Göran Fernlund

Experimental and numerical study of the effect of cure cycle, tool surface, geometry, and lay-up on the dimensional fidelity of autoclave-processed composite parts

Abstract Resin cure shrinkage and anisotropic thermal expansion cause process induced residual stresses in polymer composites. When relieved, the residual stresses cause reduction in enclosed angles of composite laminates; a phenomenon often called spring-in. Spring-in compromises the dimensional fidelity of composite parts and is often accounted for when designing the tool the part is made on. Spring-in is often estimated using past experience or simple analytical formulas that ignores many process parameters affecting the spring-in. This paper presents an experimental study that shows that spring-in can be strongly affected by a number of factors such as cure cycle, tool surface, part geometry, and lay-up. The paper also shows that by developing material models that accurately represent the stress transfer between the part and the tool at the tool-part interface, and by implementing a large deformation solution technique, the experimental results observed in this study can be predicted using finite element based process models.

Tool–part interaction in composites processing. Part I: experimental investigation and analytical model

Abstract The ability to process composite structures with a high degree of dimensional control remains a barrier to further implementation of composite materials in commercial applications. Of the numerous types of process induced deformations that occur, the warpage of flat laminates due to tool–part interaction remains a poorly understood phenomenon. This paper presents an experimental study of the effect of process conditions and part aspect ratio on tool–part interaction induced warpage. For a given lay-up and material, part aspect ratio was found to have a much greater influence than autoclave pressure on warpage, while the tool surface condition was not observed to have any significant effect. The results of the study are embodied in an empirical relation, which can be a useful guide to predict laminate warpage over a range of industrially relevant conditions. In addition, a simple analytical model is proposed which agrees well with the experimentally determined relationships. A complementary numerical model is presented in a companion paper.

Finite element based prediction of process-induced deformation of autoclaved composite structures using 2D process analysis and 3D structural analysis

Dimensional control of composite components is critical for cost effective manufacturing of large composite aerospace structures. This paper presents an engineering approach to the prediction of process-induced deformations of three-dimensional (3D) autoclaved composite components. A 6-step method that uses a two-dimensional (2D) special purpose finite element (FE) based process simulation code and a standard 3D structural FE code is presented. The approach avoids the need to develop a full 3D process model, significantly reducing the computational effort yet retaining much of the detail required for accurate analysis. The methodology is presented together with numerical examples and two case studies demonstrating the validity, utility, and limitations of the approach.

An experimental method for quantifying tool–part shear interaction during composites processing

Abstract The success with which dimensional control during processing of composite structures can be modelled depends on the level of understanding of the underlying mechanisms that drive the accumulation of residual stresses in the part. Tool–part shear interaction during processing can cause substantial warpage in initially flat laminates, yet this phenomenon remains poorly understood. This paper presents an experimental technique in which a thin tool, instrumented with strain gauges, is used for characterizing the interfacial shear stresses that arise between the tool and part during processing. The results show that a sliding interface condition occurs during the majority of the cure cycle, although, at times the tool and part adhere together resulting in high interfacial shear stresses. This tool–part interaction occurs despite the use of a release agent, though the use of a fluoroethylenepropylene (FEP) release film at the tool–part interface reduces the effect.

Tool -part interaction in composites processing. Part II: numerical modelling

Abstract When modelling deformations induced during the manufacture of composite components it is important to account for the effect which tool–part interaction has on final part shape. In the current work an existing numerical code was used to simulate the effect of tool–part interaction. A parametric study examining the effects of various model inputs on final part shape was performed and the results were compared to those of a companion experimental study. The results show that the both the tool–part interfacial shear stress distribution and the part in-plane stress distribution are critical for accurate modelling of this phenomenon.

Thermal models for MTM45-1 and Cycom 5320 out-of-autoclave prepreg resins

Out-of-autoclave prepregs require a two-step cure cycle. The first step is a low temperature cure to consolidate the laminate and build sufficient green strength to proceed to the second step, a free-standing post-cure at traditional autoclave temperatures to fully cross-link the resin. Process modeling can help design a robust cure cycle to avoid scrapping large parts in production. The focus of this article is to develop the cure kinetics, viscosity, and glass transition temperature models for two commercially available out-of-autoclave epoxy resins. Since the cure kinetics model is the basis for all other thermal models, the cure kinetics model was validated using a one-dimensional heat transfer analysis on thick prepreg laminates. Finally, the out-of-autoclave resin models were compared to a traditional autoclave resin system to highlight the difference in resin reactivity for out-of-autoclave processing.

Experimental and numerical study of the effect of caul-sheets on corner thinning of composite laminates

Caul-sheets are used to locally reduce or intensify the pressure exerted on composite laminates during autoclave processing. They are often introduced into the process in the last minute to address manufacturing problems as they occur in production. There are multitudes of caul geometries and caul materials used in industry, often quite different, but intended to give the same end result, a well consolidated laminate, free of voids and with uniform thickness. This paper presents the results of an experimental and modelling study where the thickness of composite laminates made with five different caul-sheet configurations was measured and modelled. The experimental study showed that the type of caul-sheet has a significant effect on the resulting thickness profile, and particularly corner thinning. Modelling of corner thinning was done using a special purpose finite element code. Despite the approximations made in the finite element model, the model predictions were qualitatively in agreement with the experimental results, with the model being able to rank the caul-sheets in terms of effectiveness. A simple mechanistic model illustrating what is believed to be the main mechanisms behind thickness variations is also presented.

Probability-Based Modelling of Composites Manufacturing and Its Application to Optimal Process Design

The control of process-induced deformations in composite structures is important for cost-effective manufacturing. In recent years, significant advances have been made in predicting the average deformation behaviour, but little work has been done in predicting the variability, which results from uncertainties in both the raw material properties and the manufacturing process conditions. A probability-based approach is presented in this paper for predicting the variability of process-induced deformations. A two-dimensional finite element code, which deterministically simulates the various physical phenomena during processing of composite structures, is integrated with a first-order reliability analysis method to calculate the probability of the deformations exceeding a specified allowable tolerance. The methodology is demonstrated through two case studies. In the first study, a probabilistic description of the process-induced spring-in of a channel section is achieved and the effect of variability in material properties on the final channel angle is studied. In the second study, the optimal tool-shape for the channel section is determined by coupling reliability analysis with a simple cost model of the manufacturing process.

Effect of bone graft substitute on marrow stromal cell proliferation and differentiation

Marrow stromal cells (MSCs) are ideally suited for tissue engineered bone grafts since they have the potential to regenerate bone, but may also maintain the homeostasis of the repaired tissue through their ability for self-renewal. An ideal bone graft substitute should support MSC self-renewal as well as differentiation to ensure complete bone defect regeneration and maintenance. The purpose of this investigation was to determine the effect of different substrate materials on MSC expansion and differentiation. Calcium polyphosphate (CPP), bone and hydroxyapatite/tricalcium phosphate (HA/TCP) were seeded with rat MSCs and maintained in culture conditions that promote cell expansion. At 0, 3, 7, 14, and 21 days cell numbers were determined by measuring their metabolic activity using a MTT assay and the frequency of cycling cells by 24 hr BrdU incorporation. Osteogenic, chondrogenic, and adipogenic marker expression in these cultures was measured by qRT-PCR. An initial drop in cell numbers was observed on all substrates. CPP and bone, but not HA/TCP supported an increase in proliferating cells at day 14 and 21. In addition, no upregulation of mature bone markers was observed in cells cultured on CPP and bone, which suggests that these substrates support the expansion of undifferentiated MSCs. In contrast, cell numbers on HA/TCP decreased with time and only rare BrdU positive cells were observed. This decrease in proliferation correlated with the down regulation of osteogenic progenitor markers and the substantial increase in mature osteocyte markers, indicating that HA/TCP favors MSC differentiation and maturation along the osteogenic lineage.

In vivo evaluation of calcium polyphosphate for bone regeneration

Current problems associated with bone allografts include risk of disease transmission, limited availability, and cost. Synthetic scaffolds have been proposed as substitute graft materials to address these issues. Calcium polyphosphate is a novel synthetic scaffold material that has shown good mechanical properties and biocompatibility. Here, we evaluated calcium polyphosphate in terms of its ability to support cell proliferation and differentiation in vivo. Calcium polyphosphate, morsellized cancellous bone, and hydroxyapatite/tricalcium phosphate particles were seeded with marrow stromal cells and implanted subcutaneously in the back of NOD/Scid mice. At 7, 14, and 28 days the samples were harvested and the proliferation characteristics and gene expression were analyzed. All tested graft materials had similar proliferation characteristics and gene expression. The subcutaneous environment had a stronger impact on the proliferation and differentiation of the cells than the scaffold material itself. However, it was shown that calcium polyphosphate is superior to hydroxyapatite/tricalcium phosphate and bone in its ability to support cell survival in vivo. The study confirmed that calcium polyphosphate has potential for replacing morsellized cancellous bone as a graft material for bone regeneration.