Search

A critical task of structural health monitoring is damage detection and localization. Lamb wave propagation methods have been successfully applied for damage identification in plate-like structures. However, Lamb wave processing is still a challenging task due to its multimodal and dispersive characteristics. To address this issue, data-driven machine learning approaches as artificial neural network (ANN) have been proposed. However, the effectiveness of ANN can be improved based on its architecture and the learning strategy employed to train it. The present paper proposes a Multiple Particle Collision Algorithm (MPCA) to design an optimum ANN architecture to detect and locate damages in plate-like structures. For the first time in the literature, the MPCA is applied to find damages in plate-like structures. The present work uses one piezoelectric transducer to generate Lamb wave signals on an aluminum plate structure and a linear array of four transducers to capture the scattered signals. The continuous wavelet transform (CWT) processes the captured signals to estimate the time-of-flight (ToF) that is the ANN inputs. The ANN output is the damage spatial coordinates. In addition to MPCA optimization, this paper uses a quantitative entropy-based criterion to find the best mother wavelet and the scale values. The presented experimental results show that MPCA is capable of finding a simple ANN architecture with good generalization performance in the proposed damage localization application. The proposed method is compared with the 1-dimensional convolutional neural network (1D-CNN). A discussion about the advantages and limitations of the proposed method is presented.

The impact of the data assimilation process of air temperature and relative humidity from surface meteorological stations and sounding at airports in the terminal area of Rio de Janeiro is evaluated using the Weather Research and Forecast Data Assimilation system. Synthetic data of temperature, relative humidity and wind are generated in the locations of airport sensors by applying a white-noise perturbation in the forecast data. Results show a positive overall impact of the assimilation process with the removal of part of the noise in the observation data but keeping the effect of local conditions in the later timesteps of the simulation. In addition, with the assimilation process there is a global reduction of the error between the analysis data and the observation data. In the future, a neural network will be trained to emulate the data assimilation process to speed-up the assimilation process in the WRF model. Resumo O impacto do processo de assimilação de dados de temperatura do ar e umidade relativa de estações meteorológicas de superfície e sondagem em aeroportos na área terminal do Rio de Janeiro é avaliado usando o sistema Weather Research and Forecast Data Assimilation. Dados sintéticos de temperatura, umidade relativa e vento, são gerados nas localidades dos sensores dos aeroportos a partir da aplicação de uma perturbação gaussiana nos dados de previsão. Resultados mostram um resultado positivo do processo de assimilação com remoção de parte do ruído nos dados de observação, mas mantendo o efeito das condições locais nos instantes posteriores da simulação. Além disso, com o processo de assimilação há uma redução global no domínio do erro entre os dados da análise e os dados de observação. No futuro, uma rede neural será treinada para emular o processo de assimilação de dados para acelerar o processo de assimilação no modelo WRF.

In this work, a period of 2 years (2016–2017) of ionospheric total electron content (ITEC) from ionosondes operating in Brazil is compared to the International GNSS (Global Navigation Satellite System) Service (IGS) vertical total electron content (vTEC) data. Sounding instruments from the National Institute for Space Research (INPE) provided the ionograms used, which were filtered based on confidence score (CS) and C-Level flag evaluation. Differences between vTEC from IGS maps and ionosonde TEC were accumulated in terms of root mean squared error (RMSE). As expected, we noticed that the ITEC values provided by ionosondes are systematically underestimated, which is attributed to a limitation in the electron density modeling for the ionogram topside that considers a fixed scale height, which makes density values decay too rapidly above ∼800 km, while IGS takes in account electron density from GNSS stations up to the satellite network orbits. The topside density profiles covering the plasmasphere were re-modeled using two different approaches: an optimization of the adapted α-Chapman exponential decay that includes a transition function between the F2 layer and plasmasphere and a corrected version of the NeQuick topside formulation. The electron density integration height was extended to 20 000 km to compute TEC. Chapman parameters for the F2 layer were extracted from each ionogram, and the plasmaspheric scale height was set to 10 000 km. A criterion to optimize the proportionality coefficient used to calculate the plasmaspheric basis density was introduced in this work. The NeQuick variable scale height was calculated using empirical parameters determined with data from Swarm satellites. The mean RMSE for the whole period using adapted α-Chapman optimization reached a minimum of 5.32 TECU, that is, 23 % lower than initial ITEC errors, while for the NeQuick topside formulation the error was reduced by 27 %.

In this study, a method is presented for meteorological convective event forecasting at the terminal control area of the Galeao International Airport, Rio de Janeiro, Brazil, using machine ...

Nowadays, artificial neural networks have been successfully applied on several research and application fields. An appropriate configuration for a neural network is a complex task, and it often requires the knowledge of an expert on the application. A technique for automatic configuration for a neural network is formulated as an optimization problem. Two strategies are considered: a mono-objective minimization problem, using multi-particle collision algorithm (MPCA); and a multi-objective minimization problem addressed by the non-dominated sorting genetic algorithm (NSGA-II). The proposed optimization approaches were tested for two application in geosciences: data assimilation for wave evolution equation, and the mesoscale seasonal climate prediction for precipitation. Better results with automatic configuration were obtained for data assimilation than those obtained by network defined by an expert. For climate seasonal precipitation, automatic configuration presented better predictions were presented than ones carried out by an expert. For the worked examples, the NSGA-II presented a superior result for the worked experiments.

Many natural disasters in South America are linked to meteorological phenomena. Therefore, forecasting and monitoring climatic events are fundamental issues for society and various sectors of the economy. In the last decades, machine learning models have been developed to tackle different issues in society, but there is still a gap in applications to applied physics. Here, different machine learning models are evaluated for precipitation prediction over South America. Currently, numerical weather prediction models are unable to precisely reproduce the precipitation patterns in South America due to many factors such as the lack of region-specific parametrizations and data availability. The results are compared to the general circulation atmospheric model currently used operationally in the National Institute for Space Research (INPE: Instituto Nacional de Pesquisas Espaciais), Brazil. Machine learning models are able to produce predictions with errors under 2 mm in most of the continent in comparison to satellite-observed precipitation patterns for different climate seasons, and also outperform INPE’s model for some regions (e.g., reduction of errors from 8 to 2 mm in central South America in winter). Another advantage is the computational performance from machine learning models, running faster with much lower computer resources than models based on differential equations currently used in operational centers. Therefore, it is important to consider machine learning models for precipitation forecasts in operational centers as a way to improve forecast quality and to reduce computation costs.

Emphasizes some aspects of the aviation forecasting system under construction for use by the integrated meteorological center (CIMAER) in Brazil. It consists of a set of hybrid models based on determinism and machine learning that use remote sensing data (such as lighting sensor, SODAR, satellite and soon RADAR), in situ data (from the surface weather station and radiosonde) and aircraft data (such as retransmission of aircraft weather data and vertical acceleration). The idea is to gradually operationalize the system to assist CIMAER´s meteorologists in generating their nowcasting, for example, of visibility, ceiling, turbulence, convective weather, ice, etc. with objectivity and precision. Some test results of the developed nowcasting models are highlighted as examples of nowcasting namely: a) visibility and ceiling up to 1h for Santos Dumont airport; b) 6-8h convective weather forecast for the Rio de Janeiro area and the São Paulo-Rio de Janeiro route. Finally, the steps in development and the futures are superficially covered.

Made available in DSpace on 2018-12-11T17:23:20Z (GMT). No. of bitstreams: 0 Previous issue date: 2017-01-13 Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) We present a new version of the Brazilian developments on the Regional Atmospheric Modeling System (BRAMS), in which different previous versions for weather, chemistry, and carbon cycle were unified in a single integrated modeling system software. This new version also has a new set of state-of-the-art physical parameterizations and greater computational parallel and memory usage efficiency. The description of the main model features includes several examples illustrating the quality of the transport scheme for scalars, radiative fluxes on surface, and model simulation of rainfall systems over South America at different spatial resolutions using a scale aware convective parameterization. Additionally, the simulation of the diurnal cycle of the convection and carbon dioxide concentration over the Amazon Basin, as well as carbon dioxide fluxes from biogenic processes over a large portion of South America, are shown. Atmospheric chemistry examples show the model performance in simulating near-surface carbon monoxide and ozone in the Amazon Basin and the megacity of Rio de Janeiro. For tracer transport and dispersion, the model capabilities to simulate the volcanic ash 3-D redistribution associated with the eruption of a Chilean volcano are demonstrated. The gain of computational efficiency is described in some detail. BRAMS has been applied for research and operational forecasting mainly in South America. Model results from the operational weather forecast of BRAMS on 5km grid spacing in the Center for Weather Forecasting and Climate Studies, INPE/Brazil, since 2013 are used to quantify the model skill of near-surface variables and rainfall. The scores show the reliability of BRAMS for the tropical and subtropical areas of South America. Requirements for keeping this modeling system competitive regarding both its functionalities and skills are discussed. Finally, we highlight the relevant contribution of this work to building a South American community of model developers. Centro de Previsão de Tempo e Estudos Climáticos Instituto Nacional de Pesquisas Espaciais Divisão de Ciência da Computação Instituto Tecnológico de AeronáUtica Universidade Estadual Paulista (Unesp) Faculdade de Ciências Centro de Meteorologia de Bauru (IPMet) Departamento de Ciências Ambientais Universidade Federal de São Paulo Instituto de Astronomia Geofísica e Ciências Atmosféricas Universidade de São Paulo Departamento de Ciências Atmosféricas Universidade Federal de Campina Grande Embrapa Informática Agropecuária Instituto de Ciência e Tecnologia Universidade Federal de São Paulo Departamento de Ciências Atmosféricas e Climáticas Programa de Pós Graduação em Ciências Climáticas Universidade Federal Do Rio Grande Do Norte Centro de Ciências Do Sistema Terrestre Instituto Nacional de Pesquisas Espaciais Departamento de Geociências Universidade Federal de São João Del-Rei Laboratório Associado de Computação e Matemática Aplicada Instituto Nacional de Pesquisas Espaciais Instituto de Ciências Agrárias e Ambientais Mediterrânicas Universidade de Évora Centro Interdisciplinar de Desenvolvimento em Ambiente Gestão Aplicada e Espaço Universidade Lusófona de Humanidades e Tecnologia Campo Grande Faculdade de Meteorologia Universidade Federal de Pelotas Universidade Tecnológica Federal Do Paraná Universities Space Research Association Goddard Earth Sciences Technology and Research at the Global Modeling and Assimilation Office NASA Goddard Space Flight Center Universidade Estadual Paulista (Unesp) Faculdade de Ciências FAPESP: 2014/01563-1 FAPESP: 2014/01564-8 FAPESP: 2015/10206-0 CNPq: 306340/2011-9

Unmanned Aerial Vehicles (UAV) is a technology under strong development, with application on several fields. For the UAV autonomous navigation, a standard scheme is to use signal from a Global Navigation System by Satellite (GNSS) onboard. However, such signal can suffer natural or human interference. Our approach applies image processing procedure for the UAV positioning: image edge extraction and correlation between drone image and georeferenced satellite image. A data fusion is also applied, for combining the inertial sensor data and positioning by image. The data fusion is performed by using neural network. The output from the data fusion neural network is the correction for the UAV trajectory. Here, the variance of the trajectory error is also predicted to quantify the uncertainty.

Artificial neural networks have been employed on many applications. Good results have been obtained by using neural network for the precipitation climate prediction to the Brazil. The input are some meteorological variables, as wind components for several levels, air temperature, and former precipitation. The neural network is automatically configured, by solving an optimization problem with Multi-Particle Collision Algorithm (MPCA) metaheuristic. However, it is necessary to address, beyond the prediction the uncertainty associated to the prediction. This paper is focused on two-fold. Firstly, to produce a monthly prediction for precipitation by neural network. Secondly, the neural network output is also designed to estimate the uncertainty related to neural prediction.

This study aimed to analyze 70 years historical series in the Brazilian Southeastern region, using monthly rainfall data. Statistical modeling techniques such as cross-wavelet spectra and artificial neural networks (ANN), from the R statistical package, were used to perform the analyses. Two different types of neural networks were employed: the multi-layer perceptron (MLP) and extreme learning machine (ELM). From the cited time series, the analysis shows the existence of a decadal and multi-decadal signal with cycles of 5, 11, and 22 years in the monthly rainfall in Brazilian Southeastern region, observing the existence of low-frequency variability. This shows a significant degree of modulation and association for the precipitation with solar activity. The neural networks were also used as forecasting tools, with a better performance for MLP-NN – smaller root mean square error. However, the MLP-NN presented a greater confidence interval than ELM-NN. Keywords: monthly rainfall, sunspots, multi-decadal cycles. Observando a Existencia de Variabilidade de Baixa Frequencia em Dados Mensais de Chuva no Sudeste do Brasil utilizando Ferramentas do Aplicativo R - Redes Neurais e Ondeletas RESUMO . Este estudo teve como objetivo analisar series historicas de 70 anos no sudeste do Brasil, utilizando dados mensais de precipitacao. Tecnicas de analise estatistica usando o pacote estatistico R, como espectros de wavelet cruzado e modelagem de redes neurais artificiais (RNA), foram usadas para realizar as analises. Duas implementacoes de redes neurais foram empregadas: multi-layer perceptron (MLP) e extreme learning machine (ELM). Os resultados obtidos nas analises realizadas permitem inferir que as series temporais observadas mostram a existencia de um sinal decenal e multi-decenal com ciclos de 5, 11 e 22 anos na precipitacao mensal no sudeste do Brasil, observando a existencia de variabilidade de baixa frequencia nos dados analisados. Isso mostra um grau significativo de modulacao e associacao da precipitacao com a atividade solar. A analise de series temporais longas permitem a observacao de variabilidades de baixa frequencia, evidenciando sua grande importância e relevância. Uma significativa parcela da variância total de ciclos atmosfericos decenais e modulado pela atividade solar. As redes neurais tambem foram usadas como ferramentas de previsao, com melhor desempenho para a rede MLP – como mostrado pelo erro medio quadratico. A rede MLP apresentou maior amplitude no intervalor de confianca do que a rede ELM. Palavras-chave: precipitacao mensal, manchas solares, ciclos multi-decenais.

This study investigates the use of neural networks for data assimilation of local data in the WRF model in Rio de Janeiro, Brazil. Surface and upper-air data (air temperature, relative humidity, wind speed and direction) from airport stations and 6-hour forecast from WRF are used as input for the model and the 3D-Var analysis for each grid point is used as target variable. Periods of 168h from 2014 and 2015 are used with 6h and 12h assimilation cycles for surface and upper-air data, respectively. The neural network model was built using the Multi-Particle Collision Algorithm (MPCA) where different topologies are tested until the optimum solution is found. Results show that the neural network is able to emulate the 3D-Var with root mean squared error (standard deviation), respectively, of 0.31 K (0.37 K), 3.10% (4.04%), 0.63 ms -1 (1.05 ms -1 ), 1.10 ms -1 (1.56 ms -1 ) for air temperature, relative humidity, u-component of the wind and v-component of the wind. Also, the results show the neural network method is able to run 71 times faster than the conventional method under similar hardware configurations. Keywords: data assimilation; weather research and forecasting; surface data; profile data. Redes Neurais Artificiais para Assimilacao de Dados no Modelo WRF no Rio de Janeiro, Brazil RESUMO. Este estudo investiga o uso de redes neurais para assimilacao de dados locais no modelo WRF no Rio de Janeiro. Dados de superficie e do ar superior (temperatura do ar, umidade relativa, velocidade e direcao do vento) das estacoes do aeroporto e previsao de 6 horas do WRF sao usados como entrada para o modelo, e a analise 3D-Var para cada ponto da grade e usada como variavel destino. Periodos de 168h de 2014 e 2015 sao utilizados com ciclos de assimilacao de 6h e 12h para dados de superficie e do ar superior, respectivamente. O modelo de rede neural foi construido usando o algoritmo de colisao de particulas multiplas (MPCA), onde diferentes topologias sao testadas ate que a solucao ideal seja encontrada. Os resultados mostram que a rede neural e capaz de emular o 3D-Var com raiz do erro quadratico medio (desvio padrao) de 0,31 K (0,37 K), 3,10% (4,04%), 0,63 ms -1 (1,05 ms -1 ), 1,10 ms -1 (1,56 ms -1 ) para temperatura do ar, umidade relativa, componente u do vento e componente v do vento. Alem disso, os resultados mostram que o metodo de rede neural e capaz de rodar 71 vezes mais rapido que o metodo convencional em configuracoes de hardware semelhantes. Palavras-chave: assimilacao de dados; dados de superficie; dados de perfil.

We present here the application of an image processing strategy for autonomous navigation in order to estimate the position of an Unmanned Aerial Vehicle (UAV) equipped with Synthetic Aperture Radar (SAR) sensors. Patches of the SAR images are compared to georeferenced satellite optical images, by first applying an image edge detection algorithm on all the images, and the UAV position is determined from the correlation matrix obtained from resulting segmented images. The performance of three edge detectors are compared: the Canny approach, an artificial neural network (radial base function), and a fuzzy system.

Unmanned aerial vehicle (UAV) is a technology employed for several applications nowadays. One important UAV research topic is the autonomous navigation (AN). The standard procedure for AN is to fuse the signals from an inertial navigation system (INS) and from a global navigation satellite system (GNSS). Our approach to perform the autonomous navigation uses a computer vision system, instead of GNSS signal, associated to the visual odometry. The two techniques applied to estimate the UAV position are combined by a non-extensive particle filter. However, the development of a computer vision system for estimating the UAV position in a situation of flight over water-covered areas and flight in low light conditions is a challenge. Our approach uses images from an active sensor called light detection and ranging (LiDAR) to allow the flight in such conditions.

The increasing volume of data in the area of environmental sciences needs analysis and interpretation. Among the challenges generated by this “data deluge”, the development of efficient strategies for the knowledge discovery is an important issue. Here, statistical and tools from computational intelligence are applied to analyze large data sets from meteorology and climate sciences. Our approach allows a geographical mapping of the statistical property to be easily interpreted by meteorologists. Our data analysis comprises two main steps of knowledge extraction, applied successively in order to reduce the complexity from the original data set. The goal is to identify a much smaller subset of climatic variables that might still be able to describe or even predict the probability of occurrence of an extreme event. The first step applies a class comparison technique: p-value estimation. The second step consists of a decision tree (DT) configured from the data available and the p-value analysis. The DT is used as a predictive model, identifying the most statistically significant climate variables of the precipitation intensity. The methodology is employed to the study the climatic causes of an extreme precipitation events occurred in Alagoas and Pernambuco States (Brazil) at June/2010.

In this work, a period of two years (2016–2017) of vertical total electron content (VTEC) from ionosondes operating in Brazil is compared to the International GNSS Service (IGS) data. Sounding instruments from National Institute for Space Research (INPE) provided the ionograms used, which were filtered based on confidence score (CS) and C-level flags evaluation. Differences between TEC from IGS maps and ionograms were accumulated in terms of root mean square error (RMSE). It has been noticed the TEC values provided by ionograms are systematically underestimated, which is attributed to a limitation in the electron density modeled for the ionogram topside that considers maximum height only around 800–900 Km, while IGS takes in account electron density from GNSS stations up to the satellite network orbits. The ionogram topside profiles covering the plasmasphere were re-modeled using an adaptive alpha-Chapman exponential decay that includes a transition function between the F2 layer and plasmasphere, and electron density integration height was extended to compute TEC. Chapman parameters for the F2 layer were extracted from each ionogram, and plasmaspheric scale height was set to 10,000 Km. Our analysis has shown the plasmaspheric basis electron density, assumed to be proportional to the electron peak density, plays an important role to reduce the RMSE values. Depending on the proportionality coefficient choice, mean RMSE reached a minimum of 5.32 TECU, that is 23 % lower than initial ionograms TEC errors.

The authors wish to make the following corrections to this paper [...]

Massive and super-massive black-holes (BH) can be found in the center of all galaxies, according to the opinion from the scientific community. However, an open question is to know the origin of such BHs. One explanation is to consider such objects coming from the stars with large redshifts—very old stars collapsed during the ancient eras (population-III stars: the first-born stars created in the universe). A framework to investigate this scenario is to derive a mathematical model describing the evolution of the black hole distribution. From nowadays, observations show super-massive black holes with mass about 109M⊙ are ubiquitous in galaxy centers. The regularized inverse solution can be computed using an optimizer to identify the best solution for the functional to be minimized. Preliminary results are shown using synthetic observational data.

Application and development of Unmanned Aerial Vehicles (UAV) have had a rapid growth. The flight control of these aircrafts can be performed remotely or autonomously. There are different strategies for the UAV autonomous navigation. The positioning estimation can be done by using inertial sensors and General Navigation Satellite Systems (GNSS). The use of the GNSS signal can present some difficulties: natural or not natural interference. An alternative for positioning adjustment is to use a data fusion from different sensors by a Kalman filter. A supervised artificial network (ANN) is trained to emulate the filter for reducing the computational effort. An automatic best topology for the neural network is obtained by minimizing a functional by a new meta-heuristic called Multi-Particle Collision Algorithm (MPCA). Our results show similar accuracy between the ANN and the Kalman filter, with better processing performance to the neural network.

A class of sequential Monte Carlo estimation is frequently called particle filter. This filter belongs to the Bayesian strategy for estimation, where a non-linear and non-Gaussian assumptions can be applied. Here, the Tsallis’ distribution, from the non-extensive thermo-statistics, is used to design the best likelihood operator. Therefore, no previous likelihood operator is assumed. The new filter formulation will be named as non-extensive particle filter (NEx-PF). The distribution estimated by the NEx-PF can compute the standard form of the central limit theorem, as well as the Levy-Gnedenko central limit theorem. The q-calculus formalism is employed to generalize some definitions and properties.

Parallel Genetic Algorithm (PGA) is employed to solve inverse problems. The PGA is codified considering the island model (individuals are free to migrate to any other processor, subjected to specific rules); and stepping-stone model (migration is allowed only for the closest processors). The parallel code is generated using calls to the message passing communication library MPI (Message Passing Interface). In our GA approach, a new genetic operator, named epidemic is applied. This technique is employed to solve an inverse heat conduction problem. The goal is to determine the initial temperature from the transient noisy temperature profile at a given time. This ill-posed problem requires the use of a regularization technique.

Temperature difierences data sets obtained from a telescope can be transformed into a Cosmic Microwave Background (CMB) map. Our formulation for addressing the CMB mapping is based on an inverse problem methodology minimizing an optimization problem. Noisy synthetic date representing the BEAST experiment 30-day mission were used to test our approach. The computed CMB maps) are similar to those obtained by the bin average method. Our approach produced better results in similar processing time.

Experimentally characterizing evapotranspiration (ET) in different biomes around the world is an issue of interest for different areas of science. ET in natural areas of the Brazilian Pampa biome has still not been assessed. In this study, the actual ET (ETact) obtained from eddy covariance measurements over two sites of the Pampa biome was analyzed. The objective was to evaluate the energy partition and seasonal variability of the actual ET of the Pampa biome. Results showed that the latent heat flux was the dominant component in available energy in both the autumn&ndash, winter (AW) and spring&ndash, summer (SS) periods. Evapotranspiration of the Pampa biome showed strong seasonality, with highest ET rates in the SS period. During the study period, approximately 65% of the net radiation was used for the evapotranspiration process in the Pampa biome. The annual mean ET rate was 2.45 mm d&minus, 1. ET did not show to vary significantly between sites, with daily values very similar in both sites. The water availability in the Pampa biome was not a limiting factor for ET, which resulted in a small difference between the reference ET and the actual ET. These results are helpful in achieving a better understanding of the temporal pattern of ET in relation to the landscape of the Pampa biome and its meteorological, soil, and vegetation characteristics.

The Helmholtz-Hodge decomposition is a very useful instrument for vector field interpretation because it traditionally distinguishes conservative (curl-free) components from mass-preserving (divergence-free) components. In this work, we propose to use the Helmholtz-Hodge decomposition as a basic tool for the analysis of an ensemble of 2D vector fields. Given a vector field ensemble E , our approach firstly generates the corresponding curl-free, divergence-free, and harmonic component ensembles by using the Natural Helmholtz-Hodge decomposition to each vector field in E . With these ensembles in hand, our proposal not only quantifies, via a statistical analysis, how much each component ensemble is point-wisely correlated to the original vector field ensemble, but also allows to investigate the uncertainty of curl-free, divergence-free, and harmonic components separately. We test the proposed method with a synthetic vector field ensemble and with a simulated vector field ensemble. Graphical abstractDisplay Omitted HighlightsWe propose a new tool for vector field uncertainty analysis.It is based on the Helmholtz-Hodge decomposition.It uses the curl-free and divergence-free orthogonal components for uncertainty investigation.It uses a rank correlation coefficient to quantify the contribution of each component to the total uncertainty.

Precipitation is the hardest meteorological field to be predicted. An approach based on and optimal neural network is applied for climate precipitation prediction for the Brazil. A self-configurated multi-layer perceptron neural network (MLP-NN) is used as a predictor tool. The MLP-NN topology is found by solving an optimization problem by the Multi-Particle Collision Algorithm (MPCA). Prediction for Summer and Winter seasons are shown. The neural forecasting is evaluated by using the reanalysis data from the NCEP/NCAR and data from satellite GPCP (Global Precipitation Climatology Project -- monthly precipitation dataset).

The process of data assimilation, in which meteorological observations and weather forecasts are merged to provide an analysis field, has been largely studied by the scientific community and operational centers. The 3D-Variational (3D-Var) approach available in the Weather Research and Forecast (WRF) computer model is evaluated for data assimilation for the Terminal Control Area of Rio de Janeiro (TCA-RJ). The basic goal of any variational data assimilation system is to produce an optimal estimate of the atmospheric state at analysis time. The analysis field is estimated from a first guess (previous forecast) and an observation field, weighted by the error matrices. The WRF is designed for nowcasting (forecasts up to 6h) for the TCA-RJ through assimilation cycles using surface, sounding, and wind profile data. The preliminary results show the model sensibility for each observation type and encourage the use of this technique operationally for the support of the air traffic management controlled by the Brazilian Air Force.

A new module has been developed by the limited area model BRAMS to parameterize a wind farm power plant. BRAMS is developed to be a system for numerical weather prediction, and it is currently employed for operational environmental prediction by the CPTEC/INPE. The new module was able to simulate the impact from a set of wind turbines. The additional drag from turbines changes the turbulent kinetic energy (TKE). Some simulations were carried out considering wind power plant installed on the Brazilian territory.

A hybrid metaheuristic combining the Multi-Particle Collision Algorithm (MPCA) with the Hooke–Jeeves (HJ) method is applied to identify structural damage. A new version of the MPCA is formulated with the rotation-based learning mechanism to the exploration search. The inverse problem of damage identification is formulated as an optimization problem assuming the displacement time history as experimental data. The objective function is the square difference between the measured displacement and the displacement calculated by the forward model. The approach was tested on a cantilevered beam structure. Time-invariant damages were assumed to generate the synthetic displacement data. Noiseless and noisy data were considered. Finite element method was used for solving the direct problem. The comparison with standard MPCA-HJ and the new version of the hybrid method are reported. The use of these hybrid algorithms allows to obtain good estimations using a full set of data, or using a reduced dataset with a low level of noise in data.

The dynamics of helicopter main rotor is described by a highly nonlinear equations, and it is responsible to provide the sustentation of the rotorcraft. The representative frequency to this dynamics are associated to different regimes: unstable, periodic, and even chaotic behavior. The system responses are investigate performing several simulations under a set of numerical values of frequency. One important subject is to evaluate the goodness of the prediction of the simulated dynamics. The dynamical analysis is performed by bred vector approach. The breeding technique executes the model with a perturbed initial condition. The difference between the reference and the perturbed dynamics is called bred vector. The procedure can be employed systematically, producing a time series of bred vector. The bred vector magnitude is applied for addressing the predictability of the model, i.e., the degree of confidence from the simulation.

An introduction to metaheuristics for optimization problems is presented in this chapter. In Sect. 3.1 a classification of metaheuristics is put forward. The metaheuristics Differential Evolution (DE); Particle Collision Algorithm (PCA); Ant Colony Optimization (ACO) in its version for continuous problems; and Particle Swarm Optimization (PSO) are described in Sects. 3.2, 3.3, 3.4 and 3.5, respectively. These metaheuristics were applied to he benchmark problems diagnosis described in Chap. 2, based on Fault Diagnosis-Inverse Problem Methodology (FD-IPM) as described in Chap. 2.

This chapter introduces the main concepts and ideas concerning Model based Fault Diagnosis in Sect. 1.1. Fault Diagnosis based on parameter estimation is presented in Sect. 1.2 and Inverse Problems are briefly introduced in Sect. 1.3.

This chapter presents the application of the last three steps (2, 3, and 4) of the Fault Diagnosis—Inverse Problem Methodology (FD-IPM), as described in Sect. 2.1, to the three benchmark problems that were presented in Sect. 2.4. Step 2 is to evaluate the faults for which only the detection is possible. The third step deals with the solution of an optimization problem with metaheuristics. The fourth step provides the conclusion on the diagnosis of the system, based on the results from the third step. The experiments with the DC Motor, Inverted Pendulum System and Two Tanks System are presented in Sects. 4.2, 4.3 and 4.4, respectively. The Inverted Pendulum System is affected by faults which cannot be diagnosed, only detected. For that reason, Step 2 was only applied to the The Inverted Pendulum System.

This chapter deals with FDI as an Inverse Problem. Section 2.1 presents and formalizes Fault Diagnosis as an Inverse Problem. In Sect. 2.1, it is also formalized a new methodology for Fault Diagnosis: Fault Diagnosis—Inverse Problem Methodology (FD-IPM). Section 2.2 describes the mathematical models to be used by FD-IPM. Section 2.3 describes an alternative approach for determining uniqueness in Fault Diagnosis Inverse Problems, which is based on the structural analysis of the model that describes the system.