Simulação de redes neurais artificiais para estimativa de volume de madeira florestal a partir do DAP / Simulation of artificial neural networks for estimation of forest wood volume from DAP

2021 ◽  
Vol 4 (3) ◽  
pp. 3748-3757
Author(s):  
Emanuele de Oliveira Valente ◽  
Gerson de Freitas Silva Valente

O inventário florestal usa a altura (H) e o diâmetro na altura do peito (DAP) para determinação do volume de madeira. O grande problema é ajustar equações hipsométricas adequadas para estimar a altura. As equações são testadas e avaliadas por critérios estatísticos. A medição de alturas em povoamentos florestais é uma atividade onerosa, uma vez que, em comparação com a medição do diâmetro, sua obtenção não é fácil. Uma alternativa promissora consiste no uso de redes neurais artificiais (RNA), sistemas computacionais paralelos constituídos por unidades de processamento simples conectadas entre si de maneira específica para desempenhar determinada tarefa. Diante disso, o trabalho propõe usar técnicas de inteligência artificial para estimação do volume de madeira em uma floresta usando somente o DAP. Baseado em dados da literatura (Pesquisa: Modelos para quantificação do volume de diferentes sortimentos em plantio de Eucalyptus urophylla X Eucalyptus grandis.), foram gerados dados aleatórios de H e DAP entre o mínimo do DAP e H, 2,0 cm e 4,4 m, e o máximo, 28,9 cm e 32,9 m, respectivamente, no Excel. A partir desses dados, calculou-se o volume estimado de madeira pela fórmula de Takata (H e DAP), escolhida entre outros modelos por apresentar baixo erro-padrão da estimativa, além do baixo erro-padrão da estimativa em percentagem (Syx%), o modelo de Takata tem apresentado um alto coeficiente de determinação no inventario de fazendas de eucalipto. Sendo assim, este modelo se mostrou ligeiramente superior aos demais, tornando-se o mais adequado para estimativa da variável volume total em povoamentos de Eucalyptus urograndis, em Brasília. A partir dos dados treinou-se uma rede neural artificial (RNA) utilizando como variável contínua de entrada o DAP e de saída o volume calculado pela equação de Takata. A rede neural foi obtida pelo software PYTHON usando a função MLPRegressor (verbose=True, max_iter=10000, hidden_layer_sizes=(n)). Com isso foram testadas várias redes neurais artificiais para realizar a regressão entre volume de madeira calculado pela equação de Takata e DAP, e a melhor Rede Neural Artificial foi com 10 neurônios na camada oculta, apresentou r=0,99. A RNA é uma ferramenta do machine learning para o inventario florestal na determinação do volume de madeira a partir do DAP, principalmente quando já se tem banco de dados.

Sensors ◽  
2021 ◽  
Vol 21 (5) ◽  
pp. 1654
Author(s):  
Poojitha Vurtur Badarinath ◽  
Maria Chierichetti ◽  
Fatemeh Davoudi Kakhki

Current maintenance intervals of mechanical systems are scheduled a priori based on the life of the system, resulting in expensive maintenance scheduling, and often undermining the safety of passengers. Going forward, the actual usage of a vehicle will be used to predict stresses in its structure, and therefore, to define a specific maintenance scheduling. Machine learning (ML) algorithms can be used to map a reduced set of data coming from real-time measurements of a structure into a detailed/high-fidelity finite element analysis (FEA) model of the same system. As a result, the FEA-based ML approach will directly estimate the stress distribution over the entire system during operations, thus improving the ability to define ad-hoc, safe, and efficient maintenance procedures. The paper initially presents a review of the current state-of-the-art of ML methods applied to finite elements. A surrogate finite element approach based on ML algorithms is also proposed to estimate the time-varying response of a one-dimensional beam. Several ML regression models, such as decision trees and artificial neural networks, have been developed, and their performance is compared for direct estimation of the stress distribution over a beam structure. The surrogate finite element models based on ML algorithms are able to estimate the response of the beam accurately, with artificial neural networks providing more accurate results.


Author(s):  
Odysseas Kontovourkis ◽  
Marios C. Phocas ◽  
Ifigenia Lamprou

AbstractNowadays, on the basis of significant work carried out, architectural adaption structures are considered to be intelligent entities, able to react to various internal or external influences. Their adaptive behavior can be examined in a digital or physical environment, generating a variety of alternative solutions or structural transformations. These are controlled through different computational approaches, ranging from interactive exploration ones, producing alternative emergent results, to automate optimization ones, resulting in acceptable fitting solutions. This paper examines the adaptive behavior of a kinetic structure, aiming to explore suitable solutions resulting in final appropriate shapes during the transformation process. A machine learning methodology that implements an artificial neural networks algorithm is integrated to the suggested structure. The latter is formed by units articulated together in a sequential composition consisting of primary soft mechanisms and secondary rigid components that are responsible for its reconfiguration and stiffness. A number of case studies that respond to unstructured environments are set as examples, to test the effectiveness of the proposed methodology to be used for handling a large number of input data and to optimize the complex and nonlinear transformation behavior of the kinetic system at the global level, as a result of the units’ local activation that influences nearby units in a chaotic and unpredictable manner.


Author(s):  
Mehmet Fatih Bayramoglu ◽  
Cagatay Basarir

Investing in developed markets offers investors the opportunity to diversify internationally by investing in foreign firms. In other words, it provides the possibility of reducing systematic risk. For this reason, investors are very interested in developed markets. However, developed are more efficient than emerging markets, so the risk and return can be low in these markets. For this reason, developed market investors often use machine learning techniques to increase their gains while reducing their risks. In this chapter, artificial neural networks which is one of the machine learning techniques have been tested to improve internationally diversified portfolio performance. Also, the results of ANNs were compared with the performances of traditional portfolios and the benchmark portfolio. The portfolios are derived from the data of 16 foreign companies quoted on NYSE by ANNs, and they are invested for 30 trading days. According to the results, portfolio derived by ANNs gained 10.30% return, while traditional portfolios gained 5.98% return.


Sign in / Sign up

Export Citation Format

Share Document