Wood resource management using an endocrine NARX neural network

Milovanović, Miroslav B.; Antić, Dragan; Rajić, Milena; Milosavljevic, Pedja; Pavlović, Ana; Fragassa, Cristiano

doi:10.1007/s00107-017-1223-6

Cited by 13 publications

(13 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…An efficient wood harvesting action aims at fully recovering wood as a measure to maximize the profit [3], also in terms of new methods for machining plans optimisation [4] or technical solutions for improving the equipment efficiency [5][6]. But safety prescriptions should be obeyed each time when harvesting operations are performed [7].…”

Section: Introductionmentioning

confidence: 99%

Investigating the influence of design factors on noise and vibrations in the case of chainsaws for forestry work

Kováč¹,

Krilek²,

Dado³

et al. 2018

FME Transactions

View full text Add to dashboard Cite

In the recent time forest harvesting is mainly related to the use of chainsaws. They are acceptably light, safe, cheap and multifunctional. At the same time, chainsaws can cause personal illnesses considering their vibrations and noise. This article deals with conditions and assessment evaluation of effects on workers caused by vibrations and noise in the case of professional chainsaws in operation. Mechanical vibrations and noise are mainly related to the back-and-forth movement of the piston, and to a lesser extent, to the cutting chain. The experiment revealed, in general, that design parameters have a minimal influence on chainsaw vibration and noise emissions. Specifically, the statistical analysis for frequency weighted vibration showed no difference between types of guide bars, types of saw chains and wood plant in terms of noise and vibration on handles.

show abstract

Section: Introductionmentioning

confidence: 99%

Investigating the influence of design factors on noise and vibrations in the case of chainsaws for forestry work

Kováč¹,

Krilek²,

Dado³

et al. 2018

FME Transactions

View full text Add to dashboard Cite

show abstract

“…The task is to predict two consumption variables by using an Endocrine NARX neural network (ENARX). This kind of network was already used for prediction purposes in Milovanovi c et al, 41 where the task was to predict the monthly requirements of treated wood in the country. The ENARX network is based on the artificial endocrine factor that provides an extra sensitivity of the network to external disturbances and environmental conditions.…”

Section: Endocrine Neural Network Modelmentioning

confidence: 99%

“…On the contrary, in most cases, a NARX network reduces the computation time required for processing data. The structure of the NARX model is presented in Milovanovi c et al 41 and the characteristics of the default NARX network can be found in Siegelmann et al 43…”

Section: Endocrine Neural Network Modelmentioning

confidence: 99%

Analyzing energy poverty using intelligent approach

Rajić

Milovanović

Antić

et al. 2020

Energy & Environment

View full text Add to dashboard Cite

The study represents a new approach of analyzing energy poverty using an intelligent method based on social factors. Energy poverty represents one of the most vital problems, especially in recent years, with serious implications in the social, economic, environmental, political, and health sphere. Identifying and measuring social energy parameters represents a challenge not only on the national, but also on the regional level. This research represents a new approach to neural network application in energy systems and energy resource planning activities with the aim of analyzing energy poverty with real socio-economic data. The analyzed data are associated with social factors within one country. The newly developed model includes a new data optimization framework for pre-processing and selecting the most important parameters from a raw dataset. In analyzing energy poverty, it is concluded that its weakness is based on the actual energy consumption of households (both electricity and heat consumption). The model presented in this paper uses 15 influential parameters in energy prediction.

show abstract

“…A identificação de sistemasé umaárea de estudo que lida com a obtenção de modelos matemáticos para sistemas de diversas naturezas como sistemas econômicos, processos químicos, sistemas mecânicos, elétricos entre outros (Ok et al, 2018;Chevtchenko et al, 2018;Zhao et al, 2018;Milovanović et al, 2017;Boda et al, 2016;Aguirre et al, 2011). Modelos matemáticos podem ser vistos como aproximações por meio de equações de um sistema real (Martins et al, 2013).…”

Section: Introductionunclassified

Rede Combinatória de Modelos Não Lineares - Parte II: Aplicação em Computação Aritmética

Andrade

Peixoto

Martins

2019

Anais Do 14º Simpósio Brasileiro De Automação Inteligente

View full text Add to dashboard Cite

This work is the second part of a series of two articles and applies a proposed method of parallel combination of models in the context of arithmetic computation. The combination is performed from known submodels obtained through an interval system identification procedure that uses interval arithmetic in the least squares estimator to find a set of models that best fits the data. The interpolated model was compared with a model obtained from a classic identification procedure and with the average model of the interval identification procedure obtaining a decrease of about 11 % in the RMSE index. The model was also compared to the submodels used to obtain it, reaching in this case a decrease of about 79 % in the RMSE and 22 % in the energy of the autocorrelation of the waste vector. Resumo: Este trabalhoé a segunda parte de uma série de dois artigos e aplica um método proposto de combinação paralela de modelos no contexto da computação aritmética. A combinaçãoé realizada a partir de submodelos conhecidos obtidos por meio de um procedimento de identificação intervalar de sistemas que usa aritmética intervalar no estimador de mínimos quadrados para encontrar um conjunto de modelos que melhor se ajusta aos dados. O modelo interpolado foi comparado com um modelo obtido a partir de um procedimento clássico de identificação e com o modelo médio do procedimento de identificação intervalar obtendo uma diminuição de cerca de 11% noíndice RMSE. O modelo também foi comparado com os submodelos utilizados em sua obtenção obtendo neste caso uma diminuição de cerca de 79% no RMSE e de 22% na energia da autocorrelação do vetor de resíduos.

show abstract

Wood resource management using an endocrine NARX neural network

Cited by 13 publications

References 28 publications

Investigating the influence of design factors on noise and vibrations in the case of chainsaws for forestry work

Investigating the influence of design factors on noise and vibrations in the case of chainsaws for forestry work

Analyzing energy poverty using intelligent approach

Rede Combinatória de Modelos Não Lineares - Parte II: Aplicação em Computação Aritmética

Contact Info

Product

Resources

About