In the system, however, after the 350 ~s risetime of the pulse to ~10 MA, either cables or cable adaptors experience failures at approximately 10C kv.

Transcription

1 PERFORMANCE OF THE LAGUNA PULSED POWER SYSTEM* J. H. Goforth, R. S. Caird, C. M. Fowler, A. E. Greene, H. W. Kruse, I. R. Lindemuth, H. Oona, and R. E. Reinovsky Los Alamos National Laboratory Los Alamos, NM 8755 Abstrat The goal of the LAGUNA experimental series of the Los Alamos National Laboratory TRAILMASTER program is to aelerate an annular aluminum plasma z-pinh to greater than one hundred kilojoules of implosion kineti energy. To aomplish this, an eletrial pulse )5.5 MA must be delivered to a 2 nh load in 1 s. The pulsed power system for these experiments onsists of a apaitor bank for initial energy storage, a helial explosive-driven magneti-flux ompression generator for the prime power supply and opening and losing swithes for power onditioning. While we have not yet ahieved our design goal of 15 MA delivered to the indutive store of the system, all major omponents have funtioned suessfully at the 1 MA level. Signifiant su.esses and some diffiulties experiened in these experiments are desribed. In the system, however, after the 35 s risetime of the pulse to 1 MA, either ables or able adaptors experiene failures at approximately 1C kv. Introdution We are urrently developing an explosive pulsed power system for the LAGUNA experiments of the Los Alamos National Laboratory TRAILMASTER program. The expetations for the LAGUNA experiments are presented by Greene et al.1 in another paper in this onferene. The pulsed power system makes use of an existing explosive pulsed power faility housing a 2.-MJ apaitor bank, 2 and ombines a previously tested helial explosive-driven magneti-flux ompressin generator 3 with a ylindrial explosively formed fuse for ulse ompression. A surfae disharge losing swith diverts urrent to the load. Fig. 1 is an artist's onept of the LAGUNA apparatus. To date, we have foused experiments on several aspets of the system. Initial experiments onentrated on maximizing the energy delivered to the helial generator from the apaitor bank, resulting in 73 kj of the 1.2 MJ available in one-half of our apaitor bank being oupled to the generator without leading to an internal eletro-mehanial failure. Subsequent generator tests asertained the urrent gain of our standard generator into a 1 nh dummy load, and one experiment was devoted to testing a higher urrentgain generator. With these generators in hand we onduted an initial high urrent test of the explosively formed fuse. The purpose of this test was to verify that the opening swith would behave as predited from the results of small sale tests. In this test detonator atuated losing swithes were used to divert urrent to a 1-nH stati dummy load at relatively low voltage. In the urrent series of sub-system tests we use surfae disharge losing swithes to divert urrent to exploding foil-fuse loads. These experiments are to verify the performane of the losing swith as part of this system and to subjet the remainder of the apparatus to high voltages. Although the major omponents have funtioned properly in these tests, we have not yet obtained satisfatory system performane due to failures in the able transition region. The ables have been subjeted separately to DC voltages of 17 kv and Marx Generator pulses of )5 kv without failure. *This work is performed under the auspies of the U.S. Department of Energy. 5 Figure 1. Artist's onept of a LAGUNA experiment. Cables to arry initial urrent from the apaitor bank to the generator protrude through a hole in the steel blast shield and onnet to the generator. On the output of the generator is a oaxial storage indutor with adaptor to attah transmission ables. The ables (3 in this piture) arry urrent to the vertially mounted pulse ompression/power flow/diagnostis hamber setion. The explosively formed fuse opening swith is not visible inside the lower part of the ylinder. Finally, the surfae disharge losing swith is depited by the gap in the outer transmission ylinder. Mark-IX Helial Generator Experiments Initial preditions for the performane of the LAGUNA pulsed power system were based on the availability of a large fration of the energy in our 2. MJ apaitor bank to provide the initial magneti energy in our helial generator. However, sine previous tests of our standard MK-IX helial generator were performed with only kj, we questioned the ability of the generator to funtion properly at suh high initial energies. As a result, for the initial test, only three of the four 2-kV, 3-mF modules were used to supply the initial generator urrent. 'lith this 2-kV, 9-mF apaitor bank, and a iruit dominated by the 7.2-H initial generator indutane, 1 we expeted to deliver as muh as 7 ka initial urrent to the generator with a risetime of s. The urve in Fig. 2 is the di/dt (i) waveform observed on the test as the initial urrent was fed into the generator from the apaitor bank. This urve should be a slightly damped osine wave, and the abrupt inrease seen at 19 s indiates a sudden derease in iruit indutane. Analysis indiates that suh a drop would be aused by shorting onseutive turns in the high-turns-density setion of the generator. The subsequent performane of the generator was also on-

2 Report Doumentation Page Form Approved OMB No Publi reporting burden for the olletion of information is estimated to average 1 hour per response, inluding the time for reviewing instrutions, searhing existing data soures, gathering and maintaining the data needed, and ompleting and reviewing the olletion of information. Send omments regarding this burden estimate or any other aspet of this olletion of information, inluding suggestions for reduing this burden, to Washington Headquarters Servies, Diretorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 12, Arlington VA Respondents should be aware that notwithstanding any other provision of law, no person shall be subjet to a penalty for failing to omply with a olletion of information if it does not display a urrently valid OMB ontrol number. 1. REPORT DATE JUN REPORT TYPE N/A 3. DATES COVERED -. TITLE AND SUBTITLE Performane Of The Laguna Pulsed Power System 5a. CONTRACT NUMBER 5b. GRANT NUMBER 5. PROGRAM ELEMENT NUMBER. AUTHOR(S) 5d. PROJECT NUMBER 5e. TASK NUMBER 5f. WORK UNIT NUMBER 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) Los Alamos National Laboratory Los Alamos, NM PERFORMING ORGANIZATION REPORT NUMBER 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 1. SPONSOR/MONITOR S ACRONYM(S) 12. DISTRIBUTION/AVAILABILITY STATEMENT Approved for publi release, distribution unlimited 11. SPONSOR/MONITOR S REPORT NUMBER(S) 13. SUPPLEMENTARY NOTES See also ADM IEEE Pulsed Power Conferene, Digest of Tehnial Papers , and Abstrats of the 213 IEEE International Conferene on Plasma Siene. Held in San Franiso, CA on 1-21 June 213. U.S. Government or Federal Purpose Rights Liense 1. ABSTRACT The goal of the LAGUNA experimental series of the Los Alamos National Laboratory TRAILMASTER program is to aelerate an annular aluminum plasma z-pinh to greater than one hundred kilojoules of implosion kineti energy. To aomplish this, an eletrial pulse )5.5 MA must be delivered to a 2 nh load in 1 s. The pulsed power system for these experiments onsists of a apaitor bank for initial energy storage, a helial explosive-driven magneti-flux ompression generator for the prime power supply and opening and losing swithes for power onditioning. While we have not yet ahieved our design goal of 15 MA delivered to the indutive store of the system, all major omponents have funtioned suessfully at the 1 MA level. Signifiant su.esses and some diffiulties experiened in these experiments are desribed. 15. SUBJECT TERMS 1. SECURITY CLASSIFICATION OF: 17. LIMITATION OF ABSTRACT SAR a. REPORT b. ABSTRACT. THIS PAGE 18. NUMBER OF PAGES 19a. NAME OF RESPONSIBLE PERSON Standard Form 298 (Rev. 8-98) Presribed by ANSI Std Z39-18

3 sistent with a generator having redued indutane, and whereas 15 MA had been our desired peak urrent, only 7 MA were observed. 3-1 [ Coil failure I! r Crowbar ,! i r r- J J J i! i I I I Time from bank start (t-ls) and s ), and these are retained in the Marx hookup. The tri-plate header shown in Fig. is the only hange from routine 2-module operation. At the tri-plate, the bottom plate is at the potential of the hassis of the low potential module and the middle plate is at the potential of the hassis of the high potential module. When swithes s 3 and s are losed, the high voltage side of the low potential module is onneted to the enter plate, and the top plate is onneted to the high voltage side of the high potential module. A total of kv exists between the top and bottom plates, and to suppress edge traking with this voltage, the tri-plate is bathed in 1 atm SF Figure 2. i from initial MK-IX generator test. The generator oil failure is indiated, as is the time when the generator rowbars at its input eliminating the apaitor bank from the iruit. The internal failure of the MK-IX generator ourred before flux ompression started. This problem led us to dramatially redesign the system for providing initial urrent to these high indutane generators. Our first onern was to redue the time sale of the urrent feed as muh as possible. To do this, we arranged two of the modules of our apaitor bank as a Marx generator as shown shematially in Fig. 3. This provided the energy of one-half the bank (1.2 MJ) in a onfiguration with the lumped parameters of 1.5 mf at kv. Sine the indutane of the generator is still dominant in the iruit, the primary effet was to redue the apaitane by a fator of six, and hene the urrent risetime by a fator of 2.. After onsidering damping fators, we hose to atuate the generator rowbar 13 s after initial urrent rise. L(t) -:J_ s Figure 3. Ciruit for putting two standard apaitor bank modules in series. Swithes s 1 and s 2 are ground isolation swithes that open when the bank is harged during both standard and Marx operation of the bank. Swithes s 3 and s are the same detonator atuated losing swithes used during routine operation. A ompat header for arranging the two apaitor modules in series was devised and is pitured in Fig.. In normal operation, eah bank module has independently triggered detonator atuated swithes (S 3 L Figure, Tri-plate header providing series iruit for two apaitor bank modules. The standard detonator atuated bank swithes (top and bottom on the right) have able outputs that attah to the header. Ihe ground braids from the lower right hand swith attah to the bottom plate, and the enter ondutor to the enter plate. Ihe ground braids from the upper swith attah to the enter plate and the enter ondutors to the upper plate. The output ables disappear into the shadow ast by the steel blast shield, and withstand up to kv when the detonator swithes lose. During operation, the header is enlosed in a plasti bag and bathed in SF Figure 5 shows i during the initial urrent phase of a MK-IX generator test using the Marx onfiguration. This test was performed with a total Marx voltage of 3 kv, and integration of the urve in Fig. 5 reveals that 5 ka are delivered to the generator on these tests. This is suffiient initial urrent for interesting experiments, and the generator survives this pulse with no eletro-mehanial failures. With this initial urrent, a standard MK-IX generator will drive 11 MA into a 1-nH pure indutive load. Into the omplete LAGUNA pulsed power system, the K-IX has delivered 9.5- to 1.5-MA, and the urrent and its respetive i are shown in Fig. for one of these tests. A omplete disussion of the struture observed on the i urve is beyond the sope of this paper, but we note in passing that the distint humps in the waveform are due to bifurations in the generator windings. Although one extra test was onsumed in this series in developing a dummy load that would withstand the fores of the pulse shown in Fig., all tests with the Marx onfiguration have led to experiments with adequate generator performane. On another test, the standard MK-IX helial winding pattern was hanged to produe an 11 H generator with fewer bifurations. This generator was suessfully tested and, as hoped, ahieved a higher urrent gain into the 1-nH dummy load. The standard

4 MK-IX produes a urrent gain of 2 into a 1-nH load and the 11-H generator produed a total gain of 31 in the one test. Unfortunately, the extra generator indutane redued the initial apaitor bank urrent to 35 ka, and a peak urrent of 11 MA was still observed. Although more urrent was not generated on this test, it does serve to illustrate that the MK-IX generator is not limited to 11 MA by urrent density onsiderations. We infer this beause the higher gain generator produed an equivalent 11 MA output with only one half the number of parallel windings in the helix i I!I ' ' 'i.i i i i !! i i : : : : J !! j : ::1:: II: :1:-!!!!! 25 5o Time (p,s) Figure 5. i from MK-IX test using Marx Header. While we ontinue to examine various possibilities for generating urrents of 15 MA for future experiments, the 1 MA urrents available are interesting for further system development tests, and we have proeeded with these. 1 the explosively formed fuse (EFF) whih is further disussed in another paper in this onferene. The EFF has previously been shown to ondut a few megamperes and to dissipate 5 kj eletrial energy in the urrent transfer proess. To interrupt 15 MA flowing in a 35-nH opening swith and to divert the urrent in the storage indutor to a 2-nH load, MJ must be dissipated. Our first pulse ompression test was to ouple a MK-IX generator to an EFF enlarged to meet this need and verify that the swith ould ondut the 35-s pulse, then open with harateristis similar to those projeted from small sale experiments. Detonator atuated losing swithes were used to introdue a 1-nH stati indutor in parallel with the EFF as it began to open. The swithes were atuated before the EFF had developed appreiable voltage, and the only voltage generated in the test is the Li voltage that the swith generates aross 1 nh. At 1-MA swith urrent with a 35-nH swith and a 1-nH load, 2.2 MJ must be dissipated in the swith. Figure 7 shows some of the important waveforms from the test. (These and additional urves are also shown in Ref. in this onferene.) After the urrent transfer operation, diagnostis showed that the load urrent was 2 MA less than in the storage indutor and this differene has not been aounted for. Nevertheless, the resistane profile determined from this shot was in good agreement with values projeted from small sale tests, and the swith had suessfully onduted 9.5 MA with a risetime of 35 s, then interrupted it in less than 2 s, dissipating 1.8 MJ. This test onfirmed the basi swith performane '< a... <:) dl!dt..1 'W o.o u....s2 e Tme (p,s) Figure 7. Explosively formed fuse urrent (A), load urrent (B), and Li voltage (C) sustained aross a 1-nH fixed indutane load. Approximately 2 MA were lost to an apparent transmission line failure. 1 2 Tme (p,s) Figure. I and i for MK-IX test with a omplete LAGUNA pulsed power system. Opening Swith and Current Transfer Tests With generators delivering 1-MA to 1-nH loads in hand, we began tests of the omplete pulse ompression sheme. The prinipal omponent of this sheme is 7 The next step was to investigate the performane of the surfae disharge losing swith and to subjet the system to voltages harateristi of a dynami z-pinh load. To do this, a foil-fuse load with 25-nH of indutane was used in plae of the 1-nH stati load. In addition, a surfae disharge losing swith was used to introdue the fuse load into the iruit when the opening swith had developed 1 kv. On this test the system appeared to funtion 2roperly in every respet until the 1 kv neessary to lose the surfae disharge swith were generated. At this time, or only slightly later a disrepany began t J appear among the 1 reords. 1 measured in the storage indutor should equal the sum of the i's measured in the opening swith

5 and the fuse load. After swith losure this summation failed, indiating that a failure had ourred in the ables or able adapting hardware onneting the storage indutor to the swith/load assembly. Only 1 MA was delivered to the fuse. After examining the data from the first fuse-load test, we felt that the weakest part of the interonneting able system was the termination region where the ables onneted to the base of the explosively formed fuse. Unonstrained, the ables would move a few millimeters during the urrent pulse, and the onstraints that we ould bring to bear were neither extremely massive nor strong. To alleviate this situation, the number of ables was inreased from 3 to 72, and all were positioned at an inreased radius. Both fators redue the stress imposed on the ables. >5 kv with a fast Marx generator, and have DC hi-potted the ables prior to the test to 17 kv without failure. We are urrently trying to determine the preise failure mehanism for the last of our tests, and we will proeed to LAGUNA z-pinh experiments when a transmission setion that will withstand the transfer of a urrent pulse to our load has been developed. HIGH EXPLOSIVE LOAD INSULATION Figure 9. LAGUNA pulsed power apparatus with 72-able transition and fuse load. Figure 8. Explosively Formed Fuse with 72- able transition setion. As shown, the upper oax is intended to onnet to a LAGUNA.load. The able adaptor region is filled with transformer oil. Figure 8 shows the EFF and its able input setion in the 72-able arrangement. With this arrangement, the fores would be approximately the same at 1 MA that they were on the largely suessful initial system test at 9.5 MA. Figure 9 is a photograph of one of these omplete assemblies on the firing pad. We have tested two suh assemblies, and have yet to transfer the proper fration of the urrent to the fuse load. In the first 72-able test, 3 of 9 MA were transferred to the fuse, and the fuse burst at a substantially lower than design urrent and produed only a very small voltage spike. In the seond of these tests, modifiations ditated by results of the first were inorporated, but a failure still resulted. Four of the 5 MA should have been transferred to the fuse, but only 2 MA were. The fator of two redution in urrent undersores the magnitude of our problem. This level of urrent translates to one-fourth the stress on the ables, and approahes one-tenth the stress on the ables in the initial test. The experiment nevertheless failed at the 1-kV level set for the surfae di'sharge swith to lose, just as it did in the prior two tests. We have pulse tested these ables to 8 Aknowledgements The authors wish to aknowledge the important ontributions of a large number of people and sinerely hope that the reader will turn to Ref. (in these proeedings) where we have devoted the spae to do this properly. Referenes (1] A. E. Greene, J. H. Brownell, R. S. Caird, J. H. Goforth, R. H. Prie, and R. J. Trainor, "Expetations for the LAGUNA Foil Implosion Experiments," To be presented this onferene. [2] D. J. Erikson, B. L. Freeman, J. E. Vorthman, R. S. Caird, C. M. Fowler, J. C. King, A. R. Martinez, J. v. Repa, J. B. VanMarter, and R. G. Vaughn, "Firing Complex for Explosive Pulsed Power," Fourth International Conferene on Megagauss Magneti Field Generation and Related Topis, C. M. Fowler, D. J. Erikson, and R. S. Caird, Eds., Plenum Press, (1987) to be published. [3] C. M. Fowler, R. s. Caird, B. L. Freeman, and J. B. VanMarter, "Performane of the Mark IX Helial Flux Compression Generator," M- Tehnial Note, Deember 1983, Los Alamos National Laboratory. [] J. H. Goforth, R. S. Caird, A. E. Greene, I. R. Lindemuth, S. P. Marsh, and H. Oona, "Advanes in Explosively Formed Fuse Opening Swithes," to be published this onferene. [5] R. E. Reinovsky, J. H. Goforth, A. E. Greene, and J. Graham, "Charaterization of Surfae Disharge Swithes and High Power System Appliations," to be presented this onferene.