GUEST BLOG: Closing the technology gap for accurate prediction of composite material behavior

by Roger Assaker, CEO, e-Xstream engineering, and Chief Material Strategist, MSC Software

Engineers prize fiber-reinforced plastics (FRPs) for their stiffness-to-mass ratio, which gives them the ability to reduce product weight. Too often, however, engineers give away the advantage of high strength and low mass by over-designing products to prevent failure.

FRPs’ newness is part of the over-designing problem. That will pass as engineers accrue experience working with them. In the meantime, however, they are needlessly sacrificing weight, cost and efficiency gains because they aren’t using the right technology tools.

Most of the non-linear finite element analysis (FEA) and simulation tools on the market today were created with metals in mind. They assume the isotropic properties of metals rather than the anisotropic properties of FRPs. 

Unlike the stiffness of metal parts, which is consistent throughout their geometries, the stiffness of parts designed in FRPs can vary widely in different areas of their geometries. That variation comes from the alignment of carbon fibers in FRPs’ epoxy resin matrices determined by production processes.

Engineers must be able to model production processes and fiber alignment to accurately predict how their FRP designs will respond to forces.

Predictive analysis is a fundamental element of successful engineering with FRPs. Anisotropic analysis is, in turn, a major component of predictive analysis. However, most FEA and simulation software is not geared toward analysis of anisotropic materials.

Conventional FEA tools represent FRP parts as 'black aluminum,' which is essentially a placeholder that depicts the part’s geometry but not its anisotropic properties. Traditionally, the technologies have been used to benchmark material performance. Engineers develop benchmark performance values through simulation and prototyping, then use the benchmark values to anticipate their designs’ behaviors in other designs.

This comparative approach is more relevant to metals than FRPs because metals behave more consistently from one design to another than FRPs. One benchmark for a material can carry across several designs.

Working with FRPs means dealing with much more variability within a given geometry. A single benchmark value cannot accurately represent an FRP’s stiffness from one geometry to the next. FRP stiffness varies with fiber placement. For example, a plastic part that deflects 5 mm under a 1,000 load might only deflect 0.75 mm if its upper edge is reinforced with carbon fibers. The engineer who assumes the 5 mm deflection because of a previous benchmark will use much more mass than necessary to prevent failure.

FEA and simulation tools must include fiber orientation maps and 'smart' material data models to support predictive analysis. Smart in this context means employing material datasets that, given any fiber alignment data, can tell engineers the exact stress-frame response. The fiber alignment data would come from process simulation software that depicts the 'cloud' of fibers in the epoxy resin matrix that emerges from the production process.

Working together, process simulation software and smart material modeling and FEA applications will enable engineers to predict FRP performance for each new design. This accuracy will give them the confidence to use FRPs to their best advantage and increase their rate of innovation. Rather than designing to avoid failure, engineers will be designing to create superior products. ♦ 

Roger Assaker is the co-­founder and CEO of e-­Xstream engineering, a software and engineering services company focused on advanced material modelling. He holds a PhD and MS in Aerospace Engineering with a strong focus on nonlinear computational mechanics where he totals more than 20 years of experience. Assaker is also Vice Chairman of NAFEMS' Composites Working Group and an active member of other technical associations such as SPE and SAMPE.