Full report

Original page 1

NASA/TM—2007–214959 Application of Simulated Reactivity Feedback in Nonnuclear Testing of a Direct-Drive Gas-Cooled Reactor S.M. Bragg-Sitton and K.L. Webster Marshall Space Flight Center, Marshall Space Flight Center, Alabama June 2007

Original page 2

The NASA STI Program…in Profile Since its founding, NASA has been dedicated to the advancement of aeronautics and space science. The NASA Scientific and Technical Information (STI) Program Office plays a key part in helping NASA maintain this important role. The NASA STI program operates under the auspices of the Agency Chief Information Officer. It collects, organizes, provides for archiving, and disseminates NASA’s STI. The NASA STI program provides access to the NASA Aeronautics and Space Database and its public interface, the NASA Technical Report Server, thus providing one of the largest collections of aeronautical and space science STI in the world. Results are published in both non- NASA channels and by NASA in the NASA STI Report Series, which includes the following report types: • TECHNICAL PUBLICATION. Reports of completed research or a major significant phase of research that present the results of NASA programs and include extensive data or theoretical analysis. Includes compilations of significant scientific and technical data and information deemed to be of continuing reference value. NASA’s counterpart of peerreviewed formal professional papers but has less stringent limitations on manuscript length and extent of graphic presentations. • TECHNICAL MEMORANDUM. Scientific and technical findings that are preliminary or of specialized interest, e.g., quick release reports, working papers, and bibliographies that contain minimal annotation. Does not contain extensive analysis. • CONTRACTOR REPORT. Scientific and technical findings by NASA-sponsored contractors and grantees. • CONFERENCE PUBLICATION. Collected papers from scientific and technical conferences, symposia, seminars, or other meetings sponsored or cosponsored by NASA. • SPECIAL PUBLICATION. Scientific, technical, or historical information from NASA programs, projects, and missions, often concerned with subjects having substantial public interest. • TECHNICAL TRANSLATION. Englishlanguage translations of foreign scientific and technical material pertinent to NASA’s mission. Specialized services also include creating custom thesauri, building customized databases, and organizing and publishing research results. For more information about the NASA STI program, see the following: • Access the NASA STI program home page at http://www.sti.nasa.gov • E-mail your question via the Internet to help@sti.nasa.gov • Fax your question to the NASA STI Help Desk at 301– 621–0134 • Phone the NASA STI Help Desk at 301– 621–0390 • Write to: NASA STI Help Desk NASA Center for AeroSpace Information 7115 Standard Drive Hanover, MD 21076–1320

Original page 3

NASA/TM—2007–214959 Application of Simulated Reactivity Feedback in Nonnuclear Testing of a Direct-Drive Gas-Cooled Reactor S.M. Bragg-Sitton and K.L. Webster Marshall Space Flight Center, Marshall Space Flight Center, Alabama Natonal Aeronautcs and Space Admnstraton Marshall Space Flght Center • MSFC, Alabama 35812 June 2007 

Original page 4

TRADEMARKS Trade names and trademarks are used in this report for identification only. This usage does not constitute an official endorsement, either expressed or implied, by the National Aeronautics and Space Administration. Avalable from: NASA Center for AeroSpace Informaton 7115 Standard Drve Hanover, MD 21076 –1320 301– 621– 0390 Ths report s also avalable n electronc form at <https://www2.st.nasa.gov> 

Original page 5

TAblE oF CoNTENTS 1. INTroDuCTIoN ......................................................................................................................... 1 2. SySTeM MoDel: PoINT reACTor KINeTICS .................................................................... 4 3. TeST ArTICle DeSCrIPTIoN .................................................................................................. 7 4. INSTruMeNTATIoN PlAN ....................................................................................................... 9 4.1 Temperature Measurement ...................................................................................................... 9 4.2 Deflection Measurement ......................................................................................................... 13 5. CoNTrol MeTHoDology: IMPleMeNTATIoN ................................................................ 17 5.1 Execution of Dynamic Model Within LabVIEW ................................................................... 19 5.2 Execution of Dynamic Model Within Simulink on LabVIEW PC ......................................... 19 5.3 Execution of Dynamic Model on a Target PC with User Datagram Protocol Communcatons ....................................................................................................... 19 5.4 Execution of Dynamic Model on a Target PC with Shared Memory ..................................... 21 5.5 Dynamic Model System Implementation ............................................................................... 21 5.6 Selection of a Computing Environment for Dynamic Testing ................................................ 27 6. CoNCluSIoNS ............................................................................................................................ 28 reFereNCeS ................................................................................................................................... 29 

Original page 6

lIST oF FIGURES 1. Dynamic system testing of the SAFE–100a with simulated reactivity feedback: (a) Core thermal power and calculated reactivity and (b) HX mass-flow rate, average core temperature, and average HP temperatures upstream and downstream of the HX ..... 2 2. Core block nlet end wth graphte heater elements nstalled ............................................. 7 3. DDg test artcle nstalled n the vacuum chamber ............................................................. 8 4. graphcal renderng of the DDg ........................................................................................ 8 5. (a) graphcal renderng and (b) hardware representaton of axal TC placement in the DDG core block with the central element replaced by an axial profile probe; TC locations are denoted by small circles .......................................................................... 10 6. Radial TC placement on the (a) Flow entrance (heater end) and (b) flow exit ends of the DDG core block; the central element is replaced by an axial profile probe and TC locations are denoted by small circles ................................................................... 11 7. Heater element placement nsde the DDg core block ....................................................... 11 8. TC location on graphite heater element for preliminary testing ......................................... 12 9. Screen shot taken during SAFE–100 testing identifying three axial locations for core wdth measurement ............................................................................................... 13 10. Configuration of the noncontact inductance proximity sensor for the measurement of core deflection during testing. The flange will be installed in the wall of the DDg pressure vessel, and the face of the rDP Py256C sensor wll be drected toward the core block ......................................................................................................... 15 11. Schematic diagram of data flow for the SAFE–100a test system ....................................... 18 12. xPC Target model development process ............................................................................. 19 13. (a) Data received by LabVIEW and (b) time-step delay between sent and receved data ................................................................................................................ 20 14. Simulink block diagram of the test dynamic model ........................................................... 22 v

Original page 7

lIST oF FIGURES (Continued) 15. Smulnk block dagram for the core thermal model ........................................................... 23 16. LabVIEW application: VMIC test ....................................................................................... 24 17. Simulated reactivity response to a reactivity insertion at various data rates ....................... 25 18. Simulated power response to a reactivity insertion at various data rates ............................ 25 19. Simulated temperature response to a reactivity insertion at various data rates ................... 26 lIST oF TAblES Delayed neutron properties for fast fission of 235u ............................................................. 5 1. 2. Software and hardware configurations for various data acquisition and control systems ... 18 v

Original page 8

lIST oF ACRoNYMS Ar argon API applcaton nterface AWG American wire gauge CCD charge-coupled device DDG direct-drive gas DVRT differential variable reluctance transducer EFF–TF Early Flight Fission-Test Facility He helum HeXe helium-xenon HP heat ppe HX` heat exchanger MSFC Marshall Space Flght Center Na sodum PCI perpheral component nterconnect PKE point kenetics equation SAFe Safe Affordable Fsson engne TC thermocouple TCP–TP transmission control protocol-internet protocol TM Techncal Memorandum v

Original page 9

lIST oF ACRoNYMS (Continued) uDP user datagram protocol uN uranum ntrde VMIC VME Microsystems International Corporation Xe xenon v

Original page 10

NoMENClATURE A heat transfer area (m2) Ci delayed-neutron concentration, ith precursor group Cp specific heat of the coolant Cp,struct specific heat of the core structure dL measured core deflection, ultimately a function of temperature at any given time Fs data sample rate (s) hc convective heat transfer coefficient (assumed constant) k effectve neutron multplcaton factor Mcoolant mass of the coolant n the core (kg) Mstruct mass of the core structure (kg) m mass flow rate (assumed constant) (kg/s) P reactor thermal power (W) PRx reactor thermal power (W) T temperature (K) Tin coolant nlet temperature (K) Tcoolant,avg average temperature of coolant n the core (K) Tout coolant ext temperature (K) Tstruct average structural temperature (K) t tme (s) αT temperature coefficient of reactivity v

Original page 11

NoMENClATURE (Continued) β total delayed-neutron fraction (sum of all βi) βi delayed-neutron fraction, ith precursor group Λ prompt neutron generaton tme (s) th precursor group (s–1) λi decay constant for the i ρ reactivity x

Original page 12

x

Original page 13

TeCHNICAl MeMorANDuM ApplICATIoN oF SIMUlATED REACTIvITY FEEDbACK IN NoNNUClEAR TESTING oF A DIRECT-DRIvE GAS-CoolED REACToR 1. INTRoDUCTIoN Nonnuclear testing can be used to evaluate the operation of an integrated nuclear system within a reasonable cost and schedule in order to provide valuable input to the overall system design. Various operational regimes can be studied in nonnuclear testing to validate thermal and thermal hydraulic codes, assess thermal hydraulic behavior, characterize stress/strain in the system during operation, and verify system integration processes. However, because the electrically heated core lacks neutrons, the dynamic neutronic response of the system cannot be fully simulated without the incorporation of additional system models. In early 2005, dynamic testing was performed on the Safe Affordable Fission Engine (SAFE) 100a test article at the NASA Marshall Space Flight Center (MSFC) Early Flight Fission Test Facility (EFF–TF). The SAFE–100a is a resistively heated, stainless steel heat pipe (HP) reactor core segment coupled to a prototypic gas-flow heat exchanger (HX). Core power transients that were simulated in the electrically heated system tests were controlled by a point kinetics model with reactivity feedback based on a bulk temperature reactivity feedback coefficient and core average temperature. This type of nonnuclear test s expected to provde a reasonable approxmaton of reactor transent behavor because reactivity feedback is very simple in a compact fast spectrum reactor (simple, negative, and relatively monotonic temperature feedback, caused mostly by thermal expansion) and calculations show that no significant reactivity effects are associated with fluid in the HP (the worth of the entire inventory of sodium (Na) in the core is much less than one dollar, so fluid movement and temperature changes will cause very minor effects). In previous SAFE–100 tests, the point kinetics model was based on core thermal expansion via deflection measurements.1 These tests demonstrated that core deflection was a strong functon of how the SAFe–100 modules were fabrcated and assembled, n terms of straghtness, gaps, and other tolerances. Both the SAFe–100 and 100a were constructed from ndvdual modules that nested together to form a core rather than beng constructed from a monolthc block. To remove the added variable of SAFE–100a expansion with varying temperature, as compared to a different concept, it was decided to use a temperature-based feedback model, based on several thermocouples (TCs) placed throughout the core. The bulk core temperature feedback coefficient for most reactors of this class s on the order of –0.1 to –0.2 cents per degree Kelvn. 1

Original page 14

The test matrix applied in SAFE–100a testing included changes in coolant flow rate to simulate system load variations and step reactivity insertions and decreases. The system responded as expected for all induced transients—increased/decreased coolant flow rate with reactivity feedback, increased/ decreased coolant flow rate without reactivity feedback (done for comparison), and positive/negative reactivity insertion. Expected implies that the transient response was typical of any simple reactor, whether it be cooled by a pumped fluid (gas or liquid metal), or by HPs. Figure 1 shows the system response to a series of transients in which the system load was increased several times, represented by mass-flow increase through the HX, followed by a step reactivity insertion of $0.06. An in-depth discussion of the SAFE–100a system response can be found in reference 2. 0.07 35,000 0.06 Total Core Power (W)800 Mass-Flow Rate (kg/s) 30,000 0.05 0.04 25,000 0.03 Total Core 20,000 Power 0.02 Reactivity 15,000 0.01 Reactivity ($) 0 10,000 –0.01 Increase in Mass Reactivity 5,000 –0.02 Flow Rate Insertion ($0.06) –0.03 0 0:08:38 0:17:17 0:25:55 0:34:34 0:43:12 0:51:50 1:00:29 (a) Elapsed Time (hr:min:s) 850 0.05 0.04 0.03 He Mass-Flow Rate 750 0.02 0.01 700 Average Core 0 650 Temperature –0.01 Average HX –0.02 Temperature (°C)600 Upstream/Downstream Temperature –0.03 550 –0.04 500 –0.05 0:08:38 0:17:17 0:25:55 0:34:34 0:43:12 0:51:50 1:00:29 (b) Elapsed Time Figure 1. Dynamic system testing of the SAFE–100a with simulated reactivity feedback: (a) Core thermal power and calculated reactivity and (b) HX mass-flow rate, average core temperature, and average HP temperatures upstream and downstream of the HX. During the applied transients, the HPs appear to act as very good conductors with relatively little thermal inertia, such that system response would be expected to be very similar in a reactor design in which gas directly removes power from the core, as in a direct-drive gas (DDG) cooled reactor that does not incorporate an intermediate HX. The time scale of the transient is dictated mostly by the thermal inertia of the core, with a smaller extent being dictated by the HPs and HX for the SAFE–100a. In each transient applied to the SAFE–100a, the peaks and valleys of the power and temperature occurred within 3–5 min after transient initiation, followed by modest oscillations until the system stabilized at a new steady state condition approximately 20–30 min after the transient was initiated. The magnitude of power and temperature overshoots were also modest; ≈15 K temperature overshoot for a $0.10 reactivity insertion, although the magnitude of overshoot is not of great meaning because the SAFE–100a is a low-power density core with nonprototypic thermal inertia. Similar dynamic test methodology is planned for application with the electrically heated, DDG test article currently undergoing testing at MSFC. Designed by engineers at Sandia National Laboratory and the Los Alamos National Laboratory, the DDG was previously tested to assess basic flow characteristics and system pressure drop.3The core must be refurbshed to ncorporate addtonal nstrumentaton for measurng core block temperatures and to replace the exstng graphte heater elements that have 2

Original page 15

undergone a significant amount of thermal cycling during previous checkout tests. Instrumentation will be positioned to provide extensive characterization of the core block temperature, both axially and radially, and to monitor coolant gas temperature in regions of interest. Details of the proposed instrumentation will be discussed in this Technical Memorandum (TM). More detailed characterization of core temperatures will allow implementation of a higher fidelity neutronics model that incorporates multiple, localized reactivity feedback coefficients as a function of various independent core temperature measurements. Additionally, investigation of deflection measurement techniques for application inside a pressure vessel has commenced and their potential applicability of these methods will be discussed. 3

Original page 16

1. SYSTEM MoDEl: poINT REACToR KINETICS Reactor dynamics can be modeled using the point kinetics equations (PKEs), which can be derived from transport and diffusion theory.4 The PKEs representation provides only an approximate model of the reactor; it does not provide a mechanism to describe neutron energy effects or structural details in a heterogeneous reactor. Because fast reactors are relatively small in size and do not contain any moderator to slow the neutrons to a lower energy, the PKEs are a good approximation for the dynamics in a fast reactor. In the absence of an external source and written in terms of reactor thermal power, the PKEs are given by dP (ρ − β) = dt Λ and dCi βi = P − λiCi , (2) dt Λ where the reactivity, ρ, is given by k −1 ρ = P + λiCi (1) (3) k and k is the effective neutron multiplication factor. Reactivity is often expressed in units of dollars. The reactivity scale between zero and β, the total delayed neutron fraction, is divided into 100 cents, where 100 cents of reactivity are equal to 1 dollar. Hence, the reactivity expressed in units of dollars is given by ρ / β . Studies of reactor dynamics generally recognize six distinct groups of delayed neutrons. Approximate solutions to the PKEs can be found using a smaller number of groups to represent the delayed neutron population. The dynamic model applied in the previous application of simulated reactivity feedback on the SAFE–100a included a simple dynamic model having one group of delayed neutrons, applying a weighted average decay constant and the total delayed neutron fraction. The one group decay constant (λ) is calculated as a weighted average of the six individual decay constants. A one-group representation simplified the computational model, speeding the real-time implementation of simulated reactivity feedback during test. However, to improve the fidelity of the neutronic model, a more detailed PKEs representation should be employed. A complete six-group model, requiring seven ordinary differential equations, will be applied in DDG testing. The individual decay constants and delayed neutron fractions for a fast spectrum reactor, 4

Original page 17

as reported by Hetrick, are provided in Table 1.4 To maintain real-time implementation of the neutronics model, with updates to the PKEs solution implemented on the order of 2 Hz, high-speed computing must be employed. Potential options for implementing real-time system control will be discussed in section 5. Table 1. Delayed neutron properties for fast fission of 235u. Delayed Neutron Decay Constant Relative Yield Group λi (s)–1 β /β i 1 0.0127 0.038 2 0.0317 0.213 3 0.115 0.188 4 0.311 0.407 5 1.4 0.128 6 3.87 0.026 0.0767 s–1 Total decay constant, λ = Total delayed neutron fraction, β = 0.00642 A smple model of a pont reactor wth feedback was appled n approxmatng the temperature reactivity feedback in the SAFE–100a. In that application, a single bulk reactivity feedback coefficient (αT) was appled to the entre core. The feedback calculaton was then made based on bulk core temperature changes that were determined by measuring multiple temperatures at the axial core centerlne, and calculatng the average core temperature usng these data. In ths smple case the total reactivity of the reactor core is given by equation (4) as: ρ = ρo + αT (∆T ) , (4) where αT is the temperature coefficient of reactivity, αT = dρ , (5) dT and ρo is the initial steady state reactivity before a transient is applied to the reactor power. In the previous applcaton, αT represented a bulk feedback coefficient, incorporating all reactivity feedback effects due to fuel and structural temperature changes, ncludng Doppler broadened cross sectons and thermal expansion effects. However, if highly precise temperature (and possibly deflection) measurements can be made, each of these individual effects on reactivity can be assessed separately in the feedback model. For example, equation (4) would be expanded to include several components of reactivity feedback as follows: ρ = ρo + ρdL (dL T( )) + αT ,fuel(∆Tfuel) + αT ,block bllock T ,reflector reflector (∆T ) + α (∆T ) + . (6) 5

Original page 18

In equation (6), temperature feedback was broken into components based on the measured core dimensions and the current fuel, core block, and reflector temperatures. In nonnuclear testing, fuel temperature changes must be approxmated from measured temperatures n the electrc heatng element or heater sheath, which is equivalent to the fuel-clad structure. Additionally, the nonnuclear test article may not incorporate a full reflector. In this case, the reflector temperature can be inferred from measured temperatures near the periphery of the core block and at the periphery of the pressure vessel in the case of a reactor cooled by a pumped fluid, as in the DDG. One might choose to base reactivity feedback calculations solely on temperature measurements in various locations, or one might determine reactivity feedback from measured deflection in various regions of the core. Monte Carlo calculations can be performed to assess the reactivity worth of various core deflections that occur due to elevated temperature during core operation. The largest reactivity worth in a fast-spectrum reactor is expected to be due to fuel expansion with temperature, which can be further separated into radial and axial expansion; however, in the absence of an electric heater that closely mimics the dynamic characteristics of nuclear fuel, including thermal expansion, deflectionbased reactivity feedback will likely need to be based on measured deflection of the core block, as was performed n SAFe–100 testng.1 Options for measuring the deflection inside the DDG pressure vessel wll be dscussed n secton 4.2. 6

Original page 19

1. TEST ARTIClE DESCRIpTIoN The DDG is a gas cooled, UN-fueled, pin-type fast reactor designed to use helium-xenon (HeXe) gas flowing directly into a recuperated Brayton system to produce electricity for nuclear-electric propulson.5 The DDG, which consists of 37 fuel pins in a block matrix, is designed to operate at 32 kW thermal. This is a reduced power version of a full-scale design, similar to the SAFE-30 test series that preceded the SAFe–100 and 100a test seres.6 Each fuel pin is surrounded by an annular flow channel that provides direct heat removal by flowing gas. The DDG design was initially provided by Sandia Natonal laboratores but was transferred to los Alamos Natonal laboratores for addtonal testng support. The DDG is constructed entirely of stainless steel. The core block is a 53.3-cm (21-in) long, 16.5-cm (6.5-in) (flat-to-flat) hexagonal, solid stainless steel block with a pattern of 37 gun-drilled 0.75-in-diameter holes through the length. A closeup of the core inlet region, with graphite electric heaters (thermal simulators) installed, is shown in figure 2. Thermal Core Simulators Connectors Fgure 2. Core block nlet end wth graphte heater elements nstalled. The DDg was desgned to operate at 2.4 MPa (350 psa) usng an HeAr gas coolant wth a core inlet temperature of 650 K, core exit temperature of 850 K, and bulk coolant flow of 0.1 kg/s. HeAr, which can duplicate the HeXe gas requirements, can be used in early tests to reduce the cost of the test program. Intal DDg testng was performed wth ntrogen (N2) gas to allow early system checkouts and verification of the test matrix for early problem identification, with gas-flow parameters selected based on a Reynolds number comparison. The test article is also plumbed to allow testing with any desired mxture of He and Ar. In the DDg desgn, the coolant gas enters the upper bonnet and s drected va a manifold to six 2.54-cm (1-in) inlet tubes located on the upper bonnet assembly, as shown in figures 3 and 4. The gas then flows down between the pressure vessel wall, called the downcomer region, 7

Original page 20

ExitExit InletInlet UpperUpper PipesPipes ManifoldManifold BonnetBonnet InletInlet PipesPipes TC Leads PressurePressure VesselVessel Power LeadsPower Leads Fgure 3. DDg test artcle nstalled n the vacuum chamber. Exit Inlet Upper Bonnet Core Heater Channel Pressure Insulation Vessel Shield Power Feeds Fgure 4. graphcal renderng of the DDg. keepng the vessel wall at a constant lower temperature and provdng gas nsulaton of the core. The insulation layer is constructed by including a metal shield between the core block and the pressure vessel wall, causing a stagnant gas layer to form between the core block and the metal shield. At the base of the pressure vessel, the gas collects in a plenum and is directed back up through 37 annular flow channels n the core. Flow parameters are selected such that the gas s heated ≈200 K through the core. ext gas is then collected in an exit plenum where the flow is recombined and allowed to exit through the upper bonnet center ext ppe. 8

Original page 21

1. INSTRUMENTATIoN plAN Intal DDg checkout testng was performed wth mnmal nstrumentaton of the DDg test artcle. The primary intent of these initial tests was to verify core operation, assess gas-flow temperatures, and determine experimentally the core pressure drop to validate thermal hydraulics codes. A majority of the TCs were installed to measure gas-flow temperature in various regions. A few TCs were tack welded to the core block surface to obtan an approxmate block temperature, but the temperature measurements from these TCs likely provided a combined block and flow temperature. Rather than being embedded in the core block, these TCs were mounted to the upper block surface where gas flow provided some cooling of the TC junction. No temperature measurements were made in the internal block region. During refurbishment of the DDG core, necessary to install new electric heaters for the next phase of testing, significant modifications will be made to the core instrumentation to provide the necessary data for implementation of simulated reactivity feedback in dynamic system testing. 4.1 Temperature Measurement Various options for temperature measurement in the DDG core block and heater elements are currently being investigated. Ideally, temperature measurements would be made in the gas-flow regions wthn the core block, and between the electrc heater and the sheath/fuel clad to obtan an estmate of fuel temperature. Gas-flow temperatures were measured in the initial DDG testing and will again be measured in planned tests. Additional type-K TCs will be positioned in key regions of interest to assess gas temperature: two in the gas blanket insulation region, two in the flowing gas downcomer region, two n the gas nlet plenum, and two n the gas outlet plenum. Addtonal TCs wll be placed throughout the core block to fully characterize the block temperature both radially and axially. Methods of measuring the temperature near the heater element are also beng nvestgated. 4.1.1 Core block Temperature Measurements A series of sheathed, type-K TCs will be installed throughout the DDG core block to fully characterize axial and radial temperature distribution. To minimize potential damage to the core block, machining of the block necessary for TC placement will be kept to a minimum. To characterize the axial temperature dstrbuton, t s desrable to obtan temperature measurements along the core centerlne and periphery at several axial locations. To accomplish this, the central heating element will be removed (reducing the number of heaters and flow channels to 36, representing a design configuration in which a central safety rod might be used) and will be replaced with an Inconel® sheathed axial profile probe. Manufactured by Omega, these probes can be purchased in the desired length and diameter, with the number and location of TCs located along that length also specified by the customer. The DDG core block is 52.7-cm (20.75-in) long and is equipped with 50.8-cm (20-in) heater elements. The bore holes in the core that accommodate a heater, sheath, and annular flow path are 1.9-cm (0.75-in) diameter. For DDG testing, a 51.4-cm (20.25-in), 1.7-cm (0.67-in) diameter probe will be employed with 10 TCs located at seven unique axial positions—2.54, 10.5, 18.4, 26.4, 34.3, 42.2, and 50.2 cm (1, 4.125, 7.25, 9

Original page 22

10.375, 13.5, 16.625, and 19.75 n). To ensure that measurements can be taken at these postons, even n the case of a faled TC, two ndependent TCs wll be located at three dfferent locatons: 2.54, 26.4, and 50.2 cm (1, 10.375, and 19.75 in). A 1.3-cm (0.5-in) long metal plug installed in the central channel at the gas exit end of the core will prevent gas flow along the axial profile probe. Positioning the first and last TCs 2.54 cm (1 in) from either end of the core block ensures that they will be located in the heated region, accounting for the slightly shorter heating element (50.8-cm versus 52.7-cm core block). The axial centerline probe will be complemented by seven TCs located on the core periphery, embedded ≈0.64 cm (0.25 n) nto the core block. The axal postons of these TCs are selected to correspond wth the TC positions in the profile probe, as shown in figure 5. Axial ProfileAxial Profile ProbeProbe (a) (b) Fgure 5. (a) graphcal renderng and (b) hardware representaton of axal TC placement in the DDG core block with the central element replaced by an axial profile probe; TC locations are denoted by small circles. The radial temperature profile will be obtained by embedding TCs into the core block at both the inlet and outlet surfaces. The TC positions are selected to provide radial profiles of interest in two dmensons at each core surface. These TCs wll be embedded ≈2.54 cm (1 n) nto the block to ensure that they are located in the heated region as discussed above. Complete radial profiles at either end of the core block and complete measurement of the axial profiles along the centerline and the periphery will allow the radial temperature profile to be interpolated at any axial position. Proposed TC locations are superimposed on photos of the core block in figure 6. The proposed positions on the gas exit end of the core take advantage of existing mounting holes in the core block (originally intended for placement of flow orifice pieces) that are not currently being used. These holes will be bored to the desired 2.54-cm (1-in) depth for TC placement. Dashed lines in figure 6 show the approximate temperature profile obtained from surface TC measurements. 10

Original page 23

Axial Profile Probe Axial Profile Probe (a) Flow Exit End of Core Block Axial Profile Probe (b) Figure 6. Radial TC placement on the (a) Flow entrance (heater end) and (b) flow exit ends of the DDG core block; the central element is replaced by an axial profile probe and TC locations are denoted by small circles. 4.1.2 Fuel Temperature Measurement To implement detailed reactivity feedback, it is also desirable to measure temperature in the heater region to approximate the corresponding fuel temperature at a given power level, which requires the insertion of a temperature measurement device into the 0.7-cm (0.28 in) gap between the graphite heater element and the sheath structure, shown in figure 7; the gap is currently filled by an alumina space ring at three axial locations. This may be accomplished with very fine TCs embedded in the insulator rings on the heater element, or fiber optic probes might be inserted along the length to obtain a complete descrpton of the temperature n ths regon at multple axal postons. Power Leads Sheath/Clad Graphite Heater Core Block Annular Flow Gap Alumina Spacer Fgure 7. Heater element placement nsde the DDg core block. 11

Original page 24

Fiber optic temperature probes developed by Luna Innovations, Inc. are constructed with Bragg gratings spaced evenly along the sensing length to obtain detailed temperature profiles; standard fiber optic design allows measurement of temperature at 1-cm (0.39-in) increments. These fiber optics currently have a temperature limitation of 1,273 ºK for testing over extended times and may not survive over long test times at elevated temperatures in the DDG core. Luna fiber optic probes tested on the SAFE–100 demonstrated superior ability in generating three dimensional temperature profiles of the core.7 Luna is currently working to extend the maximum operating temperature of the fiber optic sensors, which may qualify them for use as embedded temperature instrumentation in future advanced heater elements. Previous tests with Luna fiber optics were conducted by Luna, who brought the support hardware necessary to obtain and record data from the fiber optic probes to MSFC for temporary use. The EFF–TF is currently working to obtain an in-house capability to use fiber optic sensing techniques n future testng applcatons. At present, standard TCs have been nvestgated to measure temperature along the heatng element. Because TCs are relatively inexpensive and are available in a wide range of sizes and configurations, it was feasible to test a proposed configuration in a relatively short time frame. Six 0.254-mm (0.01-in) diameter, Inconel-sheathed type-K TCs with an ungrounded junction were procured to assess their applicability in DDG testing. Graphite heater elements used in the DDG testing are 0.985 cm (0.375 n) n dameter and are sngle ended (desgned such that both nlet and outlet power leads are on a single end). These heater elements fit inside a 1.4-cm (0.55-in) diameter sheath, the outer diameter of which forms the annular flow path. The heater is separated from the sheath using three alumina spacer rings to prevent contact between the graphite element and the metal sheath. The junction end of a TC was bonded to the graphte element at each alumna rng usng alumna cement to obtan three temperature measurements along each element, as shown in figure 8. During testing, several problems were noted with this configuration. Because the TC sheath is made of an electrically conductive materal (Inconel), bondng the sheath to the heater n more than one locaton set up a parallel current path in the Inconel sheath. The current required to operate the heater element caused the TC sheath to vaporize and the 52 guage TC wires did not survive. Future investigation will consider alternate methods of embedding TCs within the heater elements to obtain approximate fuel temperature measurements; e.g., measurement of heater temperature that can be directly correlated to the expected fuel temperature at comparable condtons. Graphite Heater Heater Power Leads TC Lead Wires Embedded TCs Figure 8. TC location on graphite heater element for preliminary testing. 12

Original page 25

4.2 Deflection Measurement Reactivity feedback based on core deflection measurements was implemented in SAFE–100 hardware testng.1 Note that tests discussed in section 1 were performed on the SAFE–100a, a 19-module partial array of the SAFE–100 design that utilizes Na-filled HPs and is coupled to a prototypic gasflow HX. The SAFE–100 was a predecessor to the SAFE–100a. The SAFE–100, designed to provide 100 kW power, includes 61 modules, each containing three fuel tubes and one HP. In the test article, a total of 183 graphite resistance heaters were used to simulate the heat of nuclear fission. However, the SAFE–100 test article had no mechanism for heat removal from the core; hollow stainless steel tubes were used to represent the HPs n each module. As a result, ths test artcle was used to smulate the operation of a fission system in the absence of cooling; this condition is similar to the operation of a HP reactor during the start-up period prior to complete thaw of the HP working fluid and initialization of the HXs. Deflection tests were performed on the 316 stainless steel (SS316) SAFE–100 core, manufactured accordng to the SAFe–100 core fabrcaton drawngs. Because the SAFE–100 did not require a pressure vessel, visual access to the core was maintained. In this case, deflection could be measured using high-resolution photography. Mounted outside the vacuum chamber, imagery from a 6.3 mega-pixel high-resolution camera, composed of a chargecoupled device (CCD) and a camera lens, was used to assess core deflection as a function of temperature by scanning multiple cross sections along the axial dimension of the core (fig. 9). The SAFE–100 was located on an immobile stand inside the chamber, and the CCD imaging system had a clear view of the core through a fused slca vewng port. SAFE–100 Cross-Section Diameter Current Diameter 1: 267.02 mm Current Diameter 2: 266.88 mm Current Diameter 3: 266.55 mm D3 D2 D1 Figure 9. Screen shot taken during SAFE–100 testing identifying three axal locatons for core wdth measurement. Deflection measurement becomes increasingly complex when visual access is not readily available. Because the DDG must be tested inside a pressure vessel, novel deflection measurement techniques must be investigated. Displacement can be measured using an optical proximity sensor; e.g. a laser range finder, a differential variable reluctance transducer (DVRT), or a noncontact inductance proximity sensor. Each of these was evaluated for its applicability in DDG testing. The next stage of DDG testing will 13

Original page 26

be performed at ≈ 0.34 MPa (50 psi) to accommodate a Brayton energy conversion system that may be incorporated with the DDG hardware. This reduced pressure condition somewhat relaxes the requirements for a deflection measurement system, but the gas temperature will be approximately the same as n prevous testng. Initial investigations indicate that optical proximity sensors or laser range finders may not be reliable methods of deflection measurement. In this application, the sensor would be directed at the external core block surface. This would require a window to be installed in the DDG pressure vessel, and it is difficult to find ports with viewing windows rated for the experimental pressure and temperature conditions (in the downcomer region, the inlet gas flow is ≈450 K). Additionally, these instruments are dependent on the emissivity of the surface being investigated. As the core surface temperature increases, the emissivity of the stainless steel block changes such that the accuracy of the sensor over tme can be compromsed. DVRTs were previously investigated for application to the SAFE–100 and–100a test articles. A DVRT uses a plunger that is in contact with the core surface. Deflection is then determined by measuring differential reluctance between magnetically shielded DVRT coils. Unfortunately, very few of the commercially available units are rated for high-temperature operation (>773 K). Although the sensor would not be n drect contact wth the hot core block, the plunger conducts heat to the sensor and can cause the operatng temperature to be exceeded. elmnatng drect contact wth the hot core block surface reduces the temperature that wll be seen by a sensor housing. This can be accomplished using a noncontact inductance proximity sensor. These devices are commercially available for operation at pressure and at elevated temperatures. RDP Electrosense, Inc. manufactures high-temperature, noncontact transducers that employ an inductive proximity position sensor principle and will operate with any ferritic material. The output of the displacement transducer vares as the gap between the front face of the transducer and the target changes. These sensors have electron beam-welded joints and are rated for operation from full vacuum up to 20 MPa (2,900 ps) and up to 873 K. The rDP sensor can be purchased wth a measurement range of up to 6 mm, with a quoted linearity of ±0.05 percent of full scale. Sensor model PY256C has been ordered to determine potential applicability in DDG testing. The sensor is currently being retrofitted to place a pressure feedthrough on the sgnal cable that wll allow t to be nstalled through the DDg spool. The basic configuration is shown in figure 10. The sensor head itself is 4.8-cm (1.9-in) long with a maximum diameter of 2.8 cm (1.1 in), making it a tight fit between the core and the pressure vessel. The unit must be calbrated for operaton n the desred envronment pror to mplementaton, somewhat complcatng its use. Initial calibration and testing will commence after the modified unit has been received and prior to ntegraton n the DDg. 14

Original page 27

Ferrules (Supplied by NASA) Nut (Supplied by NASA, Do Not Swage) RDP PY256C Flange With Welded Fitting (Supplied by NASA) Figure 10. Configuration of the nocontact inductance proximity sensor for the measurement of core deflection during testing. The flange will be installed in the wall of the DDG pressure vessel, and the face of the rDP Py256C sensor wll be drected toward the core block. 15

Original page 28

16

Original page 29

1. CoNTRol METhoDoloGY: IMplEMENTATIoN Specialized LabVIEW®programs have been developed at the eFF–TF to mplement all aspects of system operation. One visual interface provides power supply control while a second monitors core temperatures and controls the gas supply to remove heat from the core (either for the HX in the SAFE– 100a or the direct gas flow through the DDG). To implement reactivity feedback control, the LabVIEW control program additionally implements a solution to the PKEs in order to govern core power response to the transent. In SAFE–100a testing, which implemented the PKEs with one group of delayed neutrons, the neutronic model included only two differential equations (eqs. (1) and (2)). The simplicity of this model allowed the solution to the system of equations to be precalculated and programmed into LabVIEW in algebraic form, minimizing the computation time required by the LabVIEW data acquisition and control system. The solutions to the PKEs were then implemented in real time at each iteration of the control loop. All TCs installed in the test article were scanned at a rate of 5 Hz (approximately once every 200 ms). The power supply controller set points were updated at a rate of 1 Hz. Taking into consideration the time required to scan the TCs and depending on how well the scanning rate and the set point update rates were synchronized, the total lag between taking the temperature measurement and updating the controller set point varied between approximately 200 and 900 ms. Since there were multiple computers in the control loop, the total end-to-end lag would drift at a rate dependent on how well the clocks of each system matched. Quantitative measurements of the lag and drift were not made. It was assumed that the time constants of the physical processes were much greater than 900 ms in the tested system. Future modifications to the software architecture can be made to constrain the lag to a smaller fixed value with a small amount of jitter. With hardware changes even greater improvement can be achieved. The basic control methodology that was applied in previous applications of dynamic testing is summarized in the schematic in figure 11. The system can basically be divided into three parts: (1) Test hardware, (2) LabVIEW data acquisition system, and (3) LabVIEW power control system. In this representation the test hardware consists of the core, instrumentation, and stand-alone power supplies. The test is controlled with two computer systems, both using customized programs built in LabVIEW. The LabVIEW data acquisition system collects data from all instrumentation and broadcasts that data over a local ethernet. The second system, the power control system, reads data broadcast by the data acquisition system, computes a new power level from the PKEs solution, and updates the set points used to control the power supples. 17

Original page 30

Test Hardware LabVIEW Data Acquisition System Power Control Read Data Test Article Broadcast Data Temperature Measurement LabVIEW Power Control System Receive Data Compute Reactivity Change and Reactor Power Change Power Level Set Points Figure 11. Schematic diagram of data flow for the SAFE–100a test system. DDG testing will require increased computation time to solve a system of seven differential equations at each iteration of the control loop, due to the incorporation of six delayed neutron groups n the neutronc model. To accomplsh ths wthout mpactng the control tme, nvestgaton has been initiated to determine ways in which the computational efficacy can be improved. Potential options for dynamic system control are summarized in table 2. Currently, the most promising arrangement appears n a MATlAB®Smulnk®code that can be operated to be the development of the neutroncs model  on a separate computer that interfaces back to the governing LabVIEW data acquisition and control code. MATLAB Simulink utilizes a graphical, time-based differential equation solver. This arrangement requires that the dynamic model computer run xPC Target, an additional MATLAB product. The control computer can then communcate wth the model computer, sendng updated data to the Smulnk model at specified time steps, which then returns the new power solution to the PKEs for application to the system. Both UDP communications and shared memory are being investigated as potential communication optons between the target and control PCs. Table 2. Software and hardware configurations for various data acquisition and control systems. Data Acquisition Dynamic Model Power Control Description 1 LabVIEW—acquisition computer LabVIEW—power control computer LabVIEW—power control SAFE–100a configuration. Dynamic computer model contains limited fidelity. 2 LabVIEW—acquisition computer Simulink/LabVIEW interface toolkit- LabVIEW—power control Simulation Interface Toolkit for power control computer computer LabVIEW is required. Model is executed on same PC as LabVIEW. 3 LabVIEW—acquisition computer Simulink/xPC target (UDP)—dynamic LabVIEW—power control A separate PC running xPC target model computer 4 LabVIEW—acquisition computer Simulink/xPC target (shared computer executes the model. Data are transferred between model and LabVIEW via UDP communication. LabVIEW—power control A separate PC running xPC target memory)—dynamic model computer computer executes the model. Data are trans- 18 ferred via shared memory boards.

Original page 31

5.1 Execution of Dynamic Model Within labvIEW The first configuration listed in table 2 is the same as that described for the SAFE–100a test setup. Based on a change in temperature, which results in a change in the reactivity in a reactor, LabVIEW is used to compute a new solution to the PKEs, which is then translated to a new set of reactor power level setpoints over the projected time period. Although this method could be readily implemented on the DDG system, it becomes less desirable compared to other options that incorporate Simulink as the fidelity of the neutronic model is improved, increasing the computation time necessary at each applcaton of the control loop. 5.2 Execution of Dynamic Model Within Simulink on labvIEW pC A Simulink model may be executed on the same PC as a LabVIEW program through the use of the Simulation Interface Toolkit produced by National Instruments. This toolkit allows for communication between a LabVIEW program and a Simulink model. To assess the benefits and limitations of this configuration, further work needs to be performed. Preliminary work has shown that, as configured, the toolkit allows for limited interaction between a Simulink model and LabVIEW program. LabVIEW Real Time software may be required to allow for increased functionality between the LabVIEW program and Simulink model. Although this configuration may require the least amount of hardware changes to the system, the added complexity to the LabVIEW program required to accommodate a Simulink interface may offset the added complexity of new hardware required in the remaining configuration options. 5.3 Execution of Dynamic Model on a Target pC with User Datagram protocol Communications 5.3.1 xpC Target To remove the computatonal burden of executng a Smulnk model on the same PC as a lab- VIEW program, xPC Target may be used to execute the Simulink model on a standalone PC. A boot disk produced by the xPC Target loads the xPC Target application on the target PC at bootup. With TCP/IP communication over an ethernet, the Simulink model may then be downloaded to the target PC. This may be accomplished several ways. If the model has already been built and compiled, LabVIEW may be used to communcate wth the target PC through the use of applcaton nterface (API) drvers provded by MATLAB. Figure 12 shows the process used to build and deploy a model with xPC Target. Download Model Execute Model Develop Model Build Target Application to Target PC in Real Time in Simulink With Real-Time Workshop Fgure 12. xPC Target model development process. 19

Original page 32

5.3.2 User Datagram protocol Communication Once the model has been downloaded to the target PC, it is ready to be used in the system. To interact with the rest of the test system, the model must be capable of receiving reactivity information, then sendng the power computed from a soluton to the PKes. one possble method of communcaton between the model and target PCs is the UDP communication protocol. The SAFE–100a test system used ths protocol to dstrbute data over the local ethernet. To test the ability of the Simulink model to interact with the LabVIEW application via UDP communication, a LabVIEW application and Simulink model were created. The LabVIEW model created and sent the value of a sinusoidal signal once every 250 ms. The Simulink model was configured with a fixed step size of 250 ms. At each time step, the model read the most recent data from the communications port and then sent the data back to the LabVIEW model, where LabVIEW then read the data and displayed it. Testing indicated that synchronization of the LabVIEW software and the Simulink model was difficult in this configuration. Both the LabVIEW and Simulink systems were operating at a sample rate of 250 ms. Depending on when the Simulink model was started relative to the LabVIEW application, data received by the LabVIEW application could lag behind the sent data by as much as 250 ms. If this lag could be guaranteed, UDP communications might have been an acceptable method; however, if the LabVIEW application speeds up or slows down relative to the Simulink model (this is possible when windows are opened or closed on the LabVIEW PC), the lag between sent and received data could temporarily increase. Figure 13 shows data received by the LabVIEW application and the number of tme steps between the sent and receved data. Ths varable tme lag between sent and received data could yield incorrect results from the dynamic model. (a) (b) Figure 13. (a) Data received by LabVIEW and (b) time-step delay between sent and received data. One method that could be used to mitigate the variability of the time lag between sent and received data would be to decrease the time step size of the Simulink model. With this method, sent data would be received sooner, then resent. This solution increases the required computational power in the target PC and, if time-dependent equations are used in the Simulink model, care must be used to ensure that the equations are solved over the correct time interval. Another avenue could be to use the 20

Original page 33

data to drve the model. If Smulnk provdes programmatc hooks for controllng when a model step s executed, the executon of each model step could be ntated whenever new data s receved. In other words, the model could be driven by the data, and the variability of the time lag would be eliminated. This possibility still needs to be investigated. 5.4 Execution of Dynamic Model on a Target pC with Shared Memory To ensure better synchronization of data between the Simulink model and LabVIEW application, the use of shared memory cards was evaluated. A VMIC 5565 shared memory board was installed on the PCI bus of both the host and target PCs. The shared memory boards transfer data by writing data to a specified partition on all boards in a network when data is written to a single board; therefore, when data is available on a single memory board, it is then instantaneously available on all the memory boards in the system. The boards are connected with fiber optic cable and allow for high data transfer rates. A second feature of the shared memory boards that helps with data synchronization is that the boards are capable of sending and receiving hardware interrupts. When using xPC Target, a Simulink model executes n real tme. The use of hardware nterrupts allows for an external sgnal to be used to dctate when the model should execute its next time step. Therefore, the Simulink model may be incremented by a single time step each time the LabVIEW application increases by a time step by sending an interrupt from the LabVIEW PC. With the neutronic models tested to date, the result at each time step was computed within approximately 1×10–4 seconds. These data were then available to the LabVIEW application fractions of a second later. With this implementation, real-time execution of the dynamic model may be achieved—relative to the base system time scale of ≈200 ms. Similar to the test performed with UDP communication, shared memory boards were used to send data from LabVIEW, read data with the Simulink model, and then send data back to LabVIEW. Through the use of network nterrupts, a neglgble tme lag was found between the sent and receved data in LabVIEW. Due to the high reliability of data transfer achieved with this method, this configuration is preferred to the others presented above. Additionally, the use of a separate PC for the execution of the dynamic model increases available computational capacity and allows for the implementation of larger, more complex dynamic models in the future. 5.5 Dynamic Model System Implementation An implementation of the coupled LabVIEW-Simulink system was built to demonstrate the feasibility of implementing the improved neutronics model on a target PC. The test system was built with the shared memory configuration described in section 5.4. Although shared memory was used as the data transfer mechanism between the LabVIEW and Simulink systems, an ethernet connection was necessary between the two systems to allow for download and control of the model on the target PC. A smple Smulnk model was bult to smulate the behavor of both the reactor knetcs and the thermal system that would normally be coupled to the electrically heated test system. Figure 14 shows the Simulink block diagram of the model. The first three blocks in the data path, shown in the model, (“5565 read,” “5565 unpack,” and “change endianess”) read data from the shared memory board and prepare the data for Simulink calculations. The value of reactivity, based on the current core temperature, read from the shared memory board is then sent to the reactor power subsystem, which contains the 21

Original page 34

point kinetics representation of the reactor dynamics. The new power level calculated from the PKEs solution at that time step is then prepared and written to the shared memory board. The reactor power is also used to compute a new structural temperature with the core thermal model subsystem. This temperature is also written to the shared memory. Init. 5565 Init. Data_In Endianess Id: 1 Unpack Read UD > Bytes=24 > 5565 Read 5565 Unpack Reactivity Target Scope Power Id: 2 Reactor Power Scope (xPC) 1 Change Endianess Power_In T_struct_Out Change Target Scope Scope (xPC) Change Endianess Pack Power Write Bytes=24 5565 Write T_Struct. Change Endianess 1 5565 Pack Core_Thermal_Model Figure 14. Simulink block diagram of the test dynamic model. The core thermal model subsystem, shown in figure 15, was created to allow for a demonstration of reactivity temperature feedback effects. The thermal model consists of two energy balances. The first energy balance is for the structure of the core and may be represented analytically (eq. (7)): 1 Tstruct = ⋅ ∫ PRx Mstruct ⋅ Cp,struct 22 − hc ⋅ A ⋅(Tstruct −− Tcoolant,avg) . (7)

Original page 35

1 Power_In T_Coolant_Avg 300 Convective Heat 500 Transfer Coeff. Structural Mass 400 s 300 Heat Transfer Area Specific Heat 10 Mass Flow Rate 50 Coolant Specific Heat s 800 Coolant Inlet Temperature –C– M_Coolant Coolant Mass 0 T_Out Display1 s 1 w D:1 1 S T_Struct w T_Struct_Out Integrator 0 Display2 s 1 w S T_Coolant_Avg Integrator1 2 Gain Fgure 15. Smulnk block dagram for the core thermal model. Dummy values for the coolant and structural masses and other material properties were used in the test dynamic model. The energy balance of the coolant inside the core was computed using the relationship in equation (8): 1 ∫ struct − Tcoollant,avg) − m C⋅p(Tout − Tn) . (8) Tcoolant,avg = ⋅ hc ⋅ A ⋅(T Mcoolant ⋅ Cp A simple LabVIEW application was built to test the functionality of the shared memory interface between the LabVIEW and target PCs. The LabVIEW application functions by reading temperature output from the thermal model then computing and sending a new value for the reactivity based on the new core temperature. The output of the model is then read and displayed by the LabVIEW application. Figure 16 shows the user interface to this program. The reactivity, power, and temperature are displayed on the waveform charts. The control labeled Inserted_Reactivity allows for the addition or removal of reactivity from the system; e.g., simulating a reactor control maneuver. For simulation purposes, the timed 23

Original page 36

loops in this application are set to run once every 25 ms, causing the Simulink model to run at ten times real time. The system appeared to behave as expected. The plots show a positive reactivity insertion followed by a negative reactivity insertion. Figure 16. LabVIEW application: VMIC test. The rate at which the power may be accurately updated depends on the speed at which data is made available. In the current test system, TC data is available at 5 Hz. The test system with reactivity feedback was smulated n Smulnk at varous data rates to demonstrate the model performance when the sample rate differs from the data rate. Figures 17–19 show the simulated response of the system to a reactivity insertion at t = 100 s. The model was run at a rate of 100 Hz. However, the reactivity feedback was sampled at varous slower sample rates to demonstrate the effect of runnng the model faster than the data was being updated. Figure 17 shows the reactivity as a function of time. The black line shows the model response when the data rate s the same as the model speed (Fs=0.1s). In the smulation, temperature rises soon after the reactivity insertion, causing a drop in reactivity soon after the initial increase. When the data sample rate (Fs) s slower than the model sample rate, the model wll compute the response using the last known value of reactivity. This causes the stepped shape of the curves in figure 17. When the model computes the time-dependent power response based on this constant 24

Original page 37

reactivity, an exponential increase is seen, causing the divergence between the black line and other lines at t = 100 s in figure 18. Although running the model at a faster sample rate than the data rate allows for a higher resolution time-dependent power function, there may actually be more error in the controlled power relative to the real, continuous system than if the power was held at a constant value until new temperature data were avalable. x 10–3 6 5 4 3 2 1 0 –1 –2 –3 –4 99 100 101 102 103 Time (s) Fs  1 s Fs  0.5 s Fs  0.25 s Fs  0.1 s 104 105 106 107 108 109 Figure 17. Simulated reactivity response to a reactivity insertion at various data rates. x 105 9 8 7 6 5 4 Power (W 3 2 1 0 99 100 101 102 103 Time (s) Fs  1 s Fs  0.5 s Fs  0.25 s Fs  0.1 s 104 105 106 107 108 109 Figure 18. Simulated power response to a reactivity insertion at various data rates. 25

Original page 38

865 864.5 864 863.5 863 862.5 Temperature (K 862 861.5 861 860.5 Fs  1 s Fs  0.5 s Fs  0.25 s Fs  0.1 s 99 100 101 102 103 104 105 106 107 108 109 Time (s) Figure 19. Simulated temperature response to a reactivity insertion at various data rates. 26

Original page 39

5.6 Selection of a Computing Environment for Dynamic Testing As previously discussed, several methods can be adopted to improve the dynamic testing capability at the EFF–TF. After evaluating several potential configurations, separate model and control computers were selected over a configuration in which Simulink and LabVIEW would be run on the same PC. At present, the most desirable option that can be readily implemented will incorporate Simulink and xPC Target software and will use shared memory boards for data transfer between PCs, providing the simplest and most reliable solution. This configuration minimizes complexity within the LabVIEW program that would be required to directly interface with Simulink on a single PC. Additionally, using a single processor places a fundamental limit on the complexity that can be added to the Simulink model while maintaining reasonable calculation rates; hence, the selected configuration offers the greatest capacity for future model expansions. The capability of updating data acquisition and control information at 100-ms time steps is currently being investigated. 27

Original page 40

1. CoNClUSIoNS Integrated system tests with simulated reactivity feedback on the SAFE–100a demonstrated the potential benefits to system operation with simulated neutronic response based on measured system conditions. By simulating dynamic system response, a better understanding can be gained with respect to system integration issues, system response times and response characteristics, and overall system design potentially allowing for design improvements prior to large fiscal and manpower investments into a nuclear test program. Plans are underway to implement similar test methodology on the DDG at the EFF–TF. This implementation will incorporate higher fidelity neutronic models and will utilize more detailed measurements of the system state, as described in this TM. Necessary system modifications could be completed wthn ≈2 mo with dynamic tests commencing shortly thereafter. 28

Original page 41

REFERENCES 1. Bragg-Sitton, S.M.; and Forsbacka, M.: “Application of a Virtual Reactivity Feedback Control Loop n Nonnuclear Testng of a Fast Spectrum reactor,” Proceedngs of the International Congress on Advanced Nuclear Power Plants (ICAPP–2004), American Nuclear Society Order No. 700306, Pttsburgh, PA, pp. 2259–2268, 2004. 2. Bragg-Sitton, S.M.: “Heat Pipe Reactor Dynamic Response Tests: SAFE–100a Reactor Core Prototype,” Proceedings of the Space Nuclear Conference (SNC–2005), American Nuclear Society, San Dego, CA, pp. 781–789, 2005. 3. Godfroy, T.J.; Kapernick, R.; and Bragg-Sitton, S.M.: “Thermally Simulated 32 kW Direct-Drive Gas-Cooled Reactor: Design, Assembly and Test, ” Proceedings of Space Technology and Applications International Forum (STAIF–2004), M. El-Genk (ed.), AIP Conference Proceedings 699, New york, pp. 757–763, 2004. 4. Hetrck, D.l.: Dynamics of Nuclear Reactors, The University of Chicago Press, Chicago, IL, 1971. 5. Wright, S.A.; and Lipinski, R.J.: “Pin-Type Gas Cooled Reactor for Nuclear Electric Propulsion,” n proceedngs of Space Technology and Applications International Forum (STAIF–2003), M. El-Genk (ed.), AIP Conference Proceedings 654, New York, pp. 408–419, 2003. 6. VanDyke, M.K.; Houts, M.G.; Pedersen, K.; et al.: “Phase I Space Fission Propulsion System Testng and Development Progress, ” n proceedngs of Space Technology and Applications International Forum (STAIF–2001), M. El-Genk (ed.), AIP Conference Proceedings 552, New York, pp. 837–842, 2001. 7. Stinson-Bagby, K.L.; and Fielder, R.S.: “Fiber Bragg Gratings for High-Temperature Thermal Characterization,” in proceedings of the International Congress on Advanced Nuclear Power Plants (ICAPP–2004), American Nuclear Society Order No. 700306, Pittsburgh, PA, pp. 2288–2294, 2004. 29

Original page 42

Form Approved REPORT DOCUMENTATION PAGE OMB No. 0704-0188 Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden, to Washington Headquarters Services, Directorate for Information Operation and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503 1. AGENCY USE ONLY (Leave Blank) 2. REPORT DATE June 2007 4. TITLE AND SUBTITLE 3. REPORT TYPE AND DATES COVERED Techncal Memorandum 5. FUNDING NUMBERS Application of Simulated Reactivity Feedback in Nonnuclear Testing of a Direct-Drive Gas-Cooled Reactor 6. AUTHORS S.M. Bragg-Sitton and K.L. Webster 7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) george C. Marshall Space Flght Center Marshall Space Flght Center, Al 35812 9. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) Natonal Aeronautcs and Space Admnstraton Washington, DC 20546–0001 11. SUPPLEMENTARY NOTES 8. PERFORMING ORGANIZATION REPORT NUMBER M–1188 10. SPONSORING/MONITORING AGENCY REPORT NUMBER NASA/TM—2007–214959 Prepared by the Propulsion Systems Department, Engineering Directorate 12a. DISTRIBUTION/AVAILABILITY STATEMENT Unclassified-Unlimited Subject Category 31 Availability: NASA CASI 301–621–0390 13. ABSTRACT (Maximum 200 words) 12b. DISTRIBUTION CODE Nonnuclear testing can be a valuable tool in the development of an in-space nuclear power or propulsion system. In a nonnuclear test facility, electric heaters are used to simulate heat from nuclear fuel. Standard testing allows one to fully assess thermal, heat transfer, and stress related attributes of a given system but fails to demonstrate the dynamic response that would be present in an integrated, fueled reactor system. The integration of thermal hydraulic hardware tests with simulated neutronic response provides a bridge between electrically heated testing and full nuclear testing. By implementing a neutronic response model to simulate the dynamic response that would be expected in a fueled reactor system, one can better understand system integration issues, characterize integrated system response times and response and response characteristics, and assess potential design improvements with a relatively small fiscal investment. Initial system dynamic response testing was demonstrated on the integrated SAFE–100a heat pipe cooled, electrically heated reactor and heat exchanger hardware. This Technical Memorandum discusses the status of the planned dynamic test methodology for implementation in the direct-drive gas-cooled reactor testing and assesses the additional instrumentation needed to implement high-fidelity dynamic testing. 14. SUBJECT TERMS 15. NUMBER OF PAGES nonnuclear testing, reactivity feedback, dynamic simulation 40 17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION OF REPORT OF THIS PAGE Unclassified Unclassified NSN 7540-01-280-5500 30 16. PRICE CODE 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT OF ABSTRACT Unclassified unlmted Standard Form 298 (Rev. 2-89) Prescribed by ANSI Std. 239-18 298-102

Original page 43

Original page 44

National Aeronautics and Space Administration IS20 George C. Marshall Space Flight Center Marshall Space Flight Center, Alabama 35812