TASK 1. THERMIONIC RESEARCH
The work contained within this task was in support of Task 3, Thermionic Fuel Element Development. For background information regarding this task, see the first part of Task 3.
1.1 OUT-OF-CORE TESTING (G. O. Fitzpatrick, M. K. Yates)
The objectives of this subtask included the operation of OC-A4, a diode using a graphite-cesium sorption reservoir, and the performance mapping of an out-of-core mockup of an in-core F series cell. During the past work periods, the mapping of OC-A4 was completed and a total of 26,296 test hours was accumulated.
1.1.1 Mark VIIA OC-A4 Test
The test of Mark VIIA OC-A4 included performance mapping with a liquid cesium reservoir followed by performance mapping with a graphite sorption reservoir, life testing for 23,772 hours, and testing in association with a remotely programmable driving supply for 1426 hours. A post-operational examination was conducted after a total of 26,296 hours had been accumulated.
The performance of the cell dropped slowly, totaling approximately 5% during the first 15,229 hours of life testing. The loss was attributed largely to emissivity changes of the collector due to contamination with carbon from the graphite reservoir. Low power diagnostic measurements at 15,229 hours were followed by a loss of 35% in output power, attributed to emitter contamination with titanium from the final closure braze. The diode was operable when shut down at 29,296 hours.
The post-operational examination showed that carbon transport from the graphite reservoir to the cell had occurred. Carbon or carbide layers 3-5# thick were visible on all the niobium surfaces of the cell. No carbide or carbon was found on the emitter. The reservoir graphite had partially delaminated in a manner typical of samples with high cesium loading, but did not appear otherwise damaged. The final closure braze had largely evaporated, and the cell assembly separated at that point during handling. Further details concerning the test are included in Appendix A.
1-1
1.1.2 Laboratory F Cell
The design of the electrically heated F series converter was completed and fabrication was initiated. The design is discussed in Section 3.1 and fabrication progress in 3.2.
1.2 DIAGNOSTIC METHOD DEVELOPMENT (G. O. Fitzpatrick)
The objectives of this subtask were to develop improved methods of’ determining the operating conditions of in-core thermionic devices. This objective was achieved through the continuous incorporation of new information from in-core thermionic tests, and following the theoretical calculations using applicable diagnostic correlations.
1.2.1 Emitter Temperature Correlations
Prior to the operation of TFEs in-pile, most thermionic cells incorporated thermocouples to allow estimates of emitter temperatures. While these frequently degraded rapidly, they provided good estimates of initial operating conditions. Both E and F series TFEs, however, are fabricated without such thermocouples. In addition, they normally consist of a series string of cells, each operating with different emitter temperatures and output powers. The purpose of this subtask has been to devise ways of estimating individual emitter temperatures within these TFEs.
Emitter temperature estimation has been approached in two ways, the first based on a knowledge of the TFE’s operating environment and its thermal characteristics and the second based on its electrical characteristics.
The second approach has been used to establish the initial operating condition of a TFE. Three aspects of the TFE’s electrical characteristics are used: its output voltage in the obstructed mode (which gives an average emitter temperature), the device 's open circuit voltage (which relates the average emitter and collector temperature), and the cell current at ignition
1-2
(which provides individual emitter temperatures). Details of these correlations and their use are given in Ref.’s 1 and 2. Ignition current measurements have proven most useful, providing as they do individual estimates of’ emitter temperatures.
The use of electrical characteristics in some operating conditions is not practical (ignitions disappear above 1900°K for example), and they are also subject to variations due to changes within a cell-gas in the inter-electrode space, work function changes, emitter swelling, etc. For those reasons after initial operating temperatures were established, further temperature estimates were based on knowledge of the cell’s operating environment and thermal characteristics. The mapping data acquired with Mark VIIA OC-A4 (Ref. 2) was used to estimate the emitter temperature given the operating current density and electrode input power. The equation used was:
QE = 3.781 x 10-5Te2+(1.0245x10-3J-1.0104x10-1) Te+0.613J+71.88
where QE = Input power (Watts/cm2)
Te = Emitter temperature (°K)
J = Current density (Amps/cm2)
QE in an operating device was estimated to be the sum of the collector heat flux and electrical electrode power in a cell. An empirical expression for cell collector heat flux was developed.
QC = A (Tsh + To) b (Tsh – To) / (1 – c Tsh)
QC where is the collector heat flux
A is empirically determined
Tsh is the sheath temperature (°K)
To is the pit water temperature (°K)
b is a constant determined by the containment gas
c is a correlation parameter
1-3
The correlation parameter was determined for each cell through use of ignition current temperatures determined at various input powers early in a test.
The core location of a test and the operating reactor power give further information on the input power expected and hence the emitter temperatures. For this reason an effort has been made to map the TITR core and to develop a correlation that could be used to estimate the input power in various types of thermionic devices. This is also necessary in order to specify the enrichments of devices in fabrication and specify the core locations for devices going on test.
A linear correlation was chosen as a first-order model. This correlation has been derived for the F series TFE design, which features 40 or 80 mil (1 or 2 mm) clad tungsten emitters and a fueled length of 2 inches (51 mm) and has the following form:
Q = C x fR x fz x fr x fu x ff x fs
where Q - total input power in watts
C - a constant, derived from measured in-pile data (C = 3990 W)
fR - reactor power factor, the reactor power level divided by the value used when C was evaluated = 1.0 at a reactor power of 1.59 Mw)
fz - axial power factor: A composite axial power profile was z derived from axial mapping taken in the core using FC-l, a self-powered neutron detector, and operating TFEs. This shape is shown in Figure 1-1. The mapping data showed this shape was approximately independent of reactor power level, enrichment, core location, and control rod position. However, the position of the flux centerplane relative to that of the TRIGA fuel centerplane does depend on control rod position. For most of the data with thermionic devices at power levels of ~ 1.5 MW the control rods are almost
1-4
out of the core. The flux centerplane was assumed to be ~ 0.5 inch (1.3 cm) below the fuel centerplane in those cases.
fr - radial power factor: A small amount of radial power data was obtained during the FC-l mapping. This data is shown in Fig. 1-2 where the peak power obtained during the East-West traverse and in the DE flux trap is shown. This data shows the extent to which the graphite pack skews the flux distribution, but not enough data was taken to allow the power in other parts of the core to be estimated. However, two-dimensional calculations of the core power distribution are performed in order to follow FLIP fuel element burnup. The results of the calculations are maps showing the power generated in each fuel element. One of these core maps is shown in Fig. 1-3. The values given in this particular map have been used as the radial power factor. If a value is not available for a given position, the average of the six surrounding positions is used.
fu - Uranium 235 loading and clad thickness factor: This factor was calculated using one dimensional codes assuming that the fuel smeared over the area defined by the 1.100 inch outside diameter by 0.040 inch or 0.080 inch tungsten clad emitter. The results in terms of relative power as a function of grams U-235 per 2-inch fuel slug are given in Fig. 1-4.
ff - fuel factor, depends on fuel used
1.00 for UO2
0.97 for 90 UC-10ZrC-4W or UC-4W
fs - shim factor, depends on neutron shielding or shims added to device.
The use of this correlation in practice indicates it is accurate to within the experimental uncertainty in input power estimates based on the electrical characteristics of a TFE (± 10% based on ± 50°K errors in emitter temperature estimates and ± 30% errors in cell end loss estimates).
1- 6
Further refinements in the accuracy of the values could be made by reducing emitter temperature errors below ± 50°K, making more exact end loss calculations, and using the measured data to refine Figure 1-3. Another source of uncertainty which could be reduced is that involving relative control rod heights. In much of the testing, the rods were not kept banked; they were usually set up to favor the operation of a 6F TFE. However, this mode of operation skews the power profile in the core and changes the relative worths of the various core positions, an effect that is not taken into account in the 2D calculations whose results are given in Figure 1-3. Beginning with the test of TFE 6F4, the control rods were kept banked.
1-10
1.2.2 TFE Performance Determinations
The performance of a TFE has been described in terms of a relative power number, Pr, which is the ratio of the corrected total electrode voltage produced by a TFE to that predicted by CPOP (Ref. 3) at a specified "standard" emitter temperature and current density. The following corrections are made to obtain the corrected TFE voltage.
1. Adjust voltage upward to make up losses due to off-optimum collector temperatures. (Correction based on initial operating measurements and collector optimum curves measured for OC-A4, Ref. 4)
2. Correct the voltage to that which would be obtained if the spacing were 10 mils (Refs. 5 and 6).
3. Calculate the difference between the voltage obtained after corrections (1) and (2) and that predicted by CPOP for the TFE’s actual emitter temperatures and current density.
4. Assume this voltage difference calculated in (3) would be the same at the "standard" operating conditions and subtract it from the CPOP predicted voltage under those conditions.
Changes in performance can result from a number of causes including electrode emissivity and work function changes, inter-electrode spacing changes, inter-electrode gas, poor thermal bonds in the collector structure, shorts or parasitic electrical discharges, and input power. Determination of sources of change are made by correlating data from IV curves, cesium optimums, collector optimums, input power based on sheath temperatures and reactor location, TFE response times, voltage probe changes, and neutron radiographs.
1-11
Recently a new technique of observing individual cell cesium optimums was developed. The electrode power and the efficiency of each cell varies as the cesium temperature varies, with maximum efficiency at the optimum cesium reservoir temperature. The efficiency of each cell is reflected in the sheath temperature opposite it, and hence measurements of these temperatures can be reduced to electrode voltage changes. An example of the results of reducing data taken during cesium optimum measurements of TFE 6F4 are shown in Figure 1-5. It is clear that the optimum temperature for cell 2 is significantly higher than cells 1, 3, 5, and 6, and cell 4 appears to have an unusually low optimum temperature. Observations of individual optimums such as these can provide information on electrode work function, emitter temperature, and current density changes in specific cells within a TFE.
1.2.3 Performance Map Correlations
The performance map obtained with Mark VITA OC-A4 has been used to estimate thermal performance changes in in-pile tests. Thermionic performance has been compared to that predicted by CPOP at an inter-electrode spacing of 10 mils (0.25 mm) corrected for other spacings according to slopes derived from TECO out-of-pile cell test data (Ref. 7).
A reference performance data set generated by NASA-LeRC was received in December (Refs. 8 and 9) but its comparison with existing performance maps has not been accomplished.
1-12
REFERENCES
1. "Development of a Thermionic Reactor Space Power System," Summary Report for Period March 1, 1971 through June 30, 1971, USAEC Report GA-B11043, Gulf General Atomic.
2. "Development of a Thermionic Reactor Space Power System," Summary Report for Period July 1, 1971 through October 31, 1971, USAEC Report GA-A11061.
3. Gietzen, A., "CPOP. A Program for Thermionic Converter Performance Analysis and Optimization, USAEC Report GA-8873, Gulf General Atomic, April 30, 1969.
4. Fitzpatrick, G. O. and Yates, M. K. "The Test of a Thermionic Converter with a Graphite Sorption Cesium Reservoir," USAEC Report GA-9526, Gulf General Atomic, March 1970.
5. Wright, M. Christina, "Performance for several Variable Spacing Thermionic Converters," USAEC Report TE-4150-62-72, Thermo Electron Corp., 1972.
6. Lieb, D., and Rufeh, F., "Thermionic Performance of CVD Tungsten Emitters with Nb and Mo Collector Materials," USAEC Report TE-4137-134-71, Thermo Electron Corp., May 1971.
7. "Development of a Thermionic Reactor Space Power System," Summary Report for Period March 1, 1972 through June 30, 1972, Draft, USAEC Report GA-A12397, Gulf General Atomic.
8. Sockol, P. M., "Correlation of Experimental Performance Data for a CVD Tungsten Niobium Plasma Thermionic Converter," NASA-TM (to be published).
1-14