If it is accepted as a primary motivation that the automotive industry is committed to reducing the weight of passenger automobiles by the extensive use of aluminum then the specific character of the problem can be defined and potential solutions investigated. For example any forming method proposed must be basically capable of the production rates common for current practice [25, 41]. This production rate requirement is a sever restriction for two of the three processes which can extend the forming limits of aluminum beyond matched tools forming. These two are fluid pressure forming, described previously and super-plastic forming. that has been omitted for reasons stated previously. Conversely, the high velocity, pulsed electric power methods, described previously, operate on a much shorter time scale than matched tool stamping while providing extended forming limits. However, with the exception of axisymmetric clinching, the electric pulse energy methods are not used by Automakers since no one has yet provided a means to apply it efficiently to large, high production parts. On the other hand, fluid pressure forming is marginally employed by the auto industry. Its use has been

principally restricted to experimental and special low production of aluminum parts. In such applications the tooling cost saving provided by the single surface tools is no longer minor in comparison to the production rate penalty. In addition , cycle time in fluid pressure forming is related to the peek pressure requirements and might be improved by combination with a pulse energy method. Not to be neglected is the capital cost of a new press machines, which would be required by the adopting of a fluid pressure forming method to produce aluminum parts. A hybrid method based principally on conventional matched tools would likely not require extensive replacement of the present, installed, press machines. However, unless aluminum alloys are developed that have the plastic strain behaviors comparable to draw steels, conventional matched tool forming will need to be abandoned or integrated with another method to meet the forming performance goals require to efficiently mass produce aluminum auto bodies .
 

2.1 Combined Quasi-Static and Dynamic Forming : Hybrid Methods

It is proposed that a well designed combination of high velocity forming integrated with a quasi-static conventional forming process could meet the requirements for a reliable, cost effective method for the mass production of aluminum auto body and other commercial parts.

There is ample evidence in the literature, as reported previously, that support the claim of extended plasticity, for many alloys, at deformation velocities above 50 m/sec. Support for reduced springback and wrinkling at high deformation velocities can also be found [5, 60]. The literature also reports on the problems involved in producing large deep shells exclusively by a high velocity, electric pulse energy process . Due to the existence of an upper deformation velocity limit (see Figure 1.2) and practical limits strength of tooling materials and capacitor bank size, the power pulses cannot be made arbitrary large in order to affect deformation over larger part areas. For example, if a very large single pulse were used, the sheet deformation velocity nearest the pulse generator would likely exceed the upper limit causing the local sheet ductility to fall off sharply. The use of an array of pulse generators to provide lower peak power per individual event and more uniform distribution of deformation forces is an obvious variation of the straight high rate-forming concept. However, the actual methods of implementation and effective control of such pulse generator arrays are not obvious. In any case, the probability is still high that the forming of the larger parts by high power pulses would involve multiple sequential discharges which will obviously tend to lengthen the total cycle time. In addition, the form tools used in a straight high power pulse forming process require a greater shock resistance capacity which generally means more massive construction. This is especially true for the electro-hydraulic discharge process. Using the high power pulses only for final forming and only at the local areas of the part which require it, reduces the overall shock resistance requirements of the tools and subsequently, the construction costs.

In order to reduce the discharge energy requirements for large parts, either multiple discharges were used or simple pre-forms were made by conventional quasi-static methods and the complex features and final sizing accomplished by high velocity methods [5]. High velocity processes generally exhibit sheet stretching over draw-in during part generation. The result can be undesirable thickness variation in deep shell geometries. The inertial forces generated by the mass of the sheet in the blank holder area, outside the energy pulse zone, increase the resistance to draw-in. Concurrently the sliding friction between the work piece sheet and the blank holder surface is reduced due the increase in the draw-in velocity. For simple axisymmetric type part geometries, these conflicting effects can counter-act, resulting in very similar draw-in performance for both high and low velocity processes [47]. However, sheet draw-in is more consistent and predictable and thus can be more finely controlled in a low velocity process.

The potential benefits from the combination of the complementary attributes of static and dynamic forming methods are clear, providing that the attributes are, in practice, additive.

One possibility is to combine a near hydrostatic fluid pressure process with submerged electric arc discharge events. The majority of the sheet deformation would be done at moderate, (< 34. Mpa (5000 psi.)) pressure. High energy-rate, "electro-hydraulic" discharge events will then provide the localized high-pressure pulse to generate the fine part detail where greater total strains are required. This hybrid process will be termed Fluid Pressure- Electro-hydraulic and abbreviated as FP-EH. Figure 2.1 is a concept schematic of a FP-EH system.

Another possible hybrid process is the combination of conventional matched tool stretch-draw forming with localized electro-magnetic pulse forming. In this hybrid forming process, the part would be pre-formed, to some optimum extent by the conventional draw-in and stretch action of the match tooling. Final forming of tight corners, sharper details and sizing would be accomplished by electro-magnetic repulsion forces generated at the required areas of the part by a set of electromagnetic coils embedded in the tool halves.

This hybrid method will be referred to as Matched Tool - Electro-magnetic and will be abbreviated as MT-EM A concept schematic of a MT-EM process system is shown Figure 2.2

A third concept is the combination of a quasi-static fluid pressure process with localized shock events generated by electro-magnetically driven shock wave tube devises instead of electric arc discharges. Since there is some evidence that shock tubes are more efficient than arc discharges in diaphragm expansion, a hybrid method using electromagnetic shock tubes may be more commercially viable than one using arc discharges [68]. This hybrid forming method concept could be technically considered a combination of the fluid pressure, electro-hydraulic and electro-magnetic processes. However its sheet forming characteristics should be quite similar to FP-EH forming although its system and energy requirements will differ. It will therefore not be given a separate name here and will be lumped with FP-EH for the remainder of this discussion.

There are no fundamental reasons to dismiss any of these hybrid sheet forming concepts. Moreover, these three process concepts are by no means exhaustive, only the more obvious combinations. The common central principle of these concepts is the combination of a relatively low power process to generate the bulk of the sheet deformation with localized high power pulses which provide the final forming, where required. The gross effect can be viewed as combining a pre-form step and a final form step into a single operation with additional process design freedom provided by virtue of the different physical processes. At a more specific level, a hybrid forming process should be able to demonstrate increased forming capability of auto body size parts with localized hyperplastic effects while avoiding the problems attendant to large energy, high power pulse events.
 

2. 2 Advantages of Different Hybrid Methods

A hybrid process which combines a quasi-static Fluid Pressure forming method with multiple, distributed, Electro-Hydraulic discharges (FP-EH) has, by several measures, the greatest general performance potential. In terms of broadness of application , a FP-EH process can be used on many different types of sheet materials. For example, it is not restricted to materials which are good electrical conductors as is required by the electro-magnetic forming process. The nature of the event ( submerged arc discharge) allows it to be located further from the sheet and with less precision then the coils of a electro-magnetic process. FP-EH requires only one form tool (usually the female die). The electrode/bridge wire assemblies in a FP-EH system would be part of the press machine and not integrated into the tool as will be the coils of a Matched Tool-Electromagnetic (MT-EM) hybrid process. The fact that each MT-EM application requires a unique set of coils further increases the general complexity and cost of the process tooling of MT-EM over FP-EH . Further, MT-EM requires a pair of form tool surfaces compared to the one for the FP-EH process. Finally, the precision with which the workpiece conforms to the coil face effects the magnetic pulse pressure generated and hence the forming energy efficiency. The repulsive sheet driving force drops rapidly (~1/R4) as the sheet is moved away from the coil surface since the pressure on the sheet is proportional to the square of the flux density, B, which in turn, diminishes as the inverse of the squared distance from the current element [61]. In contrast, the pressure pulse forming effectiveness of an electro-hydraulic discharge diminishes only as the inverse of the distance squared from the discharge, (~1/R2 ) [20] thus, much less rapidly with sheet deflection. The slower attenuation of available forming pressure makes the use of sequential discharges more practical in FP-EH than MT-EM processes. In fact, a series of smaller discharges in place of a single event of much higher energy was reported to be the preferred method for producing large parts

[21]. Although the FP-EH process concept has several advantages for broad application over MT-EM, it also has several significantly greater development hurdles to overcome.

The principle development hurdle for the FP-EH process is that it cannot be easily implemented in the types of press machines existing in the auto industry. Providing the quasi-static, fluid pressure pre-form stage requires a significant amount of specialized hydraulic machine components. Moreover, the structure of many conventional presses, currently in use, may prove too light. The structural loads, at even the lower forming pressure range, when applied over the plan area of auto body panels, can be tremendously high. A tooling system which attempted a self-contained conversion of large double acting conventional presses to fluid pressure forming was patented but demonstrated only very limited success due to pressure induced structural deflection. [39, 40] The requirement of a specialized press machine for the FP-EH process represents a significant economic road block to acceptance by industry in the near term .

Another technical hurdle to the development of a FP-EH process is the modeling of multiple interacting discharge events and their effect on deformation of the part sheet. This topic has not been investigated to any significant extent. Rinehart and Pearson [62] briefly discusses the topic with respect to multiple synchronized charges for explosive forming. They suggest the use of superposition principles in the analysis of multiple charges in under water explosive forming were the shock pressures are less than 69 Mpa (10000 psi.). A robust design method for FP-EH would require a more thorough knowledge of multiple interacting events. However, modeling even a single EH discharge event is not trivial. The electro-hydraulic discharge event begins with the complex physics involved with the generation of the high temperature (5000 - 10000 K) plasma kernel of the arc path. With in a few micro seconds the expanding plasma generates shock waves whose propagation, reflection, refraction and interferences can not be neglected in order to accurately predict the process actions. Thus FP-EH employs generally more complex and harder to model physical phenomena than MT-EM with electro-magnetic pulse events. Moreover, the simple existence of the intervening liquid medium required to transfer the deformation energy in the electro-hydraulic event, adds to the potential variability and complexity of the FP-EH process.

The MT-EM process may not have the broader applicability of the FP-EH process but, for several reasons, is a better choice for an initial hybrid process development. First, the MT-EM process can be implemented using conventional mechanical or hydraulic, single or double acting presses. In principle, only minor alterations to the presses themselves should be required. The lack of a liquid

medium to transfer the deformation energy to the part, not only reduces the over all complexity of the system, it also eliminates the maintenance overhead of an additional hydraulic system.

The reduced development advantage of MT-EM over FP-EH is exemplified by the requirements for electrode assemblies of a FP-EH process. High energy arcs can quickly erode electrode tips which in turn change the pressure pulse characteristics of the discharge. Electrode problems accounted for a good deal of the trouble encountered with the old EH machines. It was found that variations in the location arc at end of the coaxial "spark plug" electrode used in one of the early systems, could cause unacceptable variations in the parts. Moreover, the spark plugs required rebuilding after only 100 discharges. The systems which used bridge wires to initiate the arc had much better repeatability but the wires required manual installation before each discharge [24, 32, 14].

Another point is that, for axisymmetric geometries at least, electromagnetic forming has been more fully development in terms of application, tooling and coil design [13, 34]. This more organized knowledge, some available in handbook form, provides additional motivation for developing the MT-EM process. Further, electromagnetic forming developed a non-aerospace, industrial niche in axisymmetric swaging. This small commercial market supported continued work on metal deformation behavior using electromagnetic pulse energy after the military aerospace efforts ceased. Although still incomplete, this existing body of knowledge is also more current than electro-hydraulic discharge forming [18]. Thus the literature of EM forming provides a slightly higher level to start the development a hybrid process .

The preceding discussion has been presented to support the decision to proceed exclusively with the investigation and development of the MT-EM hybrid process at this time. Certain aspects of the research effort to be described in the remainder of this document will be applicable to development efforts for FP-EH methods. However, no special effort will be made to formally extend these results to the other hybrid methods previously discussed.
 

2.3) Issues Involved in Developing MT-EM Forming

The hyperplasticity effect of high velocity deformation is fairly well documented and the fundamental mechanism model of inertial stabilization has not been seriously challenged [74, 17,7]. This fundamental phenomena that hybrid sheet forming processes will be utilizing to realize extended plasticity will be described here in greater detail to support the discussion of the sheet coupon tests to follow.

The inertial effect of the sheet "particle" mass which provides a force resisting the localization of strain as a necking plastic flow instability tries to form. Hu and Daehn [44] extended the understanding of the phenomena by means of a simple and rather elegant one dimensional ridged-plastic, dynamic finite element analysis of a uniaxial tension and ring expansion test specimens (Figure 2. ). The essence of the analysis formulation was simply the inclusion of a elemental mass and acceleration term in the nodal force balance (equation. 1 in Figure 2. ) which added to the internal nodal force terms obtained from the derivative of the plastic work of the element with respect to the nodal displacements (equation. 2 in Figure 2. ).

Figure 2.3: Uniaxial and ring specimen models

( from [44])

(1.1)
 

(1.2)
(1.3)

Equation (1.3) in the is the power law of the rigid-plastic, Holloman type constitutive relationship used in their analysis. Although thermal effects due to rapid plastic stains were ignored a 1% taper in the specimen geometry was included to provide a defect like inhomegeneity. In the above equations, M is the element mass, u is the displacement (axial or circumferential), Ak is the initial cross-sectional area of the element, L is initial element length. The results of this simple one dimensional model illustrated the basic effect of mass inertia on the extended ductility at high deformation velocities. Figure 2.4 shows the graphical results presented by Hu and Daehn, most pertinent to the present discussion.

Figure 2.4 illustrates that the influence of inertia is less as n and m becomes large but contributes to extending ductility for any fixed n or m. as seen by the increase of the dynamic to static strain ratio with increasing velocity. This simple model also predicts a strong coupling between total strain at failure an deformation velocity.

The inertia effect macroscopically resembles the ductility enhancing effect of strain rate hardening which is one reason that high velocity forming is suited to the working of stain rate insensitive, aluminum alloys. To qualitatively describe the suppression of localized neck formation by inertial effects as predicted by the Hu and Daehn model, consider the following. Initially the velocity distribution of material elements in uniaxial extension varies linearly from the crosshead input velocity to zero at the fixed end of the sample. As a neck starts to form, the velocity distribution approaches a step function as the material velocity between the neck and the fixed end goes to zero while the specimen material between the neck area and the crosshead assume the crosshead velocity. In order to accommodate the velocity discontinuity the material in the necking region must experience a increasingly large acceleration. The force required to accelerate the mass of a material element outward from the neck area must be transmitted though the material outside of the necking region, thus the necking tendency is diffused. This effect is, of course, always present but only significant at high deformation velocities.

The results from the simple, one dimensional model cited above, included minor geometry variations which indicates that the inertial drag suppression of necking is not critically sensitive to sheet flaws or thinning. However, variations

in sheet hardness was not addressed in that model or in any other articles reviewed. Information on the effects of these parameters on the maximum attainable strains in hybrid forming is of interest.

From the preceding, one may expect that inertial effects at high deformation velocities will only extend plastic behavior of sheet materials whose dominant failure mode is necking. Metals which exhibit little or no necking before fracture at low velocities are not expected to show a significant increase in ductility at high velocities unless there is phenomena other than inertial drag forces at work. The direct effect of this prediction to the present work is that the fully hard aluminum alloys are not expected to perform as well as a solutionized or a lightly worked condition. In the case of hybrid forming, the inertial drag model of neck suppression will be confounded by the various levels and distributions of pre-strain introduced into the sheet material during the quasi static initial forming stage of the process. In most cases, the pre-strain will introduce work hardening into the material. The work hardening thus introduced will, in general be non-uniformly distributed across the initial-form part. In addition, variation in sheet thickness could be considerable. The extent of the variations in sheet hardness and thickness will, in practice, depend heavily on the geometry of the initial-form. The suite of experiments discussed in the next chapter were designed to elucidate the relationship between the level and distribution of pre-existing strain and subsequent material strength variations and the amount of additional useful plasticity that can be obtained under high velocity deformation conditions.

In addition, the preceding discussion indicates that one should correlate inertial controlled plasticity effects with deformation velocity rather than strain rate especially for comparisons between different geometries. The simple reason is that deformation velocity varies with gage length which means that high strain rates can generated by low deformation velocities if the initial gage length is small enough. The tendency to equate high strain rates with high deformation velocities in the literature is due to the fact that nearly all researchers are conducting investigations with identical specimen geometry for which strain rate and deformation velocity are uniquely related.

The plastic behavior of any metal is temperature sensitive at to some extent. If local work sheet temperatures become high enough during forming to cause thermal softening, then neck formation can be promoted due to the subsequent strength variation in the load path. The particular case of aluminum, the deleterious effect of thermal softening is, at least partially, offset by the fact that the strain rate hardening effect ("m "in the simple power law model,) increases with increasing temperature. The MT-EH process can induce a considerable amount of electrical joule heating as well as adiabatic heating due to dynamic plastic deformation. Sheet temperature, local to the discharge event in space and time is a process variable of interest and importance to the prediction of the MT-EM performance. The transient time-temperature data local to the forming pulse is difficult to measure directly due the m second time scale of the event alone. However, changes in sheet hardness is a process variable more directly related to plastic flow which can measured easily. Care must be exercised however in the use of superficial sheet hardness due to the confounded effects of adiabatic and joule heating with the temperature induced increase in strain rate hardening of aluminum. A simple analytic model of adiabatic joule heating can be employed to obtain a upper bound of the sheet temperature in the eddy current path. The induced eddy-current in the sheet can be estimated from the measured work coil current-time history. Obviously, the numerical simulation of the high velocity event, to be discussed later, will need to provide an accurate estimate of the sheet temperature distribution to accurately model the over all process.

The data of principle importance to the assessment of the MT-EM process are the failure strain levels, distributions, and deformation velocity for the aluminum alloy sheet material acceptable for auto body use. The present investigation will be restriction the two basic aluminum alloy types, precipitation hardening and non-precipitation hardening. The specific alloys chosen are 6111-T4 and 5754. These alloys are both currently used in auto body applications. The fundamental metallurgical differences between these aluminum alloys will result in some performance variations in the MT-EM process. The variations are expected to be in rough proportion to static measured ductility and should not confuse the resulting assessment of the MT-EM process for all similar alloys. Further, if the extended dynamic plasticity effect is largely an inertial effect, then it is reasonable to expect that static-dynamic strain relationships should be found to be applicable to whole alloy groups.

The high velocity sheet forming performance cited in the literature is almost entirely for fully dynamic deformations starting from flat blanks or uniform tubes. The state of initial cold work for these cases were at least uniform and often close to zero. The material cold work condition in a hybrid process after the quasi static forming stage will definitely be non-uniform to some extent. Depending on the part geometry and static process, the cold work condition could vary widely.

The early high velocity forming literature provides considerable information on static strengths of certain alloys after dynamic, high rate, forming which has been nicely summarized by A. A. Ezra in the last chapter of his "Principles and Practices of Explosive Metalworking", [28]. The chief concern of the aerospace researchers of that time was to determine if the high rate forming processes degraded the structural properties of their alloys. Extended plasticity was recognized but less of a concern since multiple forming cycles with intermediate annealing operations are common practice in aerospace fabricating. Therefore, the literature contains quasi-static stress-strain data after dynamic pre-straining for certain aerospace alloys. Nothing was found concerning the reverse sequence of deformations. By the path dependency of plastic deformations, it would not be expected that the combined effect of static and dynamic deformations of a sheet material is symmetric or independent of application sequence. From the data currently available it would be reasonable to expect that, assuming modest initial stage strains, that a static-dynamic sequence would produce greater elongation than a dynamic-static. Interestingly, the data summarized by Ezra, [28], shows that a dynamic-static process, in comparison to a straight quasi-static process, will reduce the total elongation for mild steels and increases it for both 5052-0 and 5456-0 aluminum. The material test results reviewed by Ezra warn against too broad a generalization of the forming performance from hybrid forming experiments with any particular metal type to another .

The preceding lead to the following questions to be answered in order to establish if a fundamental benefit can be derived from a MT-EM hybrid process.

The questions listed above concern the viability of the MT-EM process from the fundamental material mechanics, plasticity aspect. The answers to these questions must reveal an enhanced forming capability with the MT-EM process to justify any further effort on any but the most academic basis. The fact is that a significant enhancement has been demonstrated, the basics of which are discussed in the following chapter and further evidence presented in Chapter five. With this knowledge in hand, discussion of the issues concerning the modeling and design tools for MT-EM is appropriate,

Conventional matched tool forming, is itself such a complex process that analytic models have been developed for only simple axisymmetric geometries and those that can be accurately represented in one or two spatial dimensions. The sheet is generally assumed to behave as a simple membrane with bending corrections possibly included. There are a number of texts covering these analytic methods such as references [43, 54]. Luckily the past ten years have seen a good deal of effort spent in the development of computer codes which are demonstrating impressive capabilities in the modeling of the conventional low velocity deep shell sheet forming processes [ ]. The design of a MT-EM process for a large part will require the use of such a code to assist in defining the best obtainable pre-form part geometry. In the ideal world, this code would have full dynamic, electromagnetic and thermodynamic capabilities along with material constitutive relations capable of accurately predicting local necking and fracture. This ideal code would also be easy and fun to use. Such a numerical modeling tool would be capable of simulating the entire MT-EM process for the designer. Although the ideal unified MT-EM simulation code is not presently available, there are codes that can model separate aspects of the process. The details of these codes and how one might arrange them into a virtual simulation package for MT-EM forming will be presented in chapter four.

It should not be assumed that hybrid forming process and MT-EM in particular can only be applied if powerful simulation tools are available. If this were the case then the commercial viability of the hybrid processes would be quite questionable despite any extended forming capacity. In fact it is quite necessary that a means of approximating the requirements of a MT-EM system exist and be outlined. A system that requires a computer simulation before anything can be known about its gross size and energy requirements is completely untenable. Such approximate design calculations are available and can suffice to produce a functioning system at the cost of additional experimentation. The trade off between simulation and experiment is as always based on the actual cost of each including the probability of errors.

The final issue in the development of a MT-EH process concerns the physical system design. The requirements of the electromagnetic pulse coils must be combined with those of the forming tool in which they are imbedded. The fatigue strength of the tool material must be sufficient to withstand the reaction forces generated by the coil pulses over the production life of the tool. Since, the electrical conductivity of the tool material effect the energy efficiency of the coil, standard iron and steel matched tool materials may not be optimum for MT-EM tools. The coils themselves must structurally absorb internal magnetic pressure, often of similar magnitude to the forming pulse. A means of replacing damaged coils with minimum down time must be considered the same as for the high wear insert sections/components of conventional tools. The replacement of coils during the production life requires reliable electrical connectors capable of peak currents of one half million amps or more. Any arcing in coil connections causes rapid deterioration at the connection interface leading to catastrophic failure in a few cycles.

Alterations to the press machines will be minimal, which is one advantage of MT-EM over the other hybrid methods, as stated above. As an issue much subordinate to the forming performance and tool design aspects, press machine alterations will be discussed in only broad terms in this work. The press machine must accommodate the energy storage capacitor sub-system either entirely or at least the ingress of the pulse power cables. Stamping plant floor space is generally at a premium which indicates that the capacitors, charging, control and pulse energy distribution will eventually be integrated into the press machine volume. In the near term, the willingness of plant managers to accommodate a home freezer size box next to a press will be an indirect measure of the industrial acceptance of the MT-EM process.

Safety of a new industrial process is an issue to be addressed at the fundamental level early, in the development cycle. The main components of the safety issue of the MT-EM process concern the high containment of the high power electrical pulses, possible high velocity debris, eye damage from arcs at connection failures and noise levels. None of the major safety concerns represent conditions or phenomena new to manufacturing or the automobile industry in particular. These hazards all currently exist in many manufacturing environments and standard practices are in place to deal with each one. The design and safety issues involve in the development of MT-EM forming will be dealt with in detail in Chapter five.

The questions to be addressed concerning the practical application of the MT-EM forming process are listed below as a summary preceding discussion. This list is appended as subordinate to the five questions already listed.