IMPULSE METALWORKING WITH VAPORIZING FOIL ACTUATORS

Abstract

Electrically driven rapid vaporization of thin metallic foils is used to conduct impulse-based metal working operations such as dynamic compaction of metal powders, collision welding, embossing, shearing, shape calibration, and closed-die forming. A metal body is sheared from a metal sheet using foil actuators operating at input electrical energies from 4 kJ to 10 kJ. During the impulse shearing operation, the sheared plugs were accelerated up to velocities of 1400 m/s within a few millimeters of travel distance. The sheared plugs were used as pistons to compress milled commercially pure titanium (CP-Ti) and Ti-6AI-4V alloy (Ti-6-4) powders, both of which had tap densities of approximately 25-28%. After the process, compaction in the range of greater than 90% was observed. When compared to a quasistatic cold compaction process, a significant gain in densification with same pressures was observed.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a non-provisional of U.S. Application No. 61/767,888, filed 22 Feb. 2013 and it is also a non-provisional of U.S. Application No. 61/929,619, filed 21 Jan. 2014. The content of each non-provisional application is incorporated by reference as if fully recited herein.

TECHNICAL FIELD

The disclosed concepts involve impulse metalworking. More specifically, some of the applications involve the use of vaporizing foil actuators, and, more specifically, the use of the pressure created from electrically driven rapid vaporization of thin conductors to compact titanium alloys.

BACKGROUND

Impulse metalworking provides several advantages over a corresponding quasi-static process. These advantages include increased forming limits, reduced springback, low cost tooling and reduced wrinkling are some of these.

Shearing at high speeds has been shown to reduce sliver formation and provide increased dimensional tolerance. Once a critical velocity is achieved, the shearing requires much less energy because of localized deformation along narrow adiabatic shear bands.

Impact welding is a solid state process that allows the joining of dissimilar metals with little to no heat affected zone. A common observation in impact welding is that the weld zone is often stronger than the parent material.

All these factors warrant a concerted effort towards developing impulse metalworking into a mainstream manufacturing technique.

While electromagnetic (EM) forming is commonly used for impulse metalworking, the inability to develop long-lasting actuators has been (and remains) a problem. It is difficult to exceed peak pressures of about 350 MPa and generally the lifetime of actuators or forming coils decreases at high pressures. The electromagnetic launch of a workpiece is based on laws of electromagnetic induction and Lorentz forces. When a conductor, considered as a secondary coil, is placed in proximity to another conductor, considered as a primary coil, carrying a transient current, then a current opposing the change in magnetic field is induced in the former. These conductors carrying opposite currents repel each other and the workpiece is accelerated to a high velocity. The primary coil is generally insulated from the workpiece by encapsulating it in an epoxy matrix. If the cycle times are low, the joule heat developed during the process can melt the epoxy material, leading to current shortage. There are also pressure limitations on the primary coil which depend on the mechanical strength of the epoxy as well as the coil material. Hence, the application of electromagnetic forming is limited at high energies and large numbers of operations by the availability of long-lived electromagnetic coils. Besides, the workpiece either has to be electrically conductive, or it has to be driven by a conductive flyer.

Another area of interest for impulse metalworking is powder compaction of metals, especially titanium alloys, due to possible reduction in processing costs. High velocity compaction (HVC) techniques which use hydraulic impact, explosives, magnets and spring loaded hammers, generally yield higher densification than conventional methods, such as hot isostatic pressing (HIP) and die compaction. Higher densification during HVC is attributed to passage of a shock wave through the powder upon impact of the piston or the flyer plate on the powder mass.

One researcher showed that the density and bending strength of a green ferrous powder compact increases significantly with increasing impact velocity. It has also been demonstrated that the force required to eject the compressed powder mass out of the die is much reduced in case of HVC as compared to a conventional compaction process. During HVC of ferrous powder, multiple low pressure strokes could be used to obtain the same densification as would be obtained by a single high pressure die compaction process.

An opposite trend, with respect to green density gain with HVC compared to conventional compaction has been observed for a ferrous alloy powder. In that situation, it was reasoned that the HVC trials did not attain shock wave or dynamic compaction conditions.

A yet further comparative study concluded that under the same condition of stress, some nominal densification is obtained with either conventional or HVC. The major difference encountered was with respect to spatial variation of densification and spring back. The most popular method for HVC is the use of hammers weighing 35 kg to 1200 kg.

In a study involving compaction of commercially pure titanium powder (CP-Ti) in different configurations and at velocities up to 10 m/s, green compacts of densities ranging from 76% to 96% were obtained. Another study compacted Ti-6AI-4V powder (Ti-6-4) to a relative density of 86% with a single stroke compaction at 8.7 m/s. Further densification of up to 99.88% was then obtained by sintering of the green compacts.

Powder compaction using explosives can be implemented by impacting a high speed projectile or a piston with the powder mass or, in some cases, by direct compaction of the powder mass lined by metal and placed inside the explosive charge. However, due to safety regulations and difficulty in controlling the loading rates, explosive compaction can be difficult to implement in industry. Hammer-based powder compaction is practiced widely now, and is well characterized. However, the hammer-based option is not easily implemented in the small scale applications that are increasingly encountered.

It is, therefore, an unfulfilled objective in the art to provide a reliable and safe technique for compacting metal powder.

SUMMARY

This and other objects of the are achieved by a method for compacting a metal powder. In the method an accelerated metal body is generated and collided into a mass of the metal powder, resulting in compaction of the metal powder.

The colliding step occurs at a velocity in the range of 200 to 2000 m/s.

The step of generating the accelerated metal body is achieved by the steps of positioning a piece of sheet metal between a consumable metal body and a stationary body and rapidly vaporizing the consumable metal body, directing vaporized metal into the piece of sheet metal; and accelerating the piece of sheet metal into collision with the stationary body at a rate sufficient to shear off a portion of the piece of sheet metal.

In some of the methods, the stationary body is a die and the piece of sheet metal is deformed by the collision to shear off the accelerated metal body therefrom.

In many of the methods, the consumable metal body comprises a foil of aluminum in many of the methods, the rapid vaporization is achieved by passing a high current rapidly into the consumable metal body. Often, this high current is provided by discharging a bank of capacitors that are connected to the consumable metal body.

BRIEF DESCRIPTION OF THE DRAWINGS

A better understanding of the invention will be had when reference is made to the accompanying drawings, wherein identical parts are identified with identical reference numerals and wherein:

FIGS. 1(A) through 1(F) show photos of a set up and results for rapid metal vaporization as a basis for impulse metalworking, with FIG. 1(G) showing a plot of current and velocity versus time;

FIG. 2 shows a schematic circuit diagram for the rapid metal vaporization;

FIG. 3 shows a workpiece for conducting a tube expansion experiment with a rapidly vaporizing metal wire;

FIGS. 4(A) and 4(B) show the arrangement for conducting the tube expansion experiment;

FIGS. 5(A) and 5(B) show photos of U-channels used in the springback experiment;

FIGS. 6(A) and 6(B) show photos of the arrangement for conducting the springback experiment;

FIGS. 7(A) and 7(B) show workpieces used in the springback experiment;

FIGS. 8(A) through 8(D) shows the set-up for shearing of ferritic stainless steel sheets, with FIG. 8(A) showing the actual set-up, FIG. 8(B) showing a schematic of the set-up in the order of assembly, FIG. 8(C) showing dimensions of the dog-bone shaped aluminum foil and FIG. 8(D) showing the insulated aluminum foil backed by steel plate;

FIGS. 9 through 12 present graphically the experimental results of Tables 1 and 2 for tube expansion experiments;

FIG. 13 shows results from springback experiments conducted using rapidly vaporizing foils;

FIG. 14 shows current/voltage results versus time for the springback experimental results of FIG. 13;

FIGS. 15(A) through 15(D) show photographic results of shearing experiments;

FIG. 16 shows photographic results of shearing experiments that resulted in fractures of the die and the workpiece;

FIG. 17 graphically depicts current and voltage results versus time for the shearing experiments;

FIG. 18 graphically shows the dependence of driving pressure on input energy, for the two capacitor banks disclosed;

FIG. 19 graphically shows the dependence of percentage strain on input energy, for the two capacitor banks disclosed;

FIG. 20 graphically shows the dependence of peak tube expansion velocity on the burst current for the two capacitor banks disclosed;

FIGS. 21(A) through 21(C) show aspects of a system for compacting a metal powder using a sheared plug;

FIG. 22(A) graphically depicts the current, voltage and flyer velocity history for a compaction experiment and FIG. 22(B) depicts a temporal history of a sheared plug's velocity at different input energies;

FIG. 23 graphically depicts deposited electrical energy as a function of time for different amounts of input energy;

FIGS. 24(A) through 24(C) provide photographic evidence of the results of powder compaction experiments;

FIG. 25(A) graphically depicts the variation of percentage relative density versus the deposited electrical energy into the foil actuator and FIG. 25(B) graphically depicts the percentage of open porosity against the same variable;

FIG. 26 graphically depicts the percentage density obtained using the vaporizing foil actuator and a extrapolated curve of a known quasistatic compaction technique; and

FIG. 27 graphically depicts the variation of both hardness and density along the height of a piece of compacted metal powder made using the disclosed method.

DETAILED DESCRIPTION

A high, short-duration current can be used to vaporize a conductor through which the current is carried. The formed gases and plasma may continue to expand by energy deposition due to the continued electrical current. This results in a very high pressure region for a short period of time. If a work piece is kept near that conductor it will be accelerated to high velocity and useful work may be done on it. This vaporizing conductor can represent a low-cost, efficient and robust disposable actuator used for impulse metalworking. This method can be used for cutting, forming, embossing, and springback calibration and is ideal for low volume production.

A simple example can show how this technique can at once produce shearing, fast acceleration, and welding, with reference to FIGS. 1A through 1G. FIGS. 1A and 1B show a schematic representation and a top view of the experimental set-up. In these figures, a full hard spring steel sheet 10 and a dogbone-shaped aluminum foil 12 (which is shown in more detail in FIG. 1C) with a 0.051 mm thickness were held in place below a die 20, the steel sheet interposed between the foil and the die. When a nominal energy level of 4 kJ was applied to the foil, a 10 mm diameter hole was sheared out from the spring steel sheet and driven upward into a US quarter dollar coin 14 at the opposite end of the die. The spring steel sheet after the experiment is shown in FIG. 1D. When the plug sheared from the spring steel impacted the is made to impact a US quarter 14, the plug not only sheared a 10 mm hole in the quarter, but it also welded with the surface of the quarter where the impact occurred, as shown in FIGS. 1E and 1F, the latter of which shows in cross-section the welded product of the impact of the spring steel and the quarter.

FIG. 1G shows a velocity-time history overlaid with a current-time history of the experiment. The velocity data were measured using Photon Doppler Velocimetry (PDV). It is noted that the sheared plug undergoes very rapid acceleration once it shears from the sheet metal and reaches a peak velocity of about 375 m/s. It starts decelerating slightly after that, most likely due to air-resistance and friction against the wall of the die.

Additional applications of this concept involve tube expansion and control of springback. Tube expansion is considered a model system for actuation using vaporizing wires. Springback is the tendency of a cold worked metallic part to partially return to its original shape after the forming force is removed. It is a very common phenomenon in sheet metal forming and is due to the elastic nature of any metal. Springback depends not only on the mechanical properties of the metal, but also on the geometrical features like bend radius, thickness, bend angle and the moment of inertia of the shape created. As will be seen, the examples presented demonstrate that very light equipment and low input electrical energies can perform ‘difficult’ metal working operations.

The experiments conducted were based on the simple circuit shown in FIG. 2. A large amount of charge stored in a capacitor bank flows across and can vaporize a thin conductor. With rapid discharge from the capacitor bank, high currents can be developed before the wire vaporizes, storing some inductive and thermal energy. This stored energy, along with the gases produced by vaporization of the conductor material and their subsequent reactions, creates a high pressure on the surrounding material.

Two capacitor banks were used for charge storage. The first one (faster capacitor bank), is a commercial design from Maxell-Magneform. It has a maximum capacity of 16 kJ at a maximum charging voltage of 8.66 kV. As will be seen later, only a fraction of the maximum energy was used in all the experiments. The unit has 8 capacitors, each triggered by an ignition switch. This unit has a total capacitance of 426 μF and internal inductance of 100 nH and a primary circuit resistance of about 10 mΩ. This gives a rise time in shorted circuit of about 12 μs. All 8 capacitors were used in the circuit in these experiments.

The second capacitor bank (slower capacitor bank) has a maximum capacity of 24 kJ at a charging voltage of 10.9 kV. This unit has 4 capacitors divided into sets of two, which can be triggered separately using spark gap switches. In the work described here, all of the capacitors were engaged in the circuit. Total capacitance of the unit is 404 μF, which is lower than that of the first bank. The inductance of the slower capacitor bank is 100 nH and its dynamic resistance is 4 mΩ. The short circuit rise time of this bank is 10 μs. However, a relatively high inductance connection was used to adapt the experiment to this capacitor bank, and gave a measured short circuit rise time of about 35 μs, much larger than that of the other system.

Copper alloy 122 tubes having a 25.4 mm (1″) inner diameter, 28.57 mm (1.125″) outer diameter, and 76.2 mm (3″) length were annealed and quenched to provide a soft and uniform starting tube condition. A tensile test was conducted in the hoop direction of the tube and it was found that the yield strength of the material was 105 MPa and the ultimate tensile strength was 206 MPa. Uniform strain of 38% and total elongation to failure of 43% was also obtained. Corona dope was applied at the ends of the tube to prevent current shorting to the specimen. A durometer 80A urethane rod with a diameter of 25.4 mm (1″) and length of 78.74 mm (3.1″) was inserted into the rod as shown in FIG. 3. Aluminum and copper wires of varying diameter were inserted in a matching hole drilled through the center of the elastomer rod.

The horizontal terminals of the capacitor bank are vertically aligned using a unique apparatus known as the FIRE (fully instrumented ring expansion), a system developed by Johnson. The system was originally set-up for material constitutive property determination using electromagnetic ring expansion; however, it can be easily adapted for ring/tube expansion using vaporizing wires. The opposite ends of the wire were connected to the positive and negative terminals of the FIRE system using split connectors made out of brass. The split connectors were held tightly in position inside plastic insulators by 4 set screws as shown in the schematic in FIG. 4(A). The set-up was vertically pressed together until the plastic insulators touched the ends of the copper tube, ensuring a minimum air gap between the wire, urethane, and tube as well as ensuring the same internal pressure at the beginning of each experiment. Aluminum and copper wires with diameter of 1.524 mm (0.06″) were used.

Current-time and voltage-time histories were recorded using a Rogowski probe and a 1000:1 voltage probe, respectively, in conjunction with an oscilloscope which can acquire data at 5 Gs/s. Tube velocity was measured using the photon doppler velocimetry (PDV) technique. The probes were protected from direct impact by any tube fragments by using a periscopic adaptation. A schematic of this arrangement is shown in FIG. 5(B). Two PDV probes were directed towards the wall of the tube and were 120° apart from each other to assess expansion symmetry.

After the experiments, tubes were visually inspected for signs of current passing through them instead of the wires. This was done by looking for black deposition on the ends of the tubes which would form if there was sparking. The maximum diameter at the middle of the tube was measured using calipers.

A simple calculation can be done to estimate the sustained initial pressure on the inside wall of the tube. Based simply on using Newton's second law and ignoring the plastic resistance of the tube, the rate of acceleration can be equated to the pressure inside the tube as:

P=ρ.{umlaut over (x)}.t

where ρ is density, {umlaut over (x)} is the rate of acceleration and t is the tube thickness. The rate of acceleration can be easily estimated from the PDV data.

To study the effect of current rise time on the final strain, similar experiments were done on slower capacitor bank at 9.6 kJ. The repeatability of this method was also verified. All of these experiments were done with aluminum wires with diameter of 1.524 mm. Single-channel PDV was implemented along with current and voltage measurements. In each experiment, efficiency was estimated as the ratio of the maximum kinetic energy imparted to the tube and the electrical energy stored in the capacitor bank prior to discharge.

Experiments with copper wires were done with the faster capacitor bank at an energy level of 8 kJ and one-channel PDV was used.

U-channels are widely used for various applications like holders, covers, brackets, rails, slides and supports. Roll forming is a common method of manufacturing U-channels. This is a multistep forming operation in which a sheet metal is progressively formed into the final shape. In order to compensate for the springback, the roll formed part has to be bent beyond 90°. The extent of this allowance depends on the material and geometrical properties of the component and tools, and has to be calculated by careful numerical simulations. The U-channel parts which are calibrated in this work were supplied by Labein Tecnalia, and were obtained by a drawing operation. This work builds on the previous work done by them as reviewed in the background section. As can be seen in FIGS. 5(A) and 5(B), there was significant springback in the drawn component. The bend angle was more than the desired angle of 90°. There is also a significant sidewall curl which is not the focus of this work. The springback angle was measured at two locations: at the bottom and top of the part wall. This was measured by tracing the edge of the part onto paper and using a protractor to find the angle. Therefore, it must be noted that this angle is an approximation only, but allows the result to be quantitatively measured quickly and easily. This technique is accurate to approximately one half of a degree. The springback angle measurement for an un-calibrated part is shown in FIG. 5(A) and 5(B). Angle α refers to the angle at the bottom of the wall and angle β refers to the angle at the top of the wall. In this case, a was 96° and β was 104°. A smaller angle after calibration means that more springback was eliminated.

The part investigated in this work is a U-channel made from high-strength DP600 steel. The steel has a yield strength of 300-470 MPa (43.5-68.2 ksi) and a tensile strength of 580-670 MPa (84.1-97.2 ksi). Polyimide tape was used as an insulator between the foil, workpiece, and die. The setup was clamped together and the foil was connected to the terminals of the faster capacitor bank. The experimental set-up is shown in FIG. 6, where FIG. 6(A) shows the parts inside a match-metal die with foil placed outside the part and FIG. 6(B) shows the assembly connected to the terminals of the capacitor bank. The current and voltage traces were taken in real time during the experiment.

The width of the narrow area of the foil and thickness of the foil were the parameters that were varied. Widths of the foils used were 12.7 mm (0.5″) and 25.4 mm (1″) while thicknesses of the foils used were 0.0508 mm (0.002″) and 0.0762 mm (0.003″). The length of the narrow section was 63.52 mm (2.5″). When used, the 12.7 mm foils were placed near the bottom of the wall while the 25.4 mm wide foils were placed near the middle. This arrangement has been shown in FIG. 7, where FIG. 7(A) shows a 12.7 mm (0.5″) wide foil at the lower wall location and FIG. 7(B) shows a 25.4 mm (1″) wide foil in the middle of the upper wall location.

Ferritic stainless steel sheets with a thickness of 0.8 mm and hardness of 58 HRC, were sheared using the apparatus shown in FIG. 8(A). A thin aluminum foil shaped in the form of a dog-bone ensures foil burst in the area of interest, shown in FIG. 8(C) and 8(D). Aluminum foils of thickness of 0.1524 mm (0.006″), 0.127 mm (0.005″), 0.0762 mm (0.003″), 0.0508 mm (0.002″) and 0.0254 mm (0.001″) and an active length of 45 mm were used. They were insulated from the workpiece which was positioned above them by polyimide tape. The side of the workpiece facing the foil also had a layer of insulator to prevent the current from shorting. The workpiece had a commercial shearing die directly above it and modest pressures were provided by bolts that held the system together. An ovoid shearing die (part number HC14713, Cleveland Steel Tool) was used. The stack of aluminum foil, workpiece and the shearing die was held together using a fixture made of mild steel. Once the fixture was bolted together, the opposite ends of the foil were connected to the faster capacitor bank. Energy levels ranging from 1.6 kJ to 10.4 kJ were used to vaporize the foils. For the 0.1524 mm thick foils, the input energy was between 7.2 kJ and 10.4 kJ, while the 0.0762 mm foils were vaporized at energies between 7.2 kJ and 4 kJ. Experiments with thinner foils were done with input energies between 4 kJ and 2.4 kJ. Some experiments were done with two thin foils sandwiched together.

Now that the experiments have been described, the results are discussed in the following tables and the graphical presentations of the attached Figures.

First, the results from the tube expansion experiment using the faster capacitor bank are presented:

TABLE 1
					Peak
Energy	Rise	Burst	Final		velocities		Peak
level	time	current	OD	%	Channel 1	Efficiency	Pressure
(kJ)	(μs)	(kAmps)	(mm)	Strain	and 2 (m/s)	(%)	(MPa)
9.6	17.2	143	42.9	50	123, 131	>6.7	282
8	18.8	131	40.6	42	112, 106	5.9	212
8	19.6	132	41.9	46	118, 112	6.5	225
6.4	20.4	120	36.1	26	90, 101	6	161
6.4	19.6	118	35.6	25	87	4.4	172

Next, the results of the tube expansion experiment using the slower capacitor bank are presented:

TABLE 2
Energy	Rise	Burst	Final				Peak
level	time	Current	OD	%	Peak	Efficiency	Pressure
(kJ)	(μs)	(kAmps)	(mm)	strain	Velocity	(%)	(MPa)
9.6	35.8	52	36.8	29	80	2.5	154
9.6	35.8	53	36.5	28	79	2.4	152
9.6	37.9	52	35.8	25	—	—

The data from Table 1 are presented graphically in FIG. 9, where the experimentally measured current, voltage and velocity are shown for an expansion of copper tube using electrical vaporization of 1.524 mm aluminum wire at 9.6 kJ input energy from the faster capacitor bank. The jump in voltage occurs at the instant when the wire vaporizes and its resistance increases suddenly, leading to excess voltage. On the right, comparison of initial and final shape of the tube is shown.

The data from Table 1 are also presented in FIG. 10, where experimentally measured current, voltage and velocity are shown for the expansion of copper tube using electrical vaporization of 1.524 mm aluminum wire at 8 kJ input energy from the faster capacitor bank.

The data from Table 2 are presented graphically in FIGS. 11 and 12. In FIG. 11, the experimentally measured current, voltage and velocity are shown for expansion of copper tube using electrical vaporization of 1.524 mm aluminum wire at 9.6 kJ input energy from slower capacitor bank. The longer time the current takes to reach its peak value must be noted.

FIG. 12 shows experimentally measured current, voltage and velocity for expansion of copper tube using electrical vaporization of 1.524 mm copper wire at 8 kJ input energy from faster capacitor bank.

The specifics of the results from the faster capacitor bank are shown in Table 1, as well as FIGS. 9 and 10 for some experiments carried out using an aluminum bridge wire. The final configuration of the tube is also inset in each case. When a tube was expanded using an aluminum wire with a diameter of 1.524 mm at a 9.6 kJ energy level, the maximum velocity of the expanding tube was found to be 123 m/s on PDV channel 1, while it was 131 m/s on channel 2. However, the tube fractured as shown in FIG. 9. At an input energy of 8 kJ the tube expanded without fracturing, as shown in FIG. 10. The expansion was relatively uniform along the length of the tube, except at the ends that flared outward. The final outer diameter was measured as 40.6 mm, resulting in a 42% strain. A nominally identical replicate experiment gave a final diameter of 41.9 mm; hence, the radial strain was 46%. Similar experiments were done at 6.4 kJ as well and the results are shown in Table 1. Efficiency of this process has been calculated based on the conversion of input electrical energy into kinetic energy of the tube. This energy converts into potential energy, i.e. plastic work, during expansion of tube. The time period required for the current to reach its peak value was in the range of 16-18 μs.

The specifics of the experiments conducted with the slower capacitor bank are shown in Table 2, and FIGS. 11 and 12. The peak velocity is in the range of 79-80 m/s while the final strain is in the range of 25-30%, which is fairly consistent. The current was also quite low as compared to those from the faster capacitor bank; it reached its peak in approximately 35 μs. However, the final strain was similar to the earlier experiments as shown in Table 2. A typical result, using the slower capacitor bank, is shown in FIG. 11.

A tube driven by the vaporization of copper wire at its axis is shown in FIG. 12. The maximum strain was found to be 15.5%, while peak velocity during expansion was 32 m/s. There was no evidence of loss of energy by a current shortage through the tube. The experiment was repeated without velocimetry and the resulting strain was again very near 15%, a much lower value than is obtained with the aluminum wire. The deformation was concentrated near the middle of the tube. Due to the low effective efficiency of the copper wire, most of the experiments were focused on aluminum wires and foils.

The impulse created by the vaporizing foils was very effective in mitigating springback, as is seen in FIG. 13, where some of the best results, graphical and numerical, are presented. For the as-stamped components, the angle at the bottom of the wall was 96°, while that at the top was 104°. The most optimal result was obtained with 0.003″ thick foil vaporized at 8 kJ and placed near the bottom of the U-channel. The top as well as bottom angles were measured to be 91° which is very close to the desired angle of 90°. The time-resolved current and voltage history of the experiment is also shown in FIG. 14. The position of the actuator did not seem to have significant effect on springback reduction. On the other hand, higher burst currents did lead to better results. In some prior work reported in the art, significant inhomogeneous deformation was introduced by the non-uniform pressure from the electromagnetic actuator path. In this case, no excess surface undulations were produced by this impact calibration exercise.

With regard to shearing, experiments were done in a range of energies and foil thicknesses. Shearing experiments at lower energies and with thinner foils were more successful in producing desirable commercial-quality shearing. The experiments done with 0.0508 mm thick foil resulted in cutouts as shown in FIG. 15. The minimum energy at which clean shearing was attained was 1.6 kJ. The smooth edges resulting from shearing with thinner foils vaporized at relatively low energies are shown in four examples depicted in FIG. 15(A). Below 1.6 kJ, the foil did not vaporize. An optical micrograph of a sheared edge is also shown in FIG. 15(B). In general, at lower energies the cut quality is better, and the force on the body of the die is reduced. The SEM images of the shear surface depict some regions of localized melting, especially along the curved edge, as shown in FIGS. 15(C), 15(D). These areas can be topic of further investigations while using this technique for shearing.

The results above 5 kJ energy input and with foil thicknesses of 0.1524 and 0.127 mm were profoundly suboptimal. It was observed that none of the cutouts had smooth edges. Also, the workpiece fractured or sheared at the rounded edge of the die, even though the pressure in that region was not intended to be high. In some cases the die was fractured. Rough fracture of the desired cutout region and fracture of the outside of the die region (similar to that seen in FIG. 16) was also found with an 0.0762 mm vaporizing foil burst at energies between 4.0 and 7.2 kJ.

FIG. 17 shows typical experimental current and voltage histories for 0.0508 mm foils burst at input energies of 2.8 and 4.0 kJ. Characteristic deviations are indicative of the time of sublimation.

Discussion of Results

Some interesting observations can be made based on the forgoing experiments. First, it was verified that aluminum is a better bridgewire material than copper for this application. This can be attributed to the fact that gaseous aluminum is very reactive and readily form oxides and nitrides. These reactions are highly exothermic leading to an increase in the temperature and pressure of the gases (i.e., the enthalpy of formation for alumina is about 1700 kJ/mol). Just to provide an estimate, if the 76.2 mm long aluminum bridgewire with 1.5 mm diameter is fully converted to alumina, about 3.1 kJ of energy is released from this reaction. On the other hand, if a similar copper wire is vaporized, less than 1 kJ of heat will be evolved from its conversion into oxides. These oxidation reactions can be quite delayed as compared to the instance of burst that was shown previously by others. However, in the tube expansion experiments presented here, there is a pressure transfer medium in between the vaporizing conductor and the workpiece. As the delayed pressure pulse travels through the medium it can get shocked up.

The use of two-channel PDV in the experiments showed that the initial expansion of the tube was axisymmetric as the velocity-time traces of two points 120° apart overlaid quite well. However, when the first pressure pulse is complete and the tube is expanding freely, a slight imbalance in initial velocities of different points can cause asymmetry of expansion. On average, the expanded tubes were symmetric along the circumference. There were some end effects that caused flaring at the ends. The velocity traces have two or more local peaks because of separate pressure impulses provided by the urethane rod. The first peak occurs due to the initial pressure impulse created by the burst of the wire. This impulse gets reflected off the wall of urethane rod and travels to a diametrically opposite point, leading to a second peak after some period of time.

The pressure impulses remain for many microseconds and result in efficient transfer of energy into the tube. There may also be a shock wave that is much shorter in duration that is not detectable using this technique. Peak pressures up to 300 MPa were estimated based on this measurement method discussed in the procedure section. There are significant error bars, as shown in FIG. 18, which shows the change in driving pressure with input energy, due to difficulties in estimating the slope of the velocity vs. time plot.

Direct measurements of pressure using fast piezoelectric gauges can help in verifying these estimates. Although these pressures are significantly less than those reported in the past, it is also important to note that higher pressures can be obtained after the tube has been accelerated to high speed when it may impact another rigid solid. This impact pressure is much greater than the launch pressure and may be useful in shearing, embossing, or springback elimination.

The present experiments show a more rapid current rise time, which is in tens of microseconds as compared to a few microseconds in the known prior art. With shorter rise time, more energy can be deposited into the wire before the instabilities start to set in. This leads to a higher burst current density. Hence the gases have more kinetic energy to exert pressure on the surrounding material. The present work shows that high pressures and launch velocities can be developed for vaporizing foil or wire experiments, even if the current pulse has rise time in tens of microseconds. As shown in FIG. 19, with 9.6 kJ input energy, experiments on the faster capacitor bank produced 50% strain while those on the slower capacitor bank produced only 25 to 29% strain. Also, if the efficiency of transfer of energy to the flyer is calculated using prior art methods, the kinetic energy is in the range of 1-5%, which is quite similar to the results of others.

One of the prior art workers argued that the Gurney energy of the vaporizing conductor depends only on the burst current density (I_b). This might be the case when the rise times are a couple of microseconds, but the data from this work shows that I_b is not the only contributing factor here. Other discharge parameters like rise time (which depends on capacitance and inductance of the system) also matter. FIG. 20 shows the peak velocities for different burst currents from experiments done on the two capacitor banks. It is observed that with the faster capacitor bank, a burst current of 120 kAmps resulted in peak velocity of nearly 80 m/s which was similar to the result obtained from the experiment done on the slower capacitor bank when the burst current was just 53 kAmps. Therefore, it is seen that for very different values of I_b from different capacitor banks, very similar peak velocities can be obtained.

From a practical standpoint, the two most important advantages of the technique presented here are that the electrically stored energies required are quite modest and that there is no limitation imposed by the potential failure of the electromagnetic actuator or coil. The springback calibration experiments were carried out at much lower energies than those used by others, who used electromagnetic force based on Lorentz repulsion to conform the U-channel into the final shape. While that work obtained the best results at an input energy of 24 kJ, the experiments presented here give similar or superior results at a mere 8 kJ. This can be attributed to the higher shock pressures that are developed more quickly using vaporizing metal as compared to electromagnetic forming. Although the driving force for springback reduction in both cases is a pressure/shock wave, higher magnitude and lower pressure rise time magnifies its effect. In this study, there is no significant current flowing through the workpiece. So, if not in general, at least in this case electro-plasticity can be eliminated from the list of possible mechanisms for springback reduction. Further work is ongoing to measure the impulse pressure with high temporal resolution. It is possible that there is a brief shock wave with a peak pressure that is much greater than the flow stress of the workpiece.

Shearing experiments using vaporizing foils showed that, to get the best results, it is not always necessary to increase the input energy level. In fact, higher input energies were found to produce negative results like cracking of die and rugged sheared edges. More specific energy needs to be deposited in a thicker foil before it bursts. So, with all the circuit parameters nearly the same, a thicker foil takes longer to destabilize and burst. Thus, even though a higher pressure can be reached with a thicker foil, the pressure evolution is slower as compared to that from the burst of a thinner foil. Therefore, in a shearing experiment, it takes longer for the pressure to reach the shear strength of the workpiece material if a thicker foil is used. In that case, the workpiece bulges into the shearing die, and a bigger area of the workpiece experiences the pressure. All this leads to undesired modes of fracture of the workpiece and sometimes even the die. This means that, for a given foil geometry, there exists an optimal energy range within which we get reproducible and practically burr-free shearing.

The shearing work finds direct application in the compaction of metal powders, specifically titanium powders. When impulse shearing is achieved at a proper rate by passage of a capacitor bank-driven current on the order of 100 kA, the resulting burr-free sheared edges with high dimensional tolerance can be used to form a plug that will travel at supersonic speeds. The sheared plug may be used as a driver for dynamic compaction of CP-Ti and Ti-6-4 powders. This technique offers characterization of powder compaction processes at velocities and pressures which cannot be easily attained with techniques that use heavy hammers. The equipment used are the capacitor banks described above. The equipment is compact and agile. There is some significant prior work on cold compaction of the same powder material that was used in the work here. The variation of percentage densification with pressure will be compared to the data obtained in this work.

The shearing apparatus discussed with regard to FIG. 1 can be adapted for the inventive concepts described here. As shown in FIG. 21, it was adapted for shearing of a 0.8 mm thick Cu110 alloy sheet through a 10 mm hole, precision-machined into a S7 tool steel block. The shape of the 0.0762 mm thick vaporizing aluminum foil, as seen in FIG. 21(A), was designed to cause preferential vaporization in the region where shearing was desired. The energy source used in these experiments was a Maxwell Magneform capacitor bank with a total capacitance of 426 μF, inductance of 100 nH, and short circuit current rise time of 12 μs.

When charged to the maximum voltage of 8.6 kV, this capacitor bank can supply 16 kJ of electrical energy. The velocities of the sheared plugs at input energies of 4 kJ, 6 kJ, 8 kJ and 10 kJ were measured using PDV, using the setup shown in FIG. 21(b). Current and voltage were also recorded. Powder compaction was implemented in a separate run of experiments, using an arrangement as at FIG. 21(C), but at the same charging energies as the velocity experiments. The sheared plug driven by the expanding plasma was used as a piston to press the Ti powder at high strain rate. These experiments were repeated once to test reproducibility. Titanium powders used in this experiment were obtained using the Armstrong process, which is a low-cost, high-output alternative to traditional methods like the Kroll's or Hunter Process. The Armstrong process was recently developed by International Titanium Powder (ITP). Details regarding powder extraction and characterization can be found in the prior art. The starting density of the milled CP-Ti powder was measured as 25% (of 4.506 g/cc) while that of the milled Ti-6-4 powder was 28% (of 4.43 g/cc). The relative density of the green compacts was measured by the Archimedes method. Percentage open porosities were also estimated by finding the difference in the masses of dry compacts and those sonicated in water. Microhardness testing was done, with a 300 g load and diamond tip, along the height of a Ti-64 compact and along the diameter of a CP-Ti compact. Since hardness of a compact is known to be directly related to the local density, these measurements also indicated the variation of relative densities of the compacts.

The data from velocity measurement experiments are shown in FIG. 22, which depicts the current, voltage and flyer velocity history for a shearing experiment done at 8 kJ input energy (at FIG. 22(A)) and the temporal history of the sheared plug's velocity for different input energies (at FIG. 22(B)).

From FIG. 22(A), it should be noted that the foil burst corresponded with a sudden rise in current and voltage traces. That moment was also when the sheared plug underwent rapid acceleration to a velocity of 1422 m/s. The temporal evolution of the velocity of the sheared plugs is shown in FIG. 22(B), and shows that the plug acceleration increased with increasing input energy. This indicates that the driving pressure followed similar trend with respect to input energy.

Based upon a mass of the sheared plug of 0.6 gram, the mechanical efficiency, calculated as the ratio of the kinetic energy of sheared plug at maximum velocity and the input energy, was found to vary from 5.9% to 8.9%. It should be noted that the maximum velocity of the plug could be higher than what was apparent from the traces that only show the data until the plugs were in the PDV probe's field of vision. Additionally, maximum velocities would be higher if the flyers were allowed to be accelerated over a longer distance. Based on the acceleration of the plugs the maximum (from point A to B in FIG. 22(A)) and average driving pressures (from point A to C in FIG. 22(A)) can be calculated.

The maximum driving pressure of 1.3 GPa was estimated for the 10 kJ experiment.

Deposited electrical energy was calculated by integrating the product of measured current and voltage over time. It can be seen from FIG. 23, where deposited electrical energy is plotted as a function of time, that the total energy deposited in the foils and plasma formed after burst was less than the intended input electrical energy. Therefore, only those amounts of energy were available for mechanical work, and should be considered while developing densification-energy relations. It is not clear why the deposited energies in the foils of the same geometry, when applied for compaction of different powder materials were different. It is possible that the higher strength of the Ti-6-4 powder led to a greater constraint on the expanding plasma, thereby leading to a higher electrical resistance. Further investigation is required to understand this phenomenon satisfactorily.

FIG. 24 provides a graphical representation of the results of the powder compaction experiments. FIG. 24(A) shows cylindrical compacts that were slightly convex on top and concave at the bottom, which is generally observed in HVC. This could lead to non-uniform density distribution as shown later in this section. These distortions were found to mirror the resulting deformation in the sheared plug and steel constraint. The convexity of the sheared plug could be attributed to its bending before shearing from the parent sheet. Therefore, in future experiments, the use of stronger materials for sheared plug and steel constrain would help avoid distortions in the compact specimens. Lubricants can also be used to reduce friction from the die wall on the sheared plug and the powder mass during the compaction process. The densest of the Ti-6-4 compacts (93% dense) was sectioned longitudinally while a transverse section of the densest CP-Ti compact (97% dense) was mounted and observed under optical microscope. The mounted sections were etched with Kroll's reagent for 10 seconds to expose the powder grains.

FIG. 21(A) shows the variation of percentage relative density with electrical energy deposited in foil actuator, and FIG. 25(b) shows the variation of percentage open porosity with electrical energy deposited in foil actuator. FIG. 25(a) demonstrates that the increase in densification with increasing input energy was more pronounced for the softer of the two materials, which was CP-Ti. Open porosity, as a percentage of total volume, ranged from 0.6% to 3.3% for CP-Ti while it was found to be from 1.8% to 12.5% in Ti-6-4 compacts. On average the open porosity decreased with increasing input energy. It can be seen that increase in densification diminished slightly upon going over 8 kJ for both materials. It must be noted that the powders were already highly densified at that energy: 97% for CP-Ti and 93% for Ti-6-4, and there was limited scope for further compaction. It has been observed in related work not reported here that, for a given foil thickness and electrical circuit parameters, increasing input energy above a certain level does not necessarily result in higher kinetic energy of the flyer. That threshold for the foil actuator used in this work seemed to be 8 kJ, above which the efficiency of foil plasma in shearing the copper sheet and propelling the flyer went down.

In one work of the prior art, Chen quasistatically cold-compacted a similar milled Ti-6-4 powder as was used here up to 690 MPa. Their density vs pressure data was well fit by the Fanelli equation, which is extrapolated here to higher pressures in FIG. 26.

The minimum mass of Ti-6-4 powder that was used by Chen was 2 g, which is similar to the amounts used in present work. According to the Chen model, for a length:diameter ratio of 0.5, which was the case in the present work, in order to get to 93% densification more than 4 GPa of quasistatic pressure would be required. However, here that degree of densification was obtained with a maximum estimated pressure of 1.3 GPa. Although this value was estimated based on the acceleration value from a separate experiment, it is not expected to change by much even when measured directly. Therefore, with this dynamic compaction method a significant gain in densification was realized as compared to a conventional compaction process.

In FIG. 27, the variations of the green compacts' hardness along diameter and height of the specimens are shown. For longitudinal section of Ti-6-4 compact, hardness was measured along the axis while that for the transverse section was measured along the diameter. The average hardness of Ti-6-4 compact was found to be 250 HV while that of CP-Ti compact was 200 HV. Hardness variation was observed in both the directions of measurement, with greater deviations observed along the axis. The densities of the powder compacts are also expected to vary in the similar manner as the hardness. FIG. 2: Hardness and density variations along the height of 93% dense Ti-64 compact and 97% dense CP-Ti compact obtained at 8 kJ input energy with VFA. Direction of measurement is shown inset in the figure.

Therefore, in FIG. 27, f(relative density), was also plotted on the y-axis. The function is expected to vary for different materials. Additionally, as pointed out in the work of Kandeil, the linear relationship between density and hardness should be taken with an appraisal for possibility of hardness increase due to strain hardening, and in the present case, due to strain rate sensitivity. However, under the reasonable assumption that the strains and strain rates remained the same throughout the mass of the powder compact, the proportionality between density and hardness should still hold true. This has been shown by Yi through a comparison between hardness/density plots for HVC, conventional compaction, and quasistatic compaction processes.

The sheared plugs reached appreciable velocities within a few millimeters of travel distance. Therefore, if high velocity compaction was to be planned in future, the powder mass could be placed at a distance of approximately 10 mm from the initial position of the sheared copper sheet. This would allow accurate comparison of data obtained by VFA compaction to HVC compaction data such as the ones obtained by Yan. In their work, they found that the relative density of green compacts varied almost lineary with respect to impact energy per unit powder mass. However, in the present run of experiments there is no impact. Rather, the sheared plug pushed on the powder mass from the start of its launch. Due to this, the process presented here is termed dynamic compaction rather than HVC.

One important factor that has not been considered in this work is the force required to eject or withdraw the powder body from the die. According to Wang, the ejection force for HVC did not have a clear relation with the impact velocity; however, it was found to be 4 times less than the ejection force required with conventional compaction of ferrous powders. Smaller radial stresses and longitudinal expansion after removal of impact force were found to be the reason for this phenomenon. Further characterization of the VFA based process, in terms of ejection force, would help in comparing it to the traditional processes.

This shows that the vaporization of small metallic cross sections is an agile, repeatable, inexpensive, and efficient impulse metalworking technique. This method avoids lifetime issues presented by electromagnetic actuators. Unlike explosive forming, this technique can be applied at smaller scales, the pressure distribution can be controlled, and the process can be implemented safely in traditional factory or laboratory environments. There are additional manufacturing applications, such as powder compaction, sheet metal forming, and collision welding which will be reported upon. This work shows the practical potential of this work. There is a clear need to develop a physics-based design science to support new manufacturing methods based on this technique. Additionally, the development of this technique is not intended to replace the existing high rate forming methods such as EHF, EXF or EMF, rather the purpose is to supplement them where their application is not suitable.

In terms of developing process fundamentals, fully instrumented experiments were carried out on the expansion of annealed 25.4 mm inner diameter and 28.6 mm outer diameter tubes of copper alloy 122 by electrically driven vaporization of coaxially placed metallic wires. Significant final deformations over a length of about 76.2 mm were obtained. About 50% and 30% increments in diameter were developed with discharges up to 9.6 kJ from separate capacitor banks. Aluminum, gave better efficiency that copper in developing mechanical impulse. The reason for this has been hypothesized to be the highly exothermic reactions aluminum vapor undergoes, to form oxides and nitrides.

Efficiency of the process was not only dependent on the wire material, but also on the discharge characteristics supplied by the capacitor banks. Increased burst currents correlated with increased mechanical work. However, it is not the only factor. It was found that by increasing the rate of increase of the current, the efficiency of the process can be increased. Although a good efficiency of about 7% was attained in the present work, faster capacitor banks which can be charged to higher voltage would make this process even more effective.

This work also demonstrated preliminary results from two prototypical manufacturing operations: springback calibration and shearing. Springback calibration was carried out at low energy levels for U-channels drawn from 1.27 mm thick DP 600 steel sheets. Drawn U-channel samples were calibrated to become very close to the desired 90° angles, close to specification, and the study demonstrated that vaporizing foil forming is a viable process for springback control of high-strength steel parts. Full hard ferritic stainless steel sheets were sheared at a low input energy of 1.6 kJ and it was shown that high energy discharge is counterproductive for this application. Circular plugs sheared from the steel plates using this technique were found to travel at high velocities and can be used to conduct impact experiments. Very low capacitor discharge energies, often under 3 kJ were shown to be very effective in shearing.

Based on this work involving dynamic compaction of Ti powders with Vaporizing Foil Actuators (VFA), several conclusions can be made.

First, VFA shearing is a feasible tool for making small scale cylindrical compacts of Ti powder. Maximum relative density of 93% with Ti-6-4, and 97% with CP-Ti was obtained with simple pulsed power equipment in conjunction with the robust and versatile VFA tool. This offers a method for laboratory scale investigation of powder compaction at strain rates which could generally be attained only with explosives.

Second, pressures of 1.3 Gpa and driver velocity of 1.4 km/s were reached and energy conversion efficiency of up to 8.9% was possible. In-situ measurement of velocity and pressure during powder compaction was not implemented and is intended as future work.

Third, significant enhancement in densification as compared to a corresponding quasistatic process at the same pressure was observed.

Claims

1. A method for compacting a metal powder, comprising the steps of:

generating an accelerated metal body; and

colliding the accelerated metal body into a mass of the metal powder, resulting in compaction of the metal powder.

2. The method of claim 1, wherein:

the colliding step occurs at a velocity in the range of 200 to 2000 m/s.

3. The method of claim 1, wherein:

the step of generating the accelerated metal body is achieved by the steps of: positioning a piece of sheet metal between a consumable metal body and a stationary body; rapidly vaporizing the consumable metal body, directing vaporized metal into the piece of sheet metal; and accelerating the piece of sheet metal into collision with the stationary body at a rate sufficient to shear off a portion of the piece of sheet metal.

4. The method of claim 3, wherein:

the stationary body is a die and the piece of sheet metal is deformed by the collision to shear off the accelerated metal body therefrom.

5. The method of claim 3, wherein:

the consumable metal body comprises a foil of aluminum.

6. The method of claim 3, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

7. The method of claim 6, wherein

the high current is achieved by discharging a bank of capacitors.

8. The method of claim 1, wherein:

the step of generating the accelerated metal body is achieved by the steps of: positioning a piece of sheet metal between a consumable metal body and a stationary body; rapidly vaporizing the consumable metal body, directing vaporized metal into the piece of sheet metal; and accelerating the piece of sheet metal into collision with the stationary body at a rate sufficient to shear off a portion of the piece of sheet metal.

9. The method of claim 3, wherein:

the stationary body is a die and the piece of sheet metal is deformed by the collision to shear off the accelerated metal body therefrom.

10. The method of claim 4, wherein:

the consumable metal body comprises a foil of aluminum.

11. The method of claim 8, wherein:

the consumable metal body comprises a foil of aluminum.

12. The method of claim 9, wherein:

the consumable metal body comprises a foil of aluminum.

13. The method of claim 4, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

14. The method of claim 10, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

15. The method of claim 8, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

16. The method of claim 11, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

17. The method of claim 9, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.

18. The method of claim 12, wherein:

the rapid vaporizing is achieved by passing a high current rapidly into the consumable metal body.