percentage. Comparative study of all the three systems provided that the results of Sugeno-ANFIS model (MSE =0.05 & Average error percentage=11.02%) are better than the two other Fuzzy Inference Systems. This ANFIS was tested with independent 28 dataset points. The results obtained were closer to training data (MSE=0.08 and Average error percentage=13.45%).

O presente artigo se trata do estudo de caso de um talude de solo na cidade de Caratinga/MG. O objetivo desta pesquisa foi avaliar a estabilidade do talude, através da análise do Fator de Segurança (FS) obtidos a partir da modelagem computacional pelo software GEO5. Assim inicialmente foi feita a análise do talude em campo, bem como a caracterização do material que compõe este talude, como as caraterísticas físicas do solo e a sondagem geotécnica. A partir dos resultados obtidos, foi observado que fatores de segurança aceitáveis (<1,50) em todos os métodos analisados. Com isso, pode-se concluir que talude se mostrar estável, mas é importante o seu monitoramento, pois qualquer mudança dentre os parâmetros avaliados no talude pode acarretar uma instabilização deste talude.

The stability of slopes is a topic of great interest to the geotechnical engineer, given the significant economic losses, and even human, resulting from the slopes collapse. It’s estimated that the landslides outbreak has already caused thousands of deaths and tens of billions of dollars in annual losses worldwide. The phenomena of instability of slopes are conditioned by many factors, such as climate, the lithology and structures of rock, the morphology, the anthropic and others. The analysis of geological and geotechnical conditions of landslides provides an appraisal of each of the factors involved in the processes of instability of slopes, allowing the achievement of results of interest with regard to the mode of action of factors. The current work aims at the use a Hybrid system that uses the Neural Network and Fuzzic Logic (Neuro-Fuzzy) to create a model that, in qualitative form, provides a prediction of the potential of slope rupture. To fulfill this objective, we studied the factors involved in the processes of instability of slopes, and how these factors are interrelated. Parametric analyzes were carried out to provide data for the Neuro-Fuzzy model. It is presented in this work, one history case well documented for the validation of the Neuro-Fuzzy Model and Among the main findings includes the capability of Hybrid Neuro-Fuzzy Modeling in predicting the potential of slope rupture, appearing as a tool capable of assisting in the detection of slopes with potential for rupture.

Se realizó un estudio para predecir el número de manchas solares en el Ciclo Solar 25 mediante el uso de dos modelos: un modelo de redes neuronales recurrentes Long short-term memory (LSTM) y un modelo Autoregressive Integrated Moving Average (ARIMA). Los datos utilizados para entrenar los modelos fueron obtenidos del sitio web del Centro Mundial de Datos SILSO, del Real Observatorio de Bélgica en Bruselas, desde 1749 hasta 2018. Nuestro modelo LSTM demostró un rendimiento excepcional (RMSE=3,6) en comparación con el mejor modelo ARIMA (RMSE=32,6). Esto demostró que nuestro modelo LSTM es significativamente más preciso en términos de predicción, con una mejora del 89% en la reducción del RMSE. Según nuestro modelo LSTM, se prevé que el número máximo de manchas solares en el Ciclo Solar 25 ocurra en marzo de 2025, alcanzando un valor máximo de 182 manchas solares. En contraste, el modelo ARIMA predice que el máximo se alcanzará en diciembre de 2024, con un valor máximo de 99 manchas ...

Este artículo, se centra en la predicción de la irradiación solar global diaria horizontal, por ser el caso más interesante en la meteorología agrícola, por ejemplo, en las previsiones de necesidades de riego, utilizando la técnica de las redes neuronales artificiales (RNAs) de la inteligencia computacional, a partir de variables accesibles en las estaciones agrometeorológicas. El lugar donde fueron medidos los datos, utilizados para entrenar las RNAs, caracterizan donde se pueden volver a utilizar este tipo de modelos, en este estudio fueron las estaciones meteorológicas de la red SIAR en Castilla y León, en concreto la situada en Mansilla Mayor (León), durante los años 2004-2010. Los modelos RNAs se construyeron en la entrada con los datos medidos de irradiación solar global diaria de uno, dos y tres días anteriores, añadiendo el día del año J(t)=1..365, para predecir su valor el día siguiente. Los resultados obtenidos, validados durante el año 2011 completo RMSE=3,8012 MJ/(m 2 d), concluyen que las RNAs estudiadas mejoran los métodos clásicos comparados: 1) año típico CENSOLAR RMSE=5,1829 MJ/(m 2 d), 2) media móvil ponderada con la autocorrelación parcial de 11 días de retardo RMSE=3,9810 MJ/(m 2 d), 3) regresión lineal sobre el valor del día anterior RMSE=4,2434 MJ/(m 2 d), 4) año típico Fourier utilizado el 1 er armónico RMSE=4,2747 MJ/(m 2 d), y 5) las matrices de transición de Markov para 33 estados posibles RMSE=4,3653 MJ/(m 2 d). Durante los días de cambio brusco en el nivel de irradiación solar, se observan los mayores errores de predicción. Se plantea utilizar en la entrada otras variables para mejorar la eficacia del modelo RNA. Una de las variables probadas fue el índice de claridad diario Kt=H/H0, resultando una mejora RMSE=3,7703 MJ/(m 2 d).

Locomotive rehabilitation devices are used by a large number of people around the world, thus, the devices reliability, its performance and quality equipment are an interesting topic for developers and their users. This paper shows how the hardware and software design of a rehabilitation locomotive platform contribute to be reliable its use, and it describes too, how the "embedded programming" is applied on the platform monitoring system. Locomotive medical devices are used for performing treatments non-invasive on the patients, they can achieve positive or negative effects on disease recovery, and then the actions for controlling monitoring those equipment, become the basic objective to achieve a successful physical recovery. This paper contains information about of the development of combined Neural Networks and Fuzzy Systems in monitoring and control using the Rehabilitation Locomotive Platform, also it relate variables such as weight-bearing on a patient, the treadmill speed system, the ECG patient's heart rate; aspects that involve programming, integration of parts, through ANFIS software use and its embedded software application using digital processors DSPs.

Resumen-Las redes neuronales en muchos casos han dado buenos resultados en la aproximación de funciones en muchas aplicaciones, aun así existen muchos problemas que no se han podido resolver. La selección de funciones de activación se realiza de acuerdo con el problema y a criterio del investigador, en ocasiones por ensayo y error. Comúnmente, la función de activación logística ha sido la más frecuentemente usada trayendo buenos resultados. En la literatura no existe un criterio estándar para la selección de estas funciones de activación en las redes neuronales, ni tampoco existe una exhaustiva investigación en este tema. Es por esta razón que el objetivo principal de este articulo es obtener un criterio de selección para tres funciones de activación en una red neuronal feedforward con una capa oculta, comparando su desempeño con múltiples neuronas, para aproximar las funciones objetivo propuestas en [4] las cuales fueron diseñadas para evaluar la capacidad de regresión de modelos de redes neuronales.

Locomotive rehabilitation devices are used by a large number of people around the world, thus, the devices reliability, its performance and quality equipment are an interesting topic for developers and their users. This paper shows how the hardware and software design of a rehabilitation locomotive platform contribute to be reliable its use, and it describes too, how the "embedded programming" is applied on the platform monitoring system. Locomotive medical devices are used for performing treatments non-invasive on the patients, they can achieve positive or negative effects on disease recovery, and then the actions for controlling monitoring those equipment, become the basic objective to achieve a successful physical recovery. This paper contains information about of the development of combined Neural Networks and Fuzzy Systems in monitoring and control using the Rehabilitation Locomotive Platform, also it relate variables such as weight-bearing on a patient, the treadmill speed system, the ECG patient's heart rate; aspects that involve programming, integration of parts, through ANFIS software use and its embedded software application using digital processors DSPs.

Objetivo: Proponer un sistema adaptativo de inferencia neuro-difusa (ANFIS) para la estimación de la radiación solar global en la ciudad de Cúcuta (Norte de Santander, Colombia) Metodología: A partir de registros históricos del IDEAM se obtuvieron las variables: temperatura ambiente, brillo solar, velocidad del viento, humedad relativa, hora de medición y radiación solar global entre los años 2005 y 2015. Usando la herramienta Fuzzy Logic Toolbox de Matlab y los datos del IDEAM se crearon redes ANFIS con diferentes configuraciones. Después de comparar los errores estadísticos, se escogió el modelo ANFIS que minimizara el RMSE. Resultados: Los errores estadísticos del modelo son: R2 = 0.9115, RMSE = 124,23 Wh/m2 y MAPE = 27,8 %, evidenciando una alta precisión para la estimación de la radiación solar global en el territorio seleccionado. Conclusiones: La red ANFIS propuesta es un modelo basado en inteligencia artificial con precisión suficiente para ser usado en el diseño de sistemas...

The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.

The aim of this study is to compare two architectures of recurrent neural networks of Elman and Jordan (RNRE and RNRJ), focus on the forecasting for two days of solar radiation and air temperature. The inputs of the forecasting model are meteorological variables as wind speed, atmospheric pressure, relative humidity and precipitation. The Research Institute for Geology and Energy of Ecuador provided the data of three meteorological stations situated in the provinces of Pichincha and Tungurahua for neural network training, validation and forecasting stages. Each network was trained with three different learning functions: backpropagation, backpropagation momentum and resilient propagation. The results shows the statistical parameters, Person correlation, mean square error and forecasting behavior on graphics for air temperature and solar radiation, according to RNRE and RNRJ model. This work shows correlation index greater than 0,9 in the validation stage. In the forecasting stage, the correlation index is higher than 0,8 and the mean square error shows values less than 0,02 kW for solar radiation and 2 ºC for air temperature.

Improved Nefclass for Datasets with Skewed Feature Values Jamileh Yousefi University of Guelph, 2018 Advisor: Dr. Andrew Hamilton-Wright Most machine learning algorithms perform poorly on datasets with skewed feature values distribution. Skewed feature values are commonly observed in biological and medical datasets. This poses a challenge for the classification of medical data. Neuro-fuzzy systems are common machine learning approaches in the medical domain because of their ability to learn fuzzy rules from training data and represent the rules in an understandable way. Therefore, addressing skewness in neuro-fuzzy systems is a topic of interest because of their applicability in the medical domain. In this thesis, the NEFCLASS neuro-fuzzy classifier is extended to provide improved classification accuracy over the original NEFCLASS classifier when trained on skewed data. In order to improve accuracy, we used two methods. Firstly, we used two alternative discretization methods. Second...

The accuracy-transparency trade-off is one of the most notable challenges when applying machine learning tools in the medical domain. NEFCLASS is a popular NEURO-FUZZY classifier in medical diagnosis systems. NEFCLASS performs increasingly poorly as the data skewness increases. This paper presents a combined approach to improve the classification accuracy and interpretability of the NEFCLASS classifier, when feature values of the training and testing datasets exhibit positive skewness. The proposed model consists of two steps. Firstly, a modified NEFCLASS classifier embedded with a choice of two alternative discretization methods, MME and CAIM is implemented. Secondly, we devised a new rule pruning method based on the Habermans' adjusted residual to reduce the size of the resulting ruleset. This rule-pruning method improves the interpretability of NEFCLASS without significant accuracy deterioration. Moreover, a hybrid approach combining the two approaches provides a considerable improvement in classification accuracy and transparency of NEFCLASS on skewed data.

This paper presents a model for the treatment of skewness effects on the accuracy of the NEFCLASS classifier by changing the embedded discretization method within the classifier. NEFCLASS is a common example of the construction of a neuro-fuzzy system. The popular NEFCLASS classifier exhibits surprising behaviour when the feature values of the training and testing data sets exhibit significant skew. As skewed feature values are commonly observed in biological data sets, this is a topic that is of interest in terms of the applicability of such a classifier to these types of problems. From this study it is clear that the effect of skewness on classification accuracy is significant, and that this must be considered in work dealing with skewed data distributions. We compared accuracy of NEFCLASS classifier with two modified versions of NEFCLASS embedded with MME and CAIM discretization methods. From this study it is found the CAIM and MME discretization methods results in greater improvements in the classification accuracy of NEFCLASS as compared to using the original EQUAL-WIDTH technique NEFCLASS uses by default.

The following article aims to classify obesity in children and male adolescents in a range from six to seventeen, using neural networks and fuzzy logic, for which the neuro-fuzzy model ANFIS (used artificial Neural Network Fuzzy Inference System) which is optimized using PSO (particle swarm optimization). Experimental tests show an error RMSE 8.31, after making 500 iterations of the algorithm PSO. This result is considered acceptable within the characteristics of this research conditions.

Nature. This e-offprint is for personal use only and shall not be self-archived in electronic repositories. If you wish to self-archive your article, please use the accepted manuscript version for posting on your own website. You may further deposit the accepted manuscript version in any repository, provided it is only made publicly available 12 months after official publication or later and provided acknowledgement is given to the original source of publication and a link is inserted to the published article on Springer's website. The link must be accompanied by the following text: "The final publication is available at link.springer.com".

In situations where maximum load demand exceeds power generation capacity, frequency would drop during cases of overload; and if appropriate under-frequency load curtailment scheme is not put in place, it could cause permanent turbine damage or affect the expected working operation of electric equipment. Various existing works on load curtailment in power systems have been done to great effect using meta-heuristic and classical control techniques. This paper seeks to comparatively analyze the Fuzzy Logic, Fuzzy-PID and Neuro-Fuzzy (ANFIS) controllers and their contribution to achieving load curtailment in a Single Area Power System via continuous load control. The transfer function model of the power system and the design of the controllers are done using Simulink, MATLAB software. The controllers are designed to output power imbalance based on frequency deviation during various system overload conditions. They also shed loads continuously until frequency is restored within the safe...

The poor management of hotspots in electrical systems often leads to very devastating consequences; an extremity is fire outbreak which could result in loss of lives and/or properties. Hence, the necessity to effectively manage or handle electrical hotspots cannot be overstated. This paper proposes a model for arresting the occurrence of the destructive activities of electrical hotspots in industries by controlling the temperature, humidity and dust density within electrical components/ equipment. To this end, a thermal imaging camera is employed for the detection of various locations and magnitudes of hotspots within electrical systems. Based on the globally approved industrial standards for the prevention of thermally induced electrical systems failure, each of the electrical components and equipment [whose thermal excursion is beyond the allowable temperature rise under measured load values (i.e. ΔT corr)] is identified and treated by adopting the recommended actions. Additionally, a fuzzy logic control (FLC) system is designed. This is further developed into an adaptive neuro-fuzzy inference system (ANFIS) for the control of the operation of the air handling unit (AHU) and the aspirator suction speed. This arrangement, thus, leads to heat reduction and dust elimination within the electrical components/equipment in industrial space, thus preventing the destructive effects of the occurrence of hotspots. However, for the sake of graphical representations of this scheme, the MATLAB environment is created for the generation of the optimum temperatures at various locations within the electrical systems. From this development, it is established that the framework has very high potential to eliminate the catastrophic effects of hotspots in electrical systems.

Hybrid models have attracted many researchers since last decade. Many hybrid models are proposed using soft computing skills. Hybrid fuzzy-neuro model is basically used for applications or projects that require high accuracy along with good comprehensibility. Neural Networks are known for their high accuracy, whereas fuzzy logic is known for its interpretable results. Fuzzy-Neuro model incorporate the goodness of both the systems, but is suffering from the curse of generating large and complex rule base. None of the existing rule pruning methods work upon an all length rule base. So a new rule pruning procedure for all length rules based on confidence and support is proposed here. A simplified rule base is obtained by the proposed system while getting the accuracy at par with existing system. The experiment carried out using the 3 well-known datasets for Credit risk analysis prove that the proposed rule pruning method for HFNN helps to improve upon its time complexity and efficiency.

El relleno de datos faltantes de precipitación diaria es un problema común en los estudios hidrológicos. El objetivo de este artículo es realizar una comparación y una evaluación de diferentes métodos que permiten rellenar los datos de precipitación diaria faltante para largos periodos en zona de llanura. Este estudio se realiza en la cuenca superior del arroyo del Azul para periodo de nueve años (2006-2014) y se emplean tres estaciones que tienen 3287 datos completos de precipitación diaria y seis estaciones con datos faltantes. Se implementaron siete métodos para el relleno de datos diarios de precipitación: el método de regresión lineal (MRL), el método de razones de distancia (MRD), el método de coeficientes de correlación con estaciones vecinas (MRC), el método de la razón promedio (MRP), método del inverso de la distancia al cuadrado (MIDW), seguido del método de cadenas de Márkov (MKV) y por último el método redes neuronales (MRN). Para la comparación y análisis de las difere...

Resumen-Las redes neuronales en muchos casos han dado buenos resultados en la aproximación de funciones en muchas aplicaciones, aun así existen muchos problemas que no se han podido resolver. La selección de funciones de activación se realiza de acuerdo con el problema y a criterio del investigador, en ocasiones por ensayo y error. Comúnmente, la función de activación logística ha sido la más frecuentemente usada trayendo buenos resultados. En la literatura no existe un criterio estándar para la selección de estas funciones de activación en las redes neuronales, ni tampoco existe una exhaustiva investigación en este tema. Es por esta razón que el objetivo principal de este articulo es obtener un criterio de selección para tres funciones de activación en una red neuronal feedforward con una capa oculta, comparando su desempeño con múltiples neuronas, para aproximar las funciones objetivo propuestas en [4] las cuales fueron diseñadas para evaluar la capacidad de regresión de modelos de redes neuronales.

Reinforced concrete slabs, just as the other structural elements, are highly affected by the high temperatures. Due to the decrease in strength of reinforced concrete members under high temperature, bearing moment capacity of reinforced concrete slabs also decreases. In this study, a prediction model is investigated in order to determine the bearing moment capacities of reinforced concrete slabs under high temperature. Pre-calculated moment capacities of slabs exposed to fire are predicted by the implementation of adaptive neuro-fuzzy inference system (ANFIS) and the prediction performance of ANFIS model is investigated. The bending capacities of slabs with different concrete characteristics and different times of exposure are calculated. High temperature resulting from the duration of fire exposure is calculated as a function of time in accordance with ISO 834. The temperature distribution inside the slab is determined by the adoption of a steady-state onedimensional heat transfer. The slab was separated into slim slices and the heat in each slice is determined depending on the time of exposure. Forces and resistance of materials under fire exposure are calculated by applying the reduction coefficients in Eurocode 2. Results confirm the high prediction capability of ANFIS model.

Additive Manufacturing describes the technologies that can produce a physical model out of a computer model with a layer-by-layer production process. Additive Manufacturing technologies, as compared to traditional manufacturing methods, have the high capability of manufacturing the complex components using minimum energy and minimum consumption. These technologies have brought about the possibility to make small pieces of raw materials in the shortest possible time without the need for a mold or tool. One of the technologies used to make pieces of the layer-by-layer process is the Gas Metal Arc Welding (GMAW). One of the basic steps in this method of making parts is the prediction of bead geometry in each pass of welding. In this study, taking into account the effective parameters on the geometry of weld bead, an empirical study has been done in this field. For this purpose, three parameters of voltage, welding speed and wire feeding rate are considered as effective parameters on th...

Esta investigación presenta la aplicación de un modelo de red neurodifusa llamado ANFIS para el problema de estimación de datos faltantes meteorológicos: temperatura, velocidad del viento, humedad relativa y radiación solar en la estimación de la Evapotranspiración de referencia ETo. ANFIS es un método que permite crear la base de reglas de un sistema difuso, utilizando el algoritmo de retro propagación a partir de los datos de un proceso. La estructura de la red neurodifusa para cada variable meteorológica consiste en dos entradas y una salida. La evaluación del relleno de datos faltantes se realiza mediante la Raíz Cuadrada del Error Cuadrático Medio (RMSE). Los resultados muestran que al usar un mayor número de iteraciones y variación de datos en el entrenamiento puede ayudar a la ANFIS a obtener resultados más precisos.

With the increasing demands for product variety and quality level the need to eliminate human operator from the feedback path for welding process correction is evident. Of the several manufacturing methods welding alone has defined true automation. Success of automation depends on effective and efficient decision making tools. Neural network is applied to intelligent weld control. In Submerged Arc Welding (SAW), selecting appropriate values for process variables is essential to control heat affected zone (HAZ) dimensions and get the required bead size and quality. Also, conditions must be selected that will ensure a predictable and reproducible weld bead. This paper proposes the modeling and prediction of dimensions of Heat-Affected Zone for SAW process using Finite Element Analysis (FEA) and Artificial Neural Network (ANN). The dimensions of HAZ for SAW are modeled are simulated using FEA using the process variables such as welding current, arc voltage, arc efficiency and welding speed and the results are used as the learning file for ANN model. The developed ANN model is forwarded to predict the dimensions of HAZ and the results are compared with simulated FEA results. The developed method is found to be time consuming, competent and cost effective.

This article presents the design process and implementation of a microprocessor for educational purposes based on basic digital components, with the goal of making it easier to understand the architecture and functioning of microprocessors. The process is carried out in three main sections, transfer instructions between registers, control flow (unconditional jumps), and logical and arithmetic operations. In order to perform the system check, analog to digital conversion ADC and digital to analog conversion DAC peripherals were added, with the goal of implementing an application of temperature measurement with an LM35 sensor, to obtain its magnitude in an industrial protocol of communication of 4mA to 20mA. The implementation of said microprocessor is carried out with the software ISE Design Suite 14.2 from XILINX, and is implemented on the Spartan-3 FPGA development board.

Meteorological time series forecasts can help decision-making processes carried out by entities in charge of disaster prevention and early warning generation before the possibility of natural events involving situations which are dangerous for communities.  There are a considerable number of methods for these forecasts, ranging from simplistic or Naïve methods to those which employ more complex techniques such as those using artificial intelligence.  This experimental study worked with a meteorological time series from the station agronomy in the city of Manizales, which provides data on the variables: precipitation, average temperature, sunlight and relative humidity.  Forecasts were employed with the Naïve approach, with artificial neuronal networks and with neuro-fuzzy networks; also comparing these with a multiple linear regression, with the goal of verifying their precision.  The results obtained in this study show firstly that it is possible to refine the models generally used...

Current classification problems that concern data sets of large and increasing size require scalable classification algorithms. In this study, we concentrate on several scalable, linear complexity classifiers that include one of the top 10 voted data mining methods, Naïve Bayes (NB), and several recently proposed semi-NB classifiers. These algorithms perform front-end discretization of the continuous features since by design they work only with nominal or discrete features. We address the lack of studies that investigate the benefits and drawbacks of discretization in the context of the subsequent classification. Our comprehensive empirical study considers 12 discretizers (two unsupervised and 10 supervised), seven classifiers (two classical NB and five semi-NB), and 16 data sets. We investigate the scalability of the discretizers and show that the fastest supervised discretizers fast class-attribute interdependency maximization (FCAIM), class-attribute interdependency maximization ...

The aim of this work is to explore and estimate the short-term prediction efficiency, using two alternative approaches: the fundamental analysis on factors affecting the Baltic Panamax Index (BPI) evolution, and a computational intelligence approach based on neuro-fuzzy technique. In order to accomplish this task, we first analyze the basic routes that compose the Baltic Panamax Index. We estimate the supply and demand on the BPI standard routes, the GDP annual growth rate of main importing countries, and the other fundamental factors and thus we conclude for the up trend or downtrend in a short time basis. Next, we develop a computer based system, consisting of a neural network and a wavelet analysis and filtering system, using as system's input, BPI time series data for the past three years. The configuration of such system, involves the application of wavelets filtering as a system preprocessor, in order to identify the underlying trend of a BPI signal. Then, applying wavelet analysis enables us to further configure a neural network, which is finally used as the anticipation engine, of this system. Results showed that both techniques are capable to produce a short-term decision support in the maritime or financial sector. Nevertheless, more research on both domains might improve the anticipation power in either the fundamental or the computational intelligence approach. As a concluding remark, we suggest that the whole procedure may be a valuable tool in order to identify the capers freight market, or other maritime time-series data, as well.

This study is aimed at obtaining a relationship between the values defining bead geometry and the welding parameters and also to select optimum welding parameters. For this reason, an experimental study has been realized. The welding parameters such as the arc current, arc voltage, and welding speed which have the most effect on bead geometry are considered, and the other parameters are held as constant. Four, three, and five different values for the arc current, the arc voltage, and welding speed are used, respectively. So, sixty samples made of St 52-3 material were prepared. The bead geometries of the samples are analyzed, and the thickness and penetration values of the weld bead are measured. Then, the relationship between the welding parameters is modeled by using artificial neural network (ANN) and neurofuzzy system approach. Each model is checked for its adequacy by using test data which are selected from experimental results. Then, the models developed are compared with rega...

Because of its high quality and reliability, Submerged Arc Welding (SAW) is one of the chief metals joining process employed in industry. In this paper, an attempt has been taken to develop a model to predict the yield characteristics (weld bead parameters) of Submerged Arc Welding (SAW) process with the help of neural network technique and analysis of various process control variables and important weld bead parameters in SAW. The SAW process has been chosen for this application because of the complex set of ...

Welding is a manufacturing process, which is carried out for joining of metals by submerged arc welding. This is the one type of Arc Welding, in which both the Arc and the weld metal are submerged beneath a layer of flux. This layer protects the weld metal from contamination and Concentrates the heat into the joint. The molten flux rises through the pool, de-oxidizing and cleaning the molten metal. It then forms a protective slag covering over the newly deposited weld. All welds will be prepared by submerged arc welding technique. We studied Design of Experiments for this work and by use of the experimental data have optimize by any one optimization techniques which has decide later. In which inputs parameters for submerged arc welding are welding current, arc voltage and welding speed and the output parameter are hardness, tensile strength and microstructure of material. We use stainless steel-304 material for welding. Small scale trial welding experiments, in the light of field jo...

To improve the corrosion resistant properties of carbon steel, usually cladding process is used. It is a process of depositing a thick layer of corrosion resistant material over carbon steel plate. Most of the engineering applications require high strength and corrosion resistant materials for long-term reliability and performance. By cladding these properties can be achieved with minimum cost. The main problem faced on cladding is the selection of optimum combinations of process parameters for achieving quality clad and hence good clad bead geometry. This paper highlights an experimental study to optimize various input process parameters (welding current, welding speed, gun angle, and contact tip to work distance and pinch) to get optimum dilution in stainless steel cladding of low carbon structural steel plates using gas metal arc welding (GMAW). Experiments were conducted based on central composite rotatable design with full replication technique, and mathematical models were dev...

This paper presents development of neuro hybrid model (NHM) to predict weld bead width in submerged arc welding. Experiments were designed using Taguchi's principles and results were used to develop a multiple regression model. Data set generated from Multiple Regression Analysis (MRA) was utilized in ANN model, which was trained with backpropagation algorithm in MATLAB platform and used to develop NHM to predict quality of weld. NHM is flexible and accurate than existing models for a better online monitoring system.

The fuzzy rule based classification system (FRBCS) design methods, whose fuzzy rules are in the form of if-then sentences, have been under intensive study during last years. One of the outstanding FRBCS design methods utilizing hedge algebras as a mathematical formalism is proposed in [9]. As in other methods, a difficult problem with the high-dimensional and multi-instance datasets needs to be solved. This paper presents an approach to tackle the high-dimensional dataset problem for the hedge algebras based classification method proposed in [9] by utilizing the feature selection algorithm proposed in [20]. The experimental results over eight high-dimensional datasets have shown that the proposed method saves much execution time than the original one, while retaining the equivalent classification performance as well as the equivalent FRBCS complexity. The proposed method is also compared with three classical classification methods based on the statistical and probabilistic approaches showing that it is a robust classifier.

This work uses Genetic Algorithm to predict the weld metal deposition in the Metal Active Gas (MAG) welding process for a given set of welding parameters; with the help of Adaptive Neuro Fuzzy Interference System (ANFIS), an artificial intelligent technique to make the input-output model a standard one for given set of data into its input range. Experiments are designed according to full factorial design of experiments and its experimental results are used to develop an ANFIS model. Multiple sets of data from experiments are utilized to train, test and validate the intelligent network. The trained network is used to predict the amount of weld metal deposition. The proposed ANFIS, developed using MATLAB functions, is flexible, and it scopes for a better online monitoring system. Genetic algorithm is used to optimize the predicted results which is validated with experimental results and found satisfactory. The optimization of the model shows the correct input range to achieve the opti...

To improve the corrosion-resistant properties of carbon steel cladding process is usually used. It is a process of depositing a thick layer of corrosion resistant material-over carbon steel plate. Most of the engineering applications require high strength and corrosion resistant materials for long-term reliability and performance. By cladding, these properties can be achieved with minimum cost. The main problem faced in cladding is the selection of optimum combinations of process parameters for achieving quality clad and hence good clad bead geometry. This paper highlights an experimental study to optimize various input process parameters (welding current, welding speed, gun angle, contact tip to work distance, and pinch) to get optimum dilution in stainless steel cladding of low-carbon structural steel plates using gas metal arc welding (GMAW). Experiments were conducted based on central composite rotatable design with full-replication technique and mathematical models were develop...

This paper presents development of neuro hybrid model (NHM) to predict weld bead width in submerged arc welding. Experiments were designed using Taguchi's principles and results were used to develop a multiple regression model. Data set generated from Multiple Regression Analysis (MRA) was utilized in ANN model, which was trained with backpropagation algorithm in MATLAB platform and used to develop NHM to predict quality of weld. NHM is flexible and accurate than existing models for a better online monitoring system.

Artificial intelligent tools such as expert systems, artificial neural network and fuzzy logic support decision-making are being used in intelligent manufacturing systems. Success of intelligent manufacturing systems depends on effective and efficient utilization of intelligent tools. Weld residual stress depends on different process parameters and its prediction and control is a challenge to the researchers. In this paper, intelligent predictive techniques artificial neural network (ANN) and fuzzy logic models are developed for weld residual stress prediction. The models are developed using Matlab toolbox functions. Data set required to train the models are obtained through finite element simulation. Results from the fuzzy model are compared with the developed artificial neural network model, and these models are also validated.

With the increasing demands for product variety and quality level the need to eliminate human operator from the feedback path for welding process correction is evident. Of the several manufacturing methods welding alone has defined true automation. Success of automation depends on effective and efficient decision making tools. Neural network is applied to intelligent weld control. In Submerged Arc Welding (SAW), selecting appropriate values for process variables is essential to control heat affected zone (HAZ) dimensions and get the required bead size and quality. Also, conditions must be selected that will ensure a predictable and reproducible weld bead. This paper proposes the modeling and prediction of dimensions of Heat-Affected Zone for SAW process using Finite Element Analysis (FEA) and Artificial Neural Network (ANN). The dimensions of HAZ for SAW are modeled are simulated using FEA using the process variables such as welding current, arc voltage, arc efficiency and welding speed and the results are used as the learning file for ANN model. The developed ANN model is forwarded to predict the dimensions of HAZ and the results are compared with simulated FEA results. The developed method is found to be time consuming, competent and cost effective.