2 Cost of Composites:a Qualitative Discussion Considering that cost is the most important aspect of an airframe structure (along with the weight),one would expect it to be among the best defined,most studied and most optimized quantities in a design.Unfortunately,it remains one of the least understood and ill-defined aspects of a structure.There are many reasons for this inconsistency some of which are:(a)cost data for different fabrication processes and types of parts are proprietary and only indirect or comparative values are usually released;(b)there seems to be no well-defined reliable method to relate design properties such as geometry and complexity to the cost of the resulting structure;(c)different companies have different methods of book-keeping the cost,and it is hard to make comparisons without knowing these differences (for example,the cost of the autoclave can be apportioned to the number of parts being cured at any given time or it may be accounted for as an overhead cost,included in the total overhead cost structure of the entire factory);(d)learning curve effects,which may or may not be included in the cost figures reported,tend to confuse the situation especially since different companies use different production run sizes in their calculations. These issues are common to all types of manufacturing technologies and not just the aerospace sector.In the case of composites the situation is further complicated by the relative novelty of the materials and processes being used,the constant emergence of new processes or variations thereof that alter the cost structure,and the high nonrecurring cost associated with switching to the new processes that,usually,acts as a deterrent towards making the switch. The discussion in this chapter attempts to bring up some of the cost considerations that may affect a design.This discussion is by no means exhaustive,in fact it is limited by the lack of extensive data and generic but accurate cost models.It serves mainly to alert or sensitize a designer to several issues that affect the cost.These issues,when appropriately accounted for,may lead to a robust design that minimizes the weight and is cost-competitive with the alternatives. Design and Analysis of Composite Structures:With Applications to Aerospace Structures Christos Kassapoglou 2010 John Wiley Sons,Ltd
2 Cost of Composites: a Qualitative Discussion Considering that cost is the most important aspect of an airframe structure (along with the weight), one would expect it to be among the best defined, most studied and most optimized quantities in a design. Unfortunately, it remains one of the least understood and ill-defined aspects of a structure. There are many reasons for this inconsistency some of which are: (a) cost data for different fabrication processes and types of parts are proprietary and only indirect or comparative values are usually released; (b) there seems to be no well-defined reliable method to relate design properties such as geometry and complexity to the cost of the resulting structure; (c) different companies have different methods of book-keeping the cost, and it is hard to make comparisons without knowing these differences (for example, the cost of the autoclave can be apportioned to the number of parts being cured at any given time or it may be accounted for as an overhead cost, included in the total overhead cost structure of the entire factory); (d) learning curve effects, which may or may not be included in the cost figures reported, tend to confuse the situation especially since different companies use different production run sizes in their calculations. These issues are common to all types of manufacturing technologies and not just the aerospace sector. In the case of composites the situation is further complicated by the relative novelty of the materials and processes being used, the constant emergence of new processes or variations thereof that alter the cost structure, and the high nonrecurring cost associated with switching to the new processes that, usually, acts as a deterrent towards making the switch. The discussion in this chapter attempts to bring up some of the cost considerations that may affect a design. This discussion is by no means exhaustive, in fact it is limited by the lack of extensive data and generic but accurate cost models. It serves mainly to alert or sensitize a designer to several issues that affect the cost. These issues, when appropriately accounted for, may lead to a robust design that minimizes the weight and is cost-competitive with the alternatives. Design and Analysis of Composite Structures: With Applications to Aerospace Structures Christos Kassapoglou 2010 John Wiley & Sons, Ltd
10 Design and Analysis of Composite Structures The emphasis is placed on recurring and nonrecurring cost.The recurring cost is the cost that is incurred every time a part is fabricated.The nonrecurring cost is the cost that is incurred once during the fabrication run. 2.1 Recurring Cost The recurring cost includes the raw material cost (including scrap)for fabricating a specific part,the labor hours spent in fabricating the part,and cost of attaching it to the rest of the structure.The recurring cost is hard to quantify,especially for complex parts.There is no single analytical model that relates specific final part attributes such as geometry,weight, volume,area,or complexity to the cost of each process step and through the summation over all process steps to the total recurring cost.One of the reasons for these difficulties and,as a result,the multitude of cost models that have been proposed with varying degrees of accuracy and none of them all-encompassing,is the definition of complexity.One of the most rigorous and promising attempts to define complexity and its effect on recurring cost of composite parts was by Gutowski et al.[1,2]. For the case of hand layup,averaging over a large quantity of parts of varying complexity ranging from simple flat laminates to compound curvature parts with co-cured stiffeners,the fraction of total cost taken up by the different process steps is shown in Figure 2.1 (taken from [3]). It can be seen from Figure 2.1 that,by far,the costliest steps are locating the plies into the mold(42%)and assembling to the adjacent structure(29%).Over the years,cost-cutting and Prepare mold 0.40% Cut material Assemble to 5.10% adjacent structure- 29.10% Trim Collect locate 0.20% into mold Remove from 41.80% mold clean (deflash) 1.50% Remove bag (debag) 6.10% Cure Debulk 7.30% 2.40% Apply vacuum bag 6.10% Figure 2.1 Process steps for hand layup and theircost as fractions of total recurring cost [3](See Plate 10 for the colour figure)
The emphasis is placed on recurring and nonrecurring cost. The recurring cost is the cost that is incurred every time a part is fabricated. The nonrecurring cost is the cost that is incurred once during the fabrication run. 2.1 Recurring Cost The recurring cost includes the raw material cost (including scrap) for fabricating a specific part, the labor hours spent in fabricating the part, and cost of attaching it to the rest of the structure. The recurring cost is hard to quantify, especially for complex parts. There is no single analytical model that relates specific final part attributes such as geometry, weight, volume, area, or complexity to the cost of each process step and through the summation over all process steps to the total recurring cost. One of the reasons for these difficulties and, as a result, the multitude of cost models that have been proposed with varying degrees of accuracy and none of them all-encompassing, is the definition of complexity. One of the most rigorous and promising attempts to define complexity and its effect on recurring cost of composite parts was by Gutowski et al. [1, 2]. For the case of hand layup, averaging over a large quantity of parts of varying complexity ranging from simple flat laminates to compound curvature parts with co-cured stiffeners, the fraction of total cost taken up by the different process steps is shown in Figure 2.1 (taken from [3]). It can be seen from Figure 2.1 that, by far, the costliest steps are locating the plies into the mold (42%) and assembling to the adjacent structure (29%). Over the years, cost-cutting and Debulk 2.40% Apply vacuum bag 6.10% Cure 7.30% Remove bag (debag) 6.10% Trim 0.20% Remove from mold & clean (deflash) 1.50% Collect & locate into mold 41.80% Prepare mold 0.40% Cut material Assemble to 5.10% adjacent structure 29.10% Figure 2.1 Process steps for hand layup and their cost as fractions of total recurring cost [3] (See Plate 10 for the colour figure) 10 Design and Analysis of Composite Structures
Cost of Composites:a Qualitative Discussion 11 optimization efforts have concentrated mostly on these two process steps.This is the reason for introducing automation.Robots,used for example in automated tape layup,take the cut plies and locate them automatically in the mold,greatly reducing the cost associated with that process step,improving the accuracy,and reducing or eliminating human error,thereby increasing consistency and quality.Since assembly accounts for about one-third of the total cost,increasing the amount of co-curing where various components are cured at the same time,reduces drastically the assembly cost.An example of this integration is shown in Figure 2.2. These improvements as well as others associated with other process steps such as automated cutting (using lasers or water jets),trimming and drilling (using numerically controlled equipment)have further reduced the cost and improved quality by reducing the human involvement in the process.Hand layup and its automated or semi-automated variations can be used to fabricate just about any piece of airframe structure.An example of a complex part with compound curvature and parts intersecting in different directions is shown in Figure 2.3. Further improvements have been brought to bear by taking advantage of the experience acquired in the textile industry.By working with fibers alone,several automated techniques such as knitting,weaving,braiding and stitching can be used to create a preform,which is then injected with resin.This is the resin transfer molding(RTM)process.The raw material cost can be less than half the raw material cost of pre-impregnated material(prepreg)used in hand layup or automated tape layup because the impregnation step needed to create the prepreg used in those processes is eliminated.On the other hand,ensuring that resin fully wets all fibers everywhere in the part and that the resin content is uniform and equal to the desired resin Figure 2.2 Integration of various parts into a single co-cured part to minimize assembly cost(Courtesy Aurora Flight Sciences)
optimization efforts have concentrated mostly on these two process steps. This is the reason for introducing automation. Robots, used for example in automated tape layup, take the cut plies and locate them automatically in the mold, greatly reducing the cost associated with that process step, improving the accuracy, and reducing or eliminating human error, thereby increasing consistency and quality. Since assembly accounts for about one-third of the total cost, increasing the amount of co-curing where various components are cured at the same time, reduces drastically the assembly cost. An example of this integration is shown in Figure 2.2. These improvements as well as others associated with other process steps such as automated cutting (using lasers or water jets), trimming and drilling (using numerically controlled equipment) have further reduced the cost and improved quality by reducing the human involvement in the process. Hand layup and its automated or semi-automated variations can be used to fabricate just about any piece of airframe structure. An example of a complex part with compound curvature and parts intersecting in different directions is shown in Figure 2.3. Further improvements have been brought to bear by taking advantage of the experience acquired in the textile industry. By working with fibers alone, several automated techniques such as knitting, weaving, braiding and stitching can be used to create a preform, which is then injected with resin. This is the resin transfer molding(RTM) process. The raw material cost can be less than half the raw material cost of pre-impregnated material (prepreg) used in hand layup or automated tape layup because the impregnation step needed to create the prepreg used in those processes is eliminated. On the other hand, ensuring that resin fully wets all fibers everywhere in the part and that the resin content is uniform and equal to the desired resin Figure 2.2 Integration of various parts into a single co-cured part to minimize assembly cost (Courtesy Aurora Flight Sciences) Cost of Composites: a Qualitative Discussion 11
12 Design and Analysis of Composite Structures Figure 2.3 Portion of a three-dimensional composite part with compound curvature fabricated using hand layup content can be hard for complex parts,and may require special tooling,complex design of injection and overflow ports,and use of high pressure.It is not uncommon,for complex RTM parts to have 10-15%less strength(especially in compression and shear)than their equivalent prepreg parts due to reduced resin content.Another problem with matched metal molding RTM is the high nonrecurring cost associated with the fabrication of the molds.For this reason, variations of the RTM process such as vacuum-assisted RTM(VARTM)where one of the tools is replaced by a flexible caul plate whose cost is much lower than an equivalent matched metal mold,or resin film infusion(RFI)where resin is drawn into dry fiber preforms from a pool or film located under it and/or from staged plies that already have resin in them,have been used successfully in several applications (Figure 2.4).Finally,due to the fact that the process operates with resin and fibers separately,the high amounts of scrap associated with hand layup can be significantly reduced. Introduction of more automation led to the development of automated fiber or tow placement.This was a result of trying to improve filament winding (see below).Robotic heads can each dispense material as narrow as 3 mm and as wide as 100 mm by manipulating individual strips(or tows)each 3 mm wide.Tows are individually controlled so the amount of material laid down in the mold can vary in real time.Starting and stopping individual tows also allows the creation of cutouts 'on the fly'.The robotic head can move in a straight line at very high rates(as high as 30 m/min).This makes automated fiber placement an ideal process for laying material down to create parts with large surface area and small variations in thickness or cutouts.For maximum efficiency,structural details (e.g.cutouts)that require starting and stopping the machine or cutting material while laying it down should be avoided.Material scrap is very low.Convex as well as concave tools can be used since the machine does not rely on constant fiber tension,as in filament winding,to lay material down.There are limitations
content can be hard for complex parts, and may require special tooling, complex design of injection and overflow ports, and use of high pressure. It is not uncommon, for complex RTM parts to have 10–15% less strength (especially in compression and shear) than their equivalent prepreg parts due to reduced resin content. Another problem with matched metal molding RTM is the high nonrecurring cost associated with the fabrication of the molds. For this reason, variations of the RTM process such as vacuum-assisted RTM (VARTM) where one of the tools is replaced by a flexible caul plate whose cost is much lower than an equivalent matched metal mold, or resin film infusion (RFI) where resin is drawn into dry fiber preforms from a pool or film located under it and/or from staged plies that already have resin in them, have been used successfully in several applications (Figure 2.4). Finally, due to the fact that the process operates with resin and fibers separately, the high amounts of scrap associated with hand layup can be significantly reduced. Introduction of more automation led to the development of automated fiber or tow placement. This was a result of trying to improve filament winding (see below). Robotic heads can each dispense material as narrow as 3 mm and as wide as 100 mm by manipulating individual strips (or tows) each 3 mm wide. Tows are individually controlled so the amount of material laid down in the mold can vary in real time. Starting and stopping individual tows also allows the creation of cutouts ‘on the fly’. The robotic head can move in a straight line at very high rates (as high as 30 m/min). This makes automated fiber placement an ideal process for laying material down to create parts with large surface area and small variations in thickness or cutouts. For maximum efficiency, structural details (e.g. cutouts) that require starting and stopping the machine or cutting material while laying it down should be avoided. Material scrap is very low. Convex as well as concave tools can be used since the machine does not rely on constant fiber tension, as in filament winding, to lay material down. There are limitations Figure 2.3 Portion of a three-dimensional composite part with compound curvature fabricated using hand layup 12 Design and Analysis of Composite Structures
Cost of Composites:a Qualitative Discussion 13 Figure 2.4 Curved stiffened panels made with the RTM process with the process associated with the accuracy of starting and stopping when material is laid down at high rates and the size and shape of the tool when concave tools are used (in order to avoid interference of the robotic head with the tool).The ability to steer fibers on prescribed paths(Figure 2.5)can also be used as an advantage by transferring the loads efficiently across Figure 2.5 Composite cylinder with steered fibers fabricated by automated fiber placement(made in a collaborative effort by TUDelft and NLR;see Plate 11 for the colour figure)
with the process associated with the accuracy of starting and stopping when material is laid down at high rates and the size and shape of the tool when concave tools are used (in order to avoid interference of the robotic head with the tool). The ability to steer fibers on prescribed paths (Figure 2.5) can also be used as an advantage by transferring the loads efficiently across Figure 2.4 Curved stiffened panels made with the RTM process Figure 2.5 Composite cylinder with steered fibers fabricated by automated fiber placement (made in a collaborative effort by TUDelft and NLR; see Plate 11 for the colour figure) Cost of Composites: a Qualitative Discussion 13