TY - JOUR

T1 - Characterizing the mechanical contribution of fiber angular distribution in connective tissue

T2 - Comparison of two modeling approaches

AU - Cortes, Daniel H.

AU - Lake, Spencer P.

AU - Kadlowec, Jennifer A.

AU - Soslowsky, Louis J.

AU - Elliott, Dawn M.

N1 - Funding Information:
Acknowledgments This study was supported by a grant from the NIH/NIAMS (AR055598) and the NIH/NIAMS-supported Penn Center for Musculoskeletal Disorders (AR050950).

PY - 2010/10

Y1 - 2010/10

N2 - Modeling of connective tissues often includes collagen fibers explicitly as one of the components. These fibers can be oriented in many directions; therefore, several studies have considered statistical distributions to describe the fiber arrangement. One approach to formulate a constitutive framework for distributed fibers is to express the mechanical parameters, such as strain energy and stresses, in terms of angular integrals. These integrals represent the addition of the contribution of infinitesimal fractions of fibers oriented in a given direction. This approach leads to accurate results; however, it requires lengthy calculations. Recently, the use of generalized structure tensors has been proposed to represent the angular distribution in the constitutive equations of the fibers. Although this formulation is much simpler and fewer calculations are required, such structure tensors can only be used when all the fibers are in tension and the angular distribution is small. However, the amount of error introduced in these cases of non-tensile fiber loading and large angular distributions have not been quantified. Therefore, the objective of this study is to determine the range of values of angular distribution for which acceptable differences (less than 10%) between these two formulations are obtained. It was found, analytically and numerically, that both formulations are equivalent for planar distributions under equal-biaxial stretch. The comparison also showed, for other loading conditions, that the differences decrease when the fiber distribution is very small. Differences of less than 10% were usually obtained when the fiber distribution was very low (κ ≈ 0.03; κ ranges between 0 and 1/3, for aligned and isotropic distributed fibers, respectively). This range of angular distribution greatly limits the types of tissue that can be accurately analyzed using generalized structure tensors. It is expected that the results from this study guide the selection of a proper approach to analyze a particular tissue under a particular loading condition.

AB - Modeling of connective tissues often includes collagen fibers explicitly as one of the components. These fibers can be oriented in many directions; therefore, several studies have considered statistical distributions to describe the fiber arrangement. One approach to formulate a constitutive framework for distributed fibers is to express the mechanical parameters, such as strain energy and stresses, in terms of angular integrals. These integrals represent the addition of the contribution of infinitesimal fractions of fibers oriented in a given direction. This approach leads to accurate results; however, it requires lengthy calculations. Recently, the use of generalized structure tensors has been proposed to represent the angular distribution in the constitutive equations of the fibers. Although this formulation is much simpler and fewer calculations are required, such structure tensors can only be used when all the fibers are in tension and the angular distribution is small. However, the amount of error introduced in these cases of non-tensile fiber loading and large angular distributions have not been quantified. Therefore, the objective of this study is to determine the range of values of angular distribution for which acceptable differences (less than 10%) between these two formulations are obtained. It was found, analytically and numerically, that both formulations are equivalent for planar distributions under equal-biaxial stretch. The comparison also showed, for other loading conditions, that the differences decrease when the fiber distribution is very small. Differences of less than 10% were usually obtained when the fiber distribution was very low (κ ≈ 0.03; κ ranges between 0 and 1/3, for aligned and isotropic distributed fibers, respectively). This range of angular distribution greatly limits the types of tissue that can be accurately analyzed using generalized structure tensors. It is expected that the results from this study guide the selection of a proper approach to analyze a particular tissue under a particular loading condition.

UR - http://www.scopus.com/inward/record.url?scp=77958093617&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77958093617&partnerID=8YFLogxK

U2 - 10.1007/s10237-010-0194-x

DO - 10.1007/s10237-010-0194-x

M3 - Article

C2 - 20148345

AN - SCOPUS:77958093617

VL - 9

SP - 651

EP - 658

JO - Biomechanics and Modeling in Mechanobiology

JF - Biomechanics and Modeling in Mechanobiology

SN - 1617-7959

IS - 5

ER -