Estimativa de vida em fadiga de estruturas submetidas a excitações randômicas e análise comparativa com resultados de ensaios de durabilidade
Carregando...
Arquivos
Data
2022-08-01
Autores
Lugli, Vinicius
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
Considerando as dificuldades da etapa de dimensionamento por fadiga de estruturas submetidas à carregamentos dinâmicos randômicos, a escolha da metodologia para estimativa de vida em fadiga destas estruturas é muito relevante para a complexidade, custo e eficiência deste processo. Portanto, a proposta principal desta dissertação é avaliar dois métodos de estimativa de fadiga distintos com auxílio de análises pelo método de elementos finitos (MEF). A primeira metodologia utiliza uma análise dinâmica transiente (TDA) pelo MEF e o algoritmo de contagem de ciclos de amplitude de tensão, pelo método de Rainflow, para estimativa do ciclo de fadiga. O segundo processo utiliza o método espectral proposto por Dirlik, onde a estimativa do ciclo de fadiga é dada por uma função densidade de probabilidade (PDF) de amplitude de tensões. Esta PDF é calculada combinando a função de resposta em frequência (FRF) da estrutura, resultado de uma análise harmônica de resposta em frequência pelo MEF, com a densidade espectral de potência (PSD) do sinal randômico de fadiga. Para complementar o trabalho, um ensaio de durabilidade foi idealizado para ensaiar vinte e quatro corpos de prova divididos em três ciclos. O resultado destes ensaios possibilita a obtenção de informações estatísticas sobre a resistência à fadiga do material, e auxiliam no processo de calibração dos parâmetros de entrada da análise pelo MEF. Com os cálculos computacionais e os ensaios de durabilidade em laboratório, chega-se à conclusão de que a estimativa de fadiga pelo método proposto por Dirlik foi 12,6% mais severa que o método de Rainflow, e 7,8% menos severa do que o resultado médio de falha da estrutura durante ensaio de durabilidade. Os resultados obtidos pelos métodos propostos saíram conforme esperado, ambos com relativa precisão, e com isso o método de Dirlik se confirmou viável para estimar o ciclo de fadiga de estruturas submetidas a processos randômicos. Além de sua eficiência na estimativa do ciclo de fadiga, o método espectral demanda um menor custo computacional para solução da análise FEA de sistemas de alta complexidade. Visando confirmar os resultados obtidos, a estrutura do paralama de uma máquina agrícola automotriz foi modelada pelo MEF e teve sua fadiga estimada pelo método de Dirlik. Esta estrutura foi submetida a um ensaio de durabilidade em pista de provas e com as medições realizadas durante este ensaio foram determinados os parâmetros de entrada da análise pelo MEF e da estimativa de fadiga pelo método espectral. Apesar da alta complexidade da estrutura analisada, o desvio relativo entre a estimativa de vida de fadiga (Dirlik) e o tempo médio de falha dos quatro paralamas ensaiados em pista de provas, foi de +10,4%. Este desvio obtido está dentro da expectativa, principalmente considerando a determinação teórica da tensão de resistência à fadiga do material.
Considering the difficulties of the fatigue design process of structures subjected to dynamic random loads, the choice of methodology to estimate the fatigue life of these structures is very relevant to the complexity, cost, and effic iency of the process. Therefore, the main goal of this research is to evaluate two different fatigue life prediction methods with the aid of computational analysis by the finite element method (FEM). The first chosen methodology is one of the most widespre ad in the structural design environment and uses the transient dynamic analysis (TDA) by the FEM and the stress amplitude cycle counting by the Rainflow algorithm method to define the fatigue cycle. The second process uses the spectral method, proposed by Dirlik, where a probability density function (PDF) of stress amplitude estimates the fatigue cycle. The PDF is calculated by combining the structure's frequency response function (FRF), from harmonic response analysis by the FEM, with the power spectral de nsity (PSD) of fatigue time history. To complement the proposed research a durability test was designed, and twentyfour specimens were tested divided into three durability cycles. With the durability test results of the specimens, it is possible statistic ally determine the structure's failure time. The purpose of this test is to calibrate the FEA input parameters, in addition to obtaining statistical information about the fatigue strength of the material. Finished the computational calculations and the ben ch tests, was concluded that the fatigue life estimate result by the Dirlik's method was 12.6% more severe than the Rainflow method, and 7.8% less severe than the average durability test result. The results obtained by the proposed methods came out as expe cted, both with relative precision, and with that the Dirlik method proved to be viable for estimating the fatigue cycle of structures subjected to random processes . In addition to its efficiency in estimating the fatigue cycle, it requires a lower computa tional cost to solve the harmonic response analysis by the FEM of high complexity systems. The fender structure of a selfpropelled agricultural machine was modeled by FEM, and to confirm the obtained results, its fatigue life was estimated by Dirlik's met hod. This structure was subject to a racetrack durability test, and measurements were performed during this test. With the obtained information were defined the input parameters of the finite element analysis (FEA) and the fatigue estimation by the Dirlik method. Despite the high complexity of the FEA model, the relative deviation between the estimated fatigue life using the Dirlik method and the average fatigue failure time of four fenders tested on the racetrack was 10.4%. This deviation is within the exp ectation, mainly considering the theoretical obtainment of the fatigue resistance stress of the material.
Considering the difficulties of the fatigue design process of structures subjected to dynamic random loads, the choice of methodology to estimate the fatigue life of these structures is very relevant to the complexity, cost, and effic iency of the process. Therefore, the main goal of this research is to evaluate two different fatigue life prediction methods with the aid of computational analysis by the finite element method (FEM). The first chosen methodology is one of the most widespre ad in the structural design environment and uses the transient dynamic analysis (TDA) by the FEM and the stress amplitude cycle counting by the Rainflow algorithm method to define the fatigue cycle. The second process uses the spectral method, proposed by Dirlik, where a probability density function (PDF) of stress amplitude estimates the fatigue cycle. The PDF is calculated by combining the structure's frequency response function (FRF), from harmonic response analysis by the FEM, with the power spectral de nsity (PSD) of fatigue time history. To complement the proposed research a durability test was designed, and twentyfour specimens were tested divided into three durability cycles. With the durability test results of the specimens, it is possible statistic ally determine the structure's failure time. The purpose of this test is to calibrate the FEA input parameters, in addition to obtaining statistical information about the fatigue strength of the material. Finished the computational calculations and the ben ch tests, was concluded that the fatigue life estimate result by the Dirlik's method was 12.6% more severe than the Rainflow method, and 7.8% less severe than the average durability test result. The results obtained by the proposed methods came out as expe cted, both with relative precision, and with that the Dirlik method proved to be viable for estimating the fatigue cycle of structures subjected to random processes . In addition to its efficiency in estimating the fatigue cycle, it requires a lower computa tional cost to solve the harmonic response analysis by the FEM of high complexity systems. The fender structure of a selfpropelled agricultural machine was modeled by FEM, and to confirm the obtained results, its fatigue life was estimated by Dirlik's met hod. This structure was subject to a racetrack durability test, and measurements were performed during this test. With the obtained information were defined the input parameters of the finite element analysis (FEA) and the fatigue estimation by the Dirlik method. Despite the high complexity of the FEA model, the relative deviation between the estimated fatigue life using the Dirlik method and the average fatigue failure time of four fenders tested on the racetrack was 10.4%. This deviation is within the exp ectation, mainly considering the theoretical obtainment of the fatigue resistance stress of the material.
Descrição
Palavras-chave
Análise por elementos finitos, Fadiga, Dirlik, Rainflow, Finite element Analysis, Fatigue