October 10, 2002
It's difficult to overemphasize the importance of cross section expansion when you're talking about successful and innovative hydroforming of steel tubing.
Overemphasizing one aspect of the tube hydroforming design process can take attention away from others and result in less than optimal design.
However, most discussion over tube hydroforming design flexibility focuses almost entirely on cross section expansion (variable-periphery design), which entails increasing the periphery of the starting tube at least along a portion of the part length.
Variable periphery is most commonly achieved in the hydroforming die during hydroexpansion.
Tube expansion adds cost, but is worth it where it adds sufficient value. Hydroexpansion can appear to be the most efficient and cheapest method, but this can only be determined after considering the alternatives. Other expansion methods that may provide design or cost benefits relative to hydroexpansion will be discussed in the next article.
There are many parts in current production that perform their function without expansion, thus avoiding several cost factors that are detailed below. Other production components have small, expanded areas that raise questions about the need and if it is the most effective and efficient design.
Where expansion is judged to be necessary, the main reasons are:
1. Structural rigidity.
2. To avoid pinching material as hydroforming die closes.
3. Fitting to mating parts.
Expansion of hydroformed parts has attracted a lot of attention and is closely associated with high forming pressure. Equation 1 shows that a mild steel tube with a 3-inch diameter (Yield Stress [YS]= 35 KSI) and wall thickness of 1.5 mm needs only 1,500 PSI to begin expanding.
It is widely believed that expansion in the die is the best and most economical way to do things. This may be true for some applications but is not always optimal. Expanding a tube by 10 percent or more is feasible only at or near the component's end, where extra material can be fed in. Where end feeding is not effective in-die expansion can cause wall thinning of 25 percent or more, as well as risk rupturing the tube.
Anything that increases friction between the die and the tube -- for example, bends; dents; low, wide rectangular shapes; or increasing distance from the end -- impedes additional material feeding because the feeding force cannot exceed the tube column strength. Large cylinders, perhaps 18 inches in diameter, are needed, and the intense abrasive forces generated may require hardened inserts. In addition, end feeding often increases cycle time, and local stretching can require special material.
You must limit the expansion amount to prevent tube rupture. Addressing this problem drives many of the ongoing development efforts in the hydroforming industry. For a particular material and part design, the areas of a tube with the least capacity for hydroexpansion are spots where formability has been depleted by prebending.
When material is stretched close to its forming limit (for example, expanded cross section corners), little elongation remains. Stress applied in service can cause cracks to start and propagate from even the smallest stress concentration, which is difficult to identify since it can't be seen.
Variations in the production process add to the uncertainty, even when parts are usually OK. Combined with the size and scope of the potential drawbacks, wall thinning must be considered carefully during design, making simulation an essential tool when using hydroexpansion.
Mild steel has been successfully expanded by 41 percent in a hydroexpansion die. Expansion of 27 percent, with 50 KSI high-strength, low-alloy (HSLA) steel and a bend of about 20 degrees, also has been done. Production feasibility of the large expanded sections of the prototype parts shown in Figure 1 would be more challenging.
Figure 2 shows a portion of a cross member with approximately 12 percent expansion combined with a more square section shape to handle loads at the joint just below the photo better. The seeming functional value of this contrasts with the 6 percent expansion, shown in Figure 3, where it is unclear why the expansion is needed. Reshaping the section without expanding would seem to have been able to do the task.
Figure 4 and Figure 5 show 15 percent expansion where end feeding has no effect because of the opposing friction due to leg length, section flatness, and large bends. The final section is shaped to increase height by 40 percent, which substantially increases vertical rigidity, in spite of wall thinning. Figure 6 shows that remaining formability is low enough that production robustness is highly questionable.
For designs with expanded sections extending beyond the first bend or decreasing in size through that bend, it may be best to expand it in the die because of bending complications. If expansion were contained to the first straight portion of the part, it would be beneficial to consider mechanical expansion before placing the blank in the die. That prevents some of the aforementioned potential drawbacks and can improve design flexibility.
Expansion to fit mating parts tends to be localized and may be done at the tube end to allow it to mate with another tube of similar size and shape.
The prevalent perception of the need to expand probably is rooted in the fact that in high-pressure hydroforming (HPH), the starting tube is slightly smaller than the smallest cross section of the finished part. Therefore, all cross sections must be expanded by at least a small amount to form the part. Alternatively, the blank may be mechanically preformed and placed in the hydroforming die where high pressure must be applied to finish forming the corners and iron out any unwanted deformation.
In other words, the tube is blown up to fill the cavity, which requires that tensile stress induced in the tube wall by internal pressure exceed the yield stress (YS).
The most common shape for automotive structural parts is rectangular, and the last portion of the cross section to stretch into place is the corners. Corner forming (Ri) increases required pressure, as does work hardening from stretching the material. Wall thinning (T) reduces it, requiring in substantially increased pressure to continue forming to completion.
Referring again to Equation 1, for mild steel (YS = 240 MPa [35,000 PSI]), forming 5T corners takes 48 MPa (7,000 PSI). Therefore, forming medium-severity corners don't require what most consider to be high pressure. It is calibration, (applying very high pressure to improve dimensional stability) that can call for more than 20,000 PSI.
Further benefits that have been attributed to expansion are this and an increase in YS and superior dimensional repeatability.
The latter allows the material (presumably inexpensive mild steel) to bend further before it stays bent, behaving structurally like the YS of HSLA steel.
The material chemistry remains that of mild steel, and the YS increase is uneven (arrow length indicates YS increase), in the approximate pattern shown in Figure 7. When a part is welded to another (common on structural parts), local annealing occurs. Since that also is commonly the location of highest stress, annealing must be accounted for in the design. A 50F HSLA is alloyed to have a minimum YS of 50,000 PSI when fully annealed; mild steel YS may be as low as 20,000 to 25,000 PSI.
The idea of superior dimensional repeatability is rooted in the idea that tensile yield over the whole surface dramatically reduces springback; but is unclear what that is compared to. It seems reasonable to presume comparison to HPH without calibration. It is unclear how it compares to other hydroforming techniques.
It also is likely that the flat areas do not stretch, because internal pressure tends to "lock" these areas to the die when forming begins.
Using higher YS steel requires higher forming pressure, which in turn increases the press and pressure delivery system's size and cost. Higher-strength steels have less elongation, reducing allowable stretching and increasing the likelihood of rupture. As corners get sharper, even in a small area, required pressure must increase proportionately, as does wall thickness variation.
Internal forming pressure has a large effect on overall hydroforming cost, regardless of the technique used (see Figure 8). It is most economical to use the minimum amount of pressure that will complete forming. Reducing wall thickness and YS and increasing cross sectional corner radii help accomplish this.
Optimizing these factors for lower pressure may conflict with design intent and compromise part function and must be rationalized with process viability and cost. When considering whether to make a product on a larger or smaller press of similar construction style, keep in mind that the smaller machine will have a lower cycle time.
Several techniques can be used to reduce wall thinning and/or make it less variable.
These include using special materials and tube-making methods, end feeding, lubrication, part or die texturing, preforming, annealing, and reducing bending severity to leave more elongation for hydroforming.
Process simulation is imperative to predict the outcome of these methods. Despite such measures, wall thinning still varies significantly in portions of many parts. Placing holes in the part can increase cost dramatically if they are made after forming. Always remember that each factor can add costs.
Pressure sequence hydroforming (PSH) avoids pinching without expansion. The die cavity is designed in such a way that the final part cross section peripheries are essentially equal to that of the start tube.
This, combined with low-pressure fluid in the tube during die closure, harnesses forming forces differently. The tube is reshaped by the die, while the water prevents it from collapsing. The result is that the wall is compressively forced into the corners. It dramatically lowers the maximum internal pressure required to form complex parts, normally 1/5 to 1/3 of that needed for a similar HPH part. As a result, wall thinning does not occur during PSH forming, and techniques to control it are unnecessary.
Application of higher pressure (greater than 10,000 PSI) after the die is closed does two things. It flattens any depressions in the flat areas of the reshaped part surface and supports the wall inside for hole punching. Water pressure does not form the corners; therefore, Equation 1 does not apply.
For best hydroforming economy ,cross sections are not expanded. The percentage of expansion that can be achieved with the PSH process is greater than with HPH, because material elongation used to avoid pinching in the latter process is not used for the former.
It is common to design parts with variable cross section peripheries on the supposition that varying the periphery makes the part more efficient, implying lower cost or better performance. This oversimplifies most part designs.
Designing the most effective and lowest-cost part is more complex. Concentrating on expansion can lead to assigning too little importance to other aspects. Designing a constant periphery along the part length allows access to a number of benefits that otherwise are unrecognized.
When designing a part to fulfill a function, it is best to knowledgeably consider several different manufacturing approaches. People often shorten their lists of options too early in the development process. As a result, they can build in additional costs and design limitations without realizing it. Expansion of hydroformed parts is an example of where this can occur.