Experimental study of the structure of the plasma current sheath on the PF-1000 facility PDF
Preview Experimental study of the structure of the plasma current sheath on the PF-1000 facility
Home Search Collections Journals About Contact us My IOPscience Experimental study of the structure of the plasma-current sheath on the PF-1000 facility This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2012 Plasma Phys. Control. Fusion 54 025010 (http://iopscience.iop.org/0741-3335/54/2/025010) View the table of contents for this issue, or go to the journal homepage for more Download details: IP Address: 194.67.73.233 The article was downloaded on 20/01/2012 at 11:21 Please note that terms and conditions apply. IOPPUBLISHING PLASMAPHYSICSANDCONTROLLEDFUSION PlasmaPhys.Control.Fusion54(2012)025010(14pp) doi:10.1088/0741-3335/54/2/025010 Experimental study of the structure of the plasma-current sheath on the PF-1000 facility VKrauz1,KMitrofanov2,MScholz3,MPaduch3,LKarpinski3, EZielinska3 andPKubes4 1NRC‘KurchatovInstitute’,Moscow,Russia 2SRCRFTRINITI,Troitsk,MoscowOblast,Russia 3IPPLM,Warsaw,Poland 4CVUT,Prague,CzechRepublic E-mail: krauz@nfi.kiae.ru Received28September2011,infinalform3December2011 Published19January2012 Onlineatstacks.iop.org/PPCF/54/025010 Abstract Theresultsofstudiesoftheplasma-currentsheathstructureonthePF-1000facilityinthestage closetotheinstantofpinchformationarepresented. Themeasurementswereperformedusing variousmodificationsofthecalibratedmagneticprobes. Studiesoftheinfluenceoftheprobe shapeanddimensionsonthemeasurementsaccuracyweredone. Thecurrentflowinginthe convergingsheathatadistanceof40mmfromtheaxisofthefacilityelectrodeswasmeasured. Intheoptimaloperatingmodes,thiscurrentisequaltothetotaldischargecurrent,which indicatesthehighefficiencyofcurrenttransportationtowardtheaxis. Insuchshotsacompact high-qualitysheathformswithshockwaveinfrontofthemagneticpiston. Itisshownthatthe neutronyielddependsonthecurrentcompressedontotheaxis. Thisdependenceagreeswell withtheknownscaling,Y ∼I4. Theuseofthetotaldischargecurrentinconstructingthe n currentscaling,especiallyforfacilitieswithalargestoredenergy,isunjustified. (Somefiguresmayappearincolouronlyintheonlinejournal) 1. Introduction capacitor bank exceeded a certain critical value, the scaling in terms of energy was violated and the neutron yield was It is well known that, in plasma focus (PF) systems, the saturatedatalevelof∼1011–1012 neutrons/shot. Thenature intensity of ionizing radiation emitted from the pinch region of this effect is one of the key subjects of study in PF depends strongly on the pinch current I. In particular, systems; in particular, it is of great importance to find out the experimental neutron scaling based on measurements whethersaturationisrelatedtonaturalphysicalconstraintsor performed at facilities with a stored energy, W, from several determinedbydisadvantagesofthefacilitydesign[13]. units of kilojoules to a few hundred kilojoules has the form We believe that saturation is not related to any physical Y ∼ I3.2–5 [1–3]. For this energy range the scaling agrees constraints. At least, on the Z-facility (Sandia National n wellwithanotherwell-knownexperimentalscalingY ∼W2. Laboratories),aneutronyieldof∼5×1013D–Dneutronsper n Suchdependence,slightlydivergentinpowerfrom2.1to2.5, shotwasachieved, however, ataveryhighdischargecurrent was found on many devices in the early stages of PF-studies of17MA[14]. Althoughtheabovescalingwasnotsatisfied [4–6]. In the beginning of the 1970s H Rapp analyzed the inthiscase,theseexperimentsdemonstratedthefundamental results obtained on more than 10 facilities in the world, and possibilityofobtaininghighneutronyieldsatrealisticcurrents showed that this scaling has a general character [7]. Later inZ-pinchsystems. this scaling has been expanded to sub-kJ machines [8,9]. The problem of achieving high discharge currents in PF However, experiments carried out on megajoule facilities systemsisbeingactivelydiscussed. Inparticular,in[15]itwas [10–12] demonstrated that, when the energy stored in the assertedthatthefurtherincreaseinthecurrentamplitudecan 0741-3335/12/025010+14$33.00 1 ©2012IOPPublishingLtd PrintedintheUK&theUSA PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal be provided only by increasing the initial discharge voltage, classical ‘shock wave-magnetic piston’ model. In particular, whichinevitablyleadstoagradualconversionofPFsystems magnetic probe measurements often demonstrate a two-peak tofastZ-pinchconfigurationsandthelossofoneofthemain PCSstructure(see,e.g.,[25]). Three-fluidMHDsimulations advantagesofPFsystems,namely,thesimplicityofdesign. In of the Z-pinch PCS [26] have shown that, since the plasma contrast,calculations,madeintheframeworkoftheLeemodel temperatureand,accordingly,plasmaconductivityinthePCS [16–18], show that there was no saturation in pinch current intheinitialstageofthedischargearelow,themagneticfield and Y : they continue to rise with condenser bank capacity diffusesintotheshockwave,whichleadstotheformationof n C althoughataslowerrate. Butthesecalculations,madefor a current density peak ahead of the shock front. In [27], it 0 PF-1000 parameters, show also, that, in spite of high power wassupposedthat,inthestageofdischargechambertraining inneutronscaling(Y ∼ I4.7 andY ∼ I3.9 )[17],theD–D (degassing),asignificantfractionofthecurrentcanflowinthe n pinch n peak neutron yield Y ∼ 1013 can be achieved only at W close to shockwaveandthecurrentisredistributedintothemagnetic n 19MJ[16]. piston only when the facility reaches its standard operating In this study, we do not consider this problem in detail, mode. In[28],itwasdemonstratedbymeansofsynchronized because it obviously requires more thorough examination. high-speed photography and magnetic probe measurements There is another (no less important) problem that manifests that, in the standard operating mode, more than 85% of the itselfalreadyatcurrentsachievableintheexistingfacilities— currentflowsinthemagneticpiston. Thus,thePCSstructure the problem of the transportation efficiency of the total depends substantially on the discharge mode. Systematic discharge current toward the system axis, where the pinch measurements of the PCS structure, however, have not been forms. The saturation effect can manifest itself if a fraction performedasyet. ofthetotalPFcurrentflowsbeyondthepinch. Measurements Inthispaper,wepresenttheresultofmeasurementsofthe of the current distribution performed on the megajoule PF PCScurrentinthestageclosetotheinstantofpinchformation facility at Frascati showed that a fraction of the total current on the PF-1000 facility [29] operating with deuterium at flowed in the residual plasma near the PF insulator and did an energy stored in the capacitor bank of 250–500kJ. The notcontributetotheprocessofpinchcompression[19]. Thus, measurements were performed using absolutely calibrated despiteinfringementoftheneutronscalingintermsofenergy, magneticprobes. ThePCSstructurewasstudiedusingoriginal scalingintermsofthecurrentflowinginthesheathcontinued magneto-optical probes simultaneously recording the dB/dt tobevalid. Itbecameclearthatthecurrentinthesheath(or signal and the plasma optical glow. Along with magneto- thepinchcurrent)sofardoesnotalwayscoincidewithatotal opticalprobes, aRogowskicoilrecordingthetotaldischarge dischargecurrent. In[20]thedistinctionofthesecurrentsfor currentandthreemagneticprobeslocatedinsidethePF-1000 variousinstallationsisestimated. Itwasconcludedthat(i)the collector were also used. The main objective of this paper relevantparameterofneutronproductionisthesheathcurrent wastostudythedependenceoftheneutronyieldonthetotal I and not total current I and (ii) all effectively operated discharge current and the current flowing in the PCS in the sh t facilities at a comparative low level of input energy have the finalstageofradialplasmacompression. relationI /I closeto1. sh t Atthesametimetheresultsreportedin[19]andobtained 2. Designofmagneticprobesandschemeofthe on very fast high-voltage facility SPEED2 [21] have shown experiment that this problem becomes more challenging with increasing discharge energy. The same situation was observed on the Magnetic probe measurements are widely used to study the PF-3facility(KurchatovInstitute,Moscow),whereabsolutely dynamicsandstructureofthePCSinPFdischarges[19,22– calibratedmagneticprobeswereusedtomeasurethecurrent 25,27,28,30–35]. Inmoststudies,however,thismethodwas distribution in PF experiments with noble gases. The results usedtoinvestigatethePCSbehaviorintheearlystageofthe ofprobemeasurementsshowedthat,inthePF-3experiments discharge in the regions located far from the pinching zone, carried out with neon, only a fraction of the total discharge because the magnetic probe diagnostics is a contact method current contributed to the process of pinch compression and questions always arise concerning the influence of the [22–24]. detectorontheplasmacharacteristicsandaccuracyofmagnetic Theefficiencyofcurrenttransportationtowardthesystem field measurements. Magnetic probes introduced directly in axisdependssubstantiallyontheefficiencyofthe‘snowplow’ thepinchformationregionmaydisturbthepinchingprocess. mechanism. Poor snowplowing of the neutral gas by the Moreover,intenseplasmaandradiationfluxescandestroythe magnetic piston in the outward expansion and the ‘runaway’ probe during one discharge, which substantially complicates stages can result in the high density of the residual plasma experiments. behind the plasma-current sheath (PCS) and, accordingly, Taking into account the aforesaid, modified absolutely in substantial current leakages. The spatial distribution of calibratedmagneticprobesforinvestigatingthePCSdynamics the plasma density and current in the final stage of plasma underthePFdischargeconditionsonthePF-1000facilitywere compression should substantially affect plasma stability and designed and manufactured. The new probes were designed theprocessesresponsibleformagneticenergydissipationand on the basis of magnetic probes developed earlier for the radiationgeneration. Therefore, studiesofthePCSstructure experiments carried out on the Angara-5-1 facility [36,37]. are of great importance. The actual current distribution in Indesigningtheseprobes,wetookintoaccountmanyfactors the PCS can differ substantially from that predicted by the limiting the application of magnetic probes in high-power 2 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal Figure1. Sensitiveelementsof(a)theplaneprobeand(b)themagneto-opticalprobe. discharges, such as the evaporation of the probe case under The probes were calibrated using a homogeneous ac the action of radiation and particle fluxes, screening of the magneticfieldwithaknownamplitudeandknownfrequency probebytheplasma,andplasmaperturbationsarisingdueto by the method described in [37]. The probe sensitivity was plasmaflowaroundtheprobe[38]. Toimproveconditionsfor ∼(1.0–5.0)×10−11VG−1s−1. Thecalibrationaccuracywas plasmaflowaroundtheprobeandminimizetheinfluenceofthe betterthan5%,andtheaccuracyindeterminingthemagnetic probeontheprocessesunderstudy,thesensitiveelementofthe induction in plasma with account of calibration was about probewasmadeplanarwithatransversesizeof0.7–0.8mm. 15–20%. Application of such probes on the Angara-5-1 facility for ThePCSstructurewasstudiedusingspeciallydeveloped measurements of magnetic fields during the implosion of magneto-optical probes. Along with the time derivative of 20mm diameter cylindrical wire arrays demonstrated their the azimuthal magnetic field, these probes also recorded the efficiency [39,40]. Later on, these probes were successfully plasma optical glow at their location. Similar probes were adapted to the experimental conditions of the PF-3 plasma earlier used to study the process of shock wave formation focusfacility[22–24]. during the compression of a transverse magnetic field by a The probes designed for measurements of the azimuthal supersonic(M > 3)plasmaflow[41,42]. Lateron, sucha A magneticfielddistributioninthestageofPCScompressionin probe was used to investigate the PCS structure in the early the PF-1000 facility consist of two ∼300µm-diameter coils stageofradialplasmacompressioninthePF-3facility[43]. wound in clockwise and counterclockwise directions. The The design of the magneto-optical probe is shown in design of the sensitive element of the probe is shown in figure 1(b). The shape of the sensitive element of the probe figure 1(a). The coils are packed in a common plane screen wasmodified:theplaneheadwasreplacedbya2mmdiameter madeofaniobium-titaniumfoilwithathicknessof10–15µm, ceramictube. Twochannelsforrecordingthemagneticfield which is less than the skin depth for the characteristic times weresupplementedwitha0.38mmdiameterpolymeroptical of the processes under study. The time of electromagnetic fiber placed in the common case with the magnetic coils. A fielddiffusionthroughsuchascreenwascalculatedinMHD shortsegmentoftheopticalfiber(∼1mm)protrudedfromthe approximation for supersonic plasma flow with a frozen-in protectingceramictubeintothedischarge. magnetic field and for parameters of the Angara 5-1-facility The second end of the optical fiber was connected to which is closed to parameters of PF-1000: n = 5×1017– e a photomultiplier tube (PMT). In processing the recorded 1019cm−3,plasmavelocity106–5×107cms−1,magneticfield signals, all the cable time delays, the propagation time of 0–100kG,T =1–30eV[37]. Forthetotalsizeofthesensitive photonsalongtheopticalfiber,andthePMTelectrontransition element of the probe in the direction perpendicular to the time(ETT)weretakenintoaccount. Themagneticchannels incidentplasmaflow0.7–0.8mm,thecalculateddiffusiontime of these probes were also calibrated by the abovementioned is∼1.5ns. Twosymmetricsignalsofdifferentpolarityfrom method[37]. two coils allow one to be sure of their magnetic origin. At TheexperimentswereperformedonthePF-1000Mather- the instant of probe destruction, the symmetry of the signals type facility. The large-scale PF-1000 facility (figure 2) violates. consistsofthefollowingmainunits: Eachprobemeasuredthetimederivativeoftheazimuthal magnetic field dB (r,ϕ)/dt at its location. To determine • The condenser bank and pulsed electrical power circuit ϕ the current flowing within the radius at which the probe was withacollectorandlow-inductancecables. installed, the signal was integrated numerically under the • The mechanical vacuum and gas system consist of the assumptionthattheplasmacurrentwassymmetricwithrespect vacuum chamber, coaxial electrodes and gas handling tothesystemaxis. system. 3 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal the recording system. To prevent high-voltage breakdowns ontotheprobe,ameasurementschemewasusedinwhichthe entire ‘probe-oscilloscope’ measurement channel was under theanodepotential. In order to monitor the discharge quality, along with magnetic probe measurements, the electrotechnical and emissioncharacteristicsofthesystemweremeasuredineach shotusingstandarddiagnosticsappliedonthePF-1000facility. Thesediagnosticsareasfollows. Neutron emission was recorded using four integral activation detectors. The integral neutron yield in these experiments was the main parameter characterizing the discharge quality. In combination with magnetic probe measurements,itallowsonetoanalyzetheneutronscalingin termsofthedischargecurrent. FourdetectorsbasedonSTS-5 and STS-6 counters covered with a silver foil and placed in a neutron moderator were used to increase the measurement Figure2. GeneralviewofthePF-1000facility. accuracy. The detectors were calibrated in their standard positionsinthepreviousPF-1000experiments. Theelectricalenergyistransferredtoacollectorandelectrodes For time-resolved measurements of radiation we used by means of low-inductance cables. The vacuum chamber, a scintillator detector (SD) based on PMT with a plastic whichsurroundstheelectrodes,hasalargevolume(1400mm scintillatorinstalledatadistanceof7mfromthepinch(neutron indiameterand2500mminlength). Twocoaxialelectrodes andhardx-rayradiation)andp–i–nsilicondiodecoveredby are shown on figure 3. The outer electrode (OE) (cathode) filter10µmBeandinstalledatadistance0.8mfromthepinch consists of 12 stainless steel rods with 80mm in diameter. (softx-rayswithenergyabove600eV). The OE and copper center electrode (CE) radii are 200mm Theinterelectrodevoltagewasmeasuredbyacapacitive and115.5mm, respectively, withCElengthof460mm. The dividerwithasensitivityof4kV/V. cylindricalaluminainsulatorsitsontheCEandthemainpart The total discharge current was recorded using a oftheinsulatorextends85mmalongtheCEintothevacuum calibratedRogowskicoilwithasensitivityof470kAV−1. chamber. The insulator prescribes the shape of the initial The time derivative of the total discharge current was current sheet between the CE and the back plate of the OE. measured using three single-loop magnetic detectors located Thecondenserbankofcapacitance1332µFcanbechargedto near the facility collector. The detectors were arranged voltagerangingbetween20and40kV,whichcorrespondedto with a step of 120◦ along the collector perimeter, which dischargeenergiesrangingfrom266to1064kJ. made it possible to estimate the current homogeneity at the During the described experiment, several shots with collector. the PF-1000 upgraded current generator and a new alumina The PCS dynamics and structure were also investigated insulatorweremade. using a 16-channel laser interferometer with a 1ns time As was mentioned above, a probe located near the resolutionandatimeintervalof10–20betweenframes[44]. system axis is subject to intense radiation and particle fluxes and, therefore, is destroyed during one shot. Before each subsequentdischarge,itisnecessarytoreplacethedestroyed 3. Studyoftheperturbationarisingduetoplasma probe with a new one. For this purpose, a vacuum lock was flowaroundthemagneticprobe createdtointroducetheprobeintheaxialregionofthesystem withoutviolatingvacuumconditionsinthedischargechamber Themagneticprobediagnosticsofplasmaisacontactmethod, of the facility. This is necessary for the initial discharge which assumes the plasma impact on the probe and vice conditions(thegaspressure,theconcentrationsofadmixtures versa. In other words, we are dealing with the plasma intheworkinggas,etc)toremainunchanged,becausethePF perturbationintroducedbytheprobe. Inthiscase,theproblem discharge is very sensitive to these conditions. The probes offlowaroundaprobebyamagnetizedplasmaisofprimary were introduced from the collector side through the vacuum importance. The accuracy and time resolution of magnetic lock along the axis of the hollow anode. The system design probe measurements depend on whether the magnetic field allowedonetoplacethesensitiveelementoftheprobenearthe has enough time to diffuse from the plasma into the probe systemaxis,namely,neartheanodesurfaceatradiiof1.2and case. The situation is especially complicated in the case of 4cm and at a distance of 0.5–1.5cm from the surface of the supersonicflowaroundtheprobebyplasmawithafrozen-in endflangeofthecentralelectrode(anode). Thearrangement magneticfield,accompaniedbytheformationofashockwave. oftheprobesinthedischargechamberisshowninfigure3. The question naturally arises concerning the characteristic It should be noted that, in these experiments, the anode scale length and relaxation time of the induced plasma was at a high voltage. In this case, high-voltage breakdown perturbation. can occur between the anode and the probe, which can lead Thepapersdevotedtostudyingtheproblemofsupersonic not only to the probe destruction, but also to a damage of flowaroundbodiesofdifferentshapesbyamagnetizedplasma 4 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal Figure3. GeometryoftheelectrodesofthePF-1000facilityandarrangementofthemagneticprobes: (1)anode,(2)cathode(12rods), (3)insulator,(4)vacuumlockthoughwhichtheprobeisintroduced,and(5,6)magneticprobesmeasuringtheazimuthalmagneticfieldat theradiiof12mmand40mm,respectively. Probe(5)isnotshownonthephoto. in high-current discharges are few in number [38,45,46]. theprobeandmakeanappreciableimpactontheaccuracyof In our experiments, we used two probe modifications with a themeasurementofthemagneticfieldsandthedegreeofthis plane(0.5×3mm)caseandacylindricalcase(withadiameter influencestronglydependsontheprobeshape. of 2.0–2.5mm). The perturbations introduced by the probes The influence of the plasma flow with a frozen-in werestudiedusinglaserinterferometry. magnetic field on the probe signal for probes of different Figures 4 and 5 show interferograms obtained in shapes was numerically calculated in [37]. It follows from dischargeswithaplaneprobe(shotno9120)andacylindrical thesecalculationsthatmagneticfieldperturbationsintroduced probe (shot no. 9127) for the same initial conditions (D , by a plane probe are about 7%. It was found that the 2 P = 1.8Torr, U = 24kV and W = 384kJ). Figure 4(a) measurementerrorofprobeswithaplanescreenislessthan 0 0 0 corresponds to the time at which the PCS passes the probe. that of probes with a conventional cylindrical screen. For a Afterpassingtheplaneprobe,thePCSperturbationoccupiesa cylindricalprobe,1mmindiameter,thiserrorisalready25%. narrowregion∼1mmwidthalongtheradius(figure4(b)),and This result was confirmed experimentally on the Angara-5-1 40nslater(figure4(c)),theperturbationcompletelyvanishes. facility[37]. Theplasmaperturbationintroducedbyacylindricalprobe It is natural to assume that the error increases with the is wider, ∼4mm, which is about 1.6 of the probe size (see increase in the probe size. We have undertaken comparative figure5). AlthoughthePCSperturbationinthiscaseislarger studiesofprobesofadifferentconfigurationsonthePF-1000 than that introduced by the plane probe, it rapidly decays as facility. In figure 7 the result of current measurements with the PCS propagates toward the system axis. An even larger thedifferentprobesobtainedinthedischargeswithpractically PCSperturbationisobservedinexperimentsperformedwith thesamevalueofthetotaldischargecurrentandtheneutron a cylindrical probe at a higher gas pressure (P = 2.4Torr, output is shown. One can see that the current measured 0 figure 6). The size of the perturbed region behind the PCS with the cylindrical probe is essentially less. The typical just after it has passed the probe is ∼6mm (figure 6(a)), valueofcurrentamplitudeunderstatingunderourconditions which is about 2.5 of the initial probe size. After 30ns, makes ∼1.75. The measurement error decreases during the the size of the perturbed region behind the PCS decreases to pulseandatthecertaininstant,theindicationsoftwoprobes ∼2.5mm(figure6(b))and,afteranother40ns,theperturbation start to coincide. It is caused by the gradual disintegration practicallyvanishes(figure6(c)). Inthiscase,anunperturbed of the plasma surrounding a probe, and shielding reduction. pinchisobservedontheaxis(seefigure6(d)). It should be noted that, in similar experiments on the FF-3 Itshouldbenotedthattheradialprobesizewithaccount facility[43],suchastronginfluenceoftheprobedimensionon oftheprojectionontotheobservationdirectionis∼3.5–4mm, theaccuracyofmeasurementswasnotrevealed. Apparently, which is somewhat larger than the initial probe size. This it is caused by the strong influence of the parameters of the can be related to the residual PCS plasma and the plasma plasmasurroundingtheprobe:themeasurementsonthePF-3 formed on the probe surface under the action of radiation in weredoneonthelargedistancesfromtheaxisunderconditions thefinalstageofplasmacompression. Thisplasmacanshield oflessdenseandlesshotplasma. 5 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal Figure4. PCSinterferogramsrecordednearthesystemaxisinshotno9120(D ,P =1.8Torr,U =24kVandW =384kJ)withplane 2 0 0 0 probeatdifferentinstantsfromthebeginningofthedischarge:(a)5824,(b)5844and(c)5884ns. TheneutronyieldisY ∼1.3×1011 n neutrons/shot. Thefollowingimportantconclusionscanbedrawnfrom sessionswithdifferentinitialconditions. Takingintoaccount theseresults: that the number of shots was limited by the number of available probes, the first experiments were performed in a (i) In the case of a plane probe, the PCS perturbation is well-reproduciblemodewithdeuteriumataninitialpressure much smaller than in the case of a cylindrical probe. of P = 3Torr and a discharge voltage of U = 27kV. It is necessary to take into account at performing the 0 0 In these experiments, the stored energy was about 485kJ. quantitativemeasurements. Beforeeachshotwiththemagneticprobe,aseriesoftraining (ii) In spite of a relatively large plasma perturbation discharges under standard conditions were performed until introducedbyacylindricalprobe,itrapidlyrelaxesasthe the neutron yield became higher than 1010 neutrons/shot. PCS propagates toward the axis, so that no perturbation Since the probes were destroyed in each shot, the working introducedbytheprobeisobservedneartheaxis. gas in the chamber was usually refreshed before the next (iii) Theprobeparametersdonotrenderanessentialinfluence experiment. onitstimecharacteristics. Itallowsusingtheprobesofthe Figure 8 shows the results of measurements with a cylindricalform,forexample,themagneto-opticalprobes magneticprobelocatedatadistanceof4cmfromthesystem describedabove,forstudiesofthePCSstructure. axisinadischargewithahighneutronyield. Inthisdischarge, theneutronyieldwas1.24×1011neutrons/shot,whichisone 4. Experimentalresults ofthebestresultsobtainedinthisexperimentalsession. Thus, itcanbeassertedthat,inthegivenexperimentalscheme,the 4.1. Studyoftheazimuthalmagneticfield probelocatedinthePCScompressionregionatadistanceof The dynamics of the azimuthal magnetic field during PCS 4cmfromtheaxisinsignificantlyaffectsthepinchingprocess radial compression was investigated in four experimental and,accordingly,theneutronyield. Itfollowsfromfigure8(a) 6 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal Figure5. PCSinterferogramsrecordednearthesystemaxisinshotno9127(D ,P =1.8Torr,U =24kVandW =384kJ)with 2 0 0 0 cylindricalprobeatdifferentinstantsfromthebeginningofthedischarge: (a)5793and(b)5913ns. TheneutronyieldisY ∼1.7×1011 n neutrons/shot. thatthesignalfromthemagneticprobeappearsabout6.2µs A more abrupt drop in the signal from the Rogowski coil at after the beginning of the discharge. As a rule, it has one or the instant of dip (a peak in the time derivative of the total twomaxima,whichindicatesthatthecurrentdistributioninthe current) can be attributed to the partial differentiation of this PCSisrathercomplicated. Inparticular,inthisshottwopeaks signal. were recorded, the full width at half-maximum (FWHM) of Undertheconditionsinwhichthetotaldischargecurrent eachpeakbeing∼35–40ns. Aprecursorintheformofasmall- varies insignificantly while the PCS passes the probe, the amplitudepulsebeforethemainsignalwasalsoobserved. The signalmaximumcorrespondstotheinstantatwhichtheentire presenceofsuchaprecursor, whichcancarry∼5–7%ofthe PCS current begins to flow inside the 4cm radius region. total PCS current, is typical of many our discharges and its After this, until the probe is destroyed, the behavior of the originisunclear. currentmeasuredbytheprobecoincideswiththatofthetotal The signals from two different probe coils remain current,i.e.themeasuredcurrentdecreasesinaccordancewith practically symmetric over ∼200ns, which confirms their the reduction in the total discharge current. An important magnetic nature. Later on, the symmetry of these signals is consequence of this result is that the discharge current is abruptlyviolated(seetheverticaldashedlineinfigure8(a)), efficientlycompressedtowardtheaxis(atleasttoaradiusof which indicates that at least one coil is destroyed. After this 4cm),whichresultsinthehighvalueofneutronyieldinthis instant, the signals were not processed. However, even this discharge. timeintervalisquitesufficienttodeterminethecurrentflowing At the instant of dip (at the peak of the total current inthePCS.Thiscurrentwasrecoveredundertheassumption derivative), an abrupt increase in the voltage up to 50kV is thatthePCSwasaxisymmetric. observedinthevoltagewaveform. Itcanbeseenthat,over∼100ns,thePCScurrentreaches Figure 9 shows the results of similar measurements itsmaximumvalueof∼1.7MA,whichcorresponds,towithin performedwiththeuseoftwomagneticprobes,oneofwhich measurement errors, to the total discharge current measured wasinstalledatadistanceof12mmfromtheaxis. by the Rogowski coil or recovered using the signals from Thesignalfromthemagneticprobelocatedattheradius the loop detectors measuring the time derivative of the total of40mmalsoappearsabout6.2µsafterthebeginningofthe current. A certain difference in the signals from of the discharge, whereas the signal from the probe located at the loop detectors can be caused by the inhomogeneous current radius of 12mm appears 130ns later. From the time delay distribution at the collector, because the loop detectors are between these signals, we find that the average PCS radial arranged with a step of 120◦ along the collector perimeter. velocitybetweentheseprobesis≈2.1×107cms−1. Forthis 7 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal Figure6. PCSinterferogramsrecordednearthesystemaxisinshotno9131(D ,P =2.4Torr,U =24kVandW =384kJ)with 2 0 0 0 cylindricalprobeatdifferentinstantsfromthebeginningofthedischarge:(a)6479,(b)6509,(c)6549and(d)6569ns. Theneutronyieldis Y ∼8.7×1010neutrons/shot. n velocity,theskindepthestimatedfromtheformula which resulted in a reduction in the neutron yield. In this (cid:1) (cid:2) discharge, the discharge current was also almost completely V I˙≈I × r compressed to the radius of 4cm, but the neutron yield was δ skin about one-half order of magnitude lower than in the above is δ ≈ 1.6cm. Under the assumption of the Spitzer discharge. Unfortunately, theprobeinstalledattheradiusof skin conductivity, this value corresponds to the plasma electron 12mmwasdestroyedverysoon;therefore,itwasimpossible temperature of T ≈ 50eV. Discharge current somewhat tomeasurethecurrentamplitudeinthisregion. Itisnaturalto e increases (to 1.9MA in our case); however, the fraction of assumethattheprobeinstalledsoclosetotheaxiscouldaffect the current compressed to the axis decreases substantially, the process of pinch formation. This conclusion, however, 8 PlasmaPhys.Control.Fusion54(2012)025010 VKrauzetal 2 1 1.5 A 2 M 1 I, 0.5 0 6200 6400 6600 6800 7000 7200 7400 t,ns Figure7. Totaldischargecurrent(1)andcurrentrecoveredfromthe probemeasurement(2)obtainedunderthesameinitialconditions (D ,P =1.8Torr,U =24kVandW =384kJ)bytheplane 2 0 0 0 probe(solidline,shot9341,Y =6.9×1010n/pulse)andbythe n cylindricalprobe(dashedline,shot9340,Y =9.8×1010n/pulse). n Probelocationis4cmfromthedischargesystemaxis. can occur too hastily. First, earlier experiments carried out on the PF-3 facility with the use of x-ray pinhole cameras and frame cameras [22,24,43] demonstrated that a similar probe located at a distance of 2cm from the system axis practicallydidnotaffecttheprocessofpinchformation. On theotherhand,thelowerneutronyieldinthisdischargecanbe explainedbyotherfactors,inparticular,alowerPCScurrent (∼1.5MA) and a broader PCS with a complicated structure. The neutron yield measured in this discharge lies within the admissiblescatterobservedinaseriesofreferenceexperiments performed without magnetic probes. Nevertheless, taking into account fast destruction of the probe located so close to the axis and, accordingly, the impossibility of obtaining reliableinformationonthePCSparametersatthisradius,we abandonedtheapplicationofsuchprobesatthisposition. Another experimental session was performed after the power supply of the PF-1000 facility was upgraded. This made it possible to attain discharge currents of 1.5–1.9MA atalowerbankvoltageof20–24kV,whichcorrespondstoa lowerdischargeenergy,W =260–380kJ.However,although the energy deposited in the discharge was lower than in the previousseriesofexperiments,theneutronyieldremainedat thesamelevelof1010–1011 neutrons/shot. Thisindicatesthe conditionalcharacteroftheenergyscalingandtheimportance ofthecurrentscaling. This fact was clearly demonstrated in our experiments. In particular, we performed measurements in the stage of dischargesystemtraining,duringwhichthedischargevoltage wasgraduallyincreased. Inthecourseoftraining,theneutron yield gradually increased with increasing discharge voltage from20to23kV(figures10(a)and(b)). However,whenthe voltageincreasedto24kV,theneutronyielddegradedabruptly. Figure8. Resultsofthemeasurementoftheazimuthalmagnetic fieldinshotno8220(D ,P =3Torr,U =27kVand Probemeasurementsshowedthat,inthiscase,theefficiencyof 2 0 W =485kJ):(a)currenttimederivativesmeasuredbythetwocoils currenttransportationtowardtheaxisdecreasedsubstantially. oftheprobe(theverticaldashedlineshowstheinstantofprobe The probe located at the radius of 40mm recorded only destruction)and(b)currentmeasuredbythemagneticprobe(I ), p aboutone-halfofthetotalcurrentmeasuredbytheRogowski totaldischargecurrentmeasuredbytheintegratingRogowskicoil coil (figure 10(c)). It should be noted that, along with the (IR),totaldischargecurrentmeasuredbytheloopdetectorsatthe collector(I ),timederivativeofthetotaldischargecurrent upgrade of the power supply, the insulator of the discharge loop (dI/dt) andvoltage(U). TheprobeisatR=40mmand systemwasalsoreplaced. Itiswellknownthat,inthiscase, Z=10tmm. Theneutronyieldis1.24×1011neutrons/shot. itisnecessarytograduallytrainthedischargesysteminorder 9
The list of books you might like
