1. Introduction
Accidents related to load rejection are an important technical problem faced by nuclear power plant interfaces and electricity grids, as they may challenge the grid reliability and nuclear safety. Generally, a full load rejection could be caused by either the sudden loss of external power demand from the grid, or by accidental opening of the breaker connecting the generator and the grid [1]. A load rejection test is, therefore, always listed as compulsory during the commissioning phase of a nuclear power plant [2]. The main purpose of this type of commissioning test is to verify the interconnection control and interlock protection among the reactor, the turbine, and the generator.
The HTR-10 is an experimental modular high temperature reactor (HTR) located at Tsinghua University, and is the first of its kind in China. As a prototype power plant, it is configured with a 10 MWth pebble bed reactor, once-through steam generator, and 3 MWe steam turbine generator. This type of plant could be used for electricity generation and could provide heat to the district. When the generator and the grid are decoupled, the power supply by the small steam turbine generator is not stable, so the house load operation was not considered by the design [3]. After the reactor ran continuously for 72 h with full power in the commissioning stage, a full load rejection test was conducted by manually turning on the breaker of the turbo generator. Following the initiation of the load rejection, the generator was disconnected from the grid, and the turbine was subsequently tripped due to the closure of the main steam valve. Meanwhile, the reactor was scrammed because the protection system was triggered by the signal from the conventional island, and other engineering safety measures were also implemented. By doing this, the interconnection control and interlock protection of the HTR-10 plant could be fully proven with the test results. In addition, the transient reactor parameters were recorded, and the data was used to validate and benchmark the calculation models originally adopted in the design of the reactor.
In this study, we simulated an HTR-10 full load rejection test. The THERMIX system analysis code was used for the post-test analysis. This code was originally introduced by FZJ in Germany. After adding certain analysis functions and modifying the source code, an in-house version of THERMIX was developed and applied in the design of the HTR-10 and HTR-PM. The main analysis functions of THERMIX are based on the neutron kinetics of the core, heat conduction calculations of both the solid materials in the reactor and a single fuel pebble, convection calculations of the coolant in the reactor, and a fluid flow calculation of the coolant in the primary circuit [4, 5]. Furthermore, the code can couple specific modules for the chemical reaction between graphite and air/water, and for the steam generator dynamics [6, 7]. In the past, the THERMIX code was benchmarked in various ways, including a series of out-of-pile experiments conducted at FZJ, simulation tests of the AVR loss of coolant accident, as part of extensive experimental programs of the THTR-300, in the benchmarking calculations for the GT-MHR, and so on [8-11]. In general, in these past studies, THERMIX produced satisfactory results in comparison with the experimental results or with those of other codes. Our modified version of the code was used by China to participate in several coordinated research projects (CRPs) initiated by the IAEA. For example, CRP-3 was used to research heat transport and afterheat removal for gas-cooled reactors, and CRP-5 was used for the evaluation of HTR performance. Comparisons of both code-to-code and code-to-experiment showed that China’s analysis results were satisfactory and acceptable [12-14]. Besides that, the HTR-10 measured data made it possible to include more prototypic reactor conditions in the code validation. THERMIX has been tested against most of the tests conducted on this reactor, such as a steady-state full power operation test, a power ascension test, and safety demonstration tests [15-18]. To build upon previous validation work, in this study we present the post-analysis results of the HTR-10 full load rejection.
The simulation used this study is based on the actual test conditions of the full load rejection, and focuses on the reactor transient response to the test initiation. As a modular HTR, the HTR-10 has inherent safety features, and among these, the self-acting residual heat removal capability is particularly important. The important thermal-fluid phenomena that occurred during the heat transfer process after a reactor shutdown was analyzed (e.g., natural convection of helium established in the core and the temperature redistribution of the reactor). There were several thermocouples installed in the reactor, and temperature curves were reproduced from these measuring points and were then compared with real data measured by the thermocouples. The simulation and test results agreed, which demonstrated that the code is capable of simulating reactor transient behaviors during a HTR-10 full load rejection test. The maximum possible fuel temperature is the most important safety feature of the reactor. Based on the simulation, there was always a large temperature margin between this parameter and the safety limit, due to the low residual heat level of the core, the rapid cooling effect of the reactor based on its geometry, and the heat storage capability provided by the heat capacity of the ceramic internals.
In Section 2, the primary and secondary HTR-10 systems are briefly introduced, and the test conditions of the full load rejection are described in detail. After describing the analysis method and calculation models, Section 3 presents the predicted temperature response of the HTR-10 during the full load rejection test. The differences between the code results and the test data are then discussed. At last, a summary of the full load rejection test and some conclusions related to the code capability are made in Sect. 4.
2. The HTR-10 load rejection test
As mentioned in Sect. 1, the HTR-10 plant mainly consists of a pebble bed reactor, a once-through steam generator, and a steam turbine generator, as shown in Fig. 1 [19]. The reactor pressure vessel and the steam generator pressure vessel were arranged side by side, and were connected by another horizontal pressure vessel where the hot gas duct was installed. The thermal power of the reactor was rated to 10 MW, with an average power density of 2 MW/m3. When the reactor was under full power operation, the primary circuit had a helium pressure of 3 MPa and coolant inlet/outlet temperatures of 250/700 oC. In the secondary circuit, the inlet temperature and mass flow rate of the feedwater were 104 oC and 3.49 kg/s, respectively. Superheated steam of 4 MPa and 440 oC could be generated.
The pebble bed reactor core contained about 27000 spherical fuel elements. The pebble bed was cylindrical with a cone shaped bottom. The equivalent core diameter was 1.8 m, with an average height of 1.97 m. The diameter of one of the fuel balls was 6 cm. Each fuel element was composed of two parts, the inner fuel zone with a diameter of 5 cm, and the outer graphite shell with a thickness of 5 mm. Three kinds of graphite reflectors were arranged to enclose the pebble bed core, and were categorized as the top reflector, the bottom reflector, and the side reflector. In the top reflector, there was a plenum whose function was to collect cold helium before it flowed into the core. Since the coolant at the core outlet may have a non-uniform radial temperature distribution, another plenum was placed in the bottom reflector to sufficiently mix the hot helium with different temperatures. In the side reflector, various channels were drilled, for example, control rod guiding channels, small absorber ball channels, and cold helium flow channels. Carbon bricks were arranged around the graphite reflectors. The carbon material thermally insulated the core to reduce heat loss, and also shielded radiation from hitting the reactor pressure vessel (RPV) to protect it from neutron irradiation. All of the reactor internals, including the ceramic metallic internals, were supported and enveloped by the RPV. Arrows in Fig. 1 show the coolant flow direction under normal operating conditions. The cold helium was driven by the pressure head from the helium circulator, and first entered the bottom cavity of the RPV. The coolant then passed through the metallic structures in the bottom cavity and was distributed by the cold helium flow channels, where the helium flowed upwards until it merged into the upper plenum in the top reflector. After that, the mainstream flowed through the pebble ped in a top-down direction, and was thus heated by the fuel balls in the reactor core. Subsequently, the hot helium flowed from the core outlet to the lower plenum, where the helium achieved an average temperature of 700 oC. After completing the heat exchange with the secondary water in the steam generator, the primary coolant was pumped by the helium circulator again, and moved back into the RPV to start a new circulation. A thermal measurement system was designed to monitor parameters of concern under normal operation and after reactor shutdown [20]. These parameters included primary pressure and temperatures of different components and positions, e.g., the ceramic internals and metallic internals, the RPV, the cavity concrete, and the cavity cooling system. Figure 1 gives the distribution of the thermocouples in the reactor [21]. TR1 to TR6 are the thermocouples in the top internals and were arranged in two symmetrical columns. SR1 to SR6 were installed in a row in the upper part of the side internals, while SR7 to SR12 were in another row in the lower part of the side internals. In the bottom internals, the thermocouples were arranged in three main parts. BR1 to BR4 were designed to measure the temperature of the bottom reflector. They were arranged in a row and were symmetrical pairwise. FD1 to FD6, which were in two symmetrical columns, were fixed to surround the fuel discharging tube. CB1 to CB8 were installed in two symmetrical columns in the bottom carbon brick.
Figure 2 illustrates the closed water-steam cycle in the secondary circuit of the plant. The HTR-10 steam generator works in once-through operation. There was a total of 30 modular assemblies for the heat transfer between the primary helium and the secondary water/steam. Each assembly was made up of a helical tube bundle and included three sections, the preheating section, the evaporation section, and the superheat section [22]. The feedwater was driven by an electric pump and was heated by the primary helium in the steam generator. When the reactor was operated under 30% to 100% rated power, the outlet steam could be supplied to the turbine for electricity generation. When the power was lower than 30% rated power, the steam was bypassed via a special startup and shutdown loop, due to technical issues related to two-phase flow stability and steam quality.
Before the full load rejection test, the HTR-10 operated in steady state. The power level, hot helium temperature, and primary pressure were 10 MWth, 700 oC, and 3 MPa, respectively. The superheated steam had a pressure and temperature of 3.45 MPa and 430 oC, respectively. The test was initiated by manually opening the breaker of the generator. After the initiation, the main steam valve of the turbine was closed immediately, because of the instantaneous loss of the external load. The emergency shutdown was triggered after the reactor protection system received the turbine trip signal from the conventional island [23]. The reactor scram was actuated by the first shutdown system of the HTR-10 (i.e., the control rod system). In the test, ten control rods fell into their guiding channels by gravity within 7 s. At the same time, the helium circulator automatically stopped by the switch-off of its transducer, and any forced cooling of the core was lost. The blower baffle was also closed within 15 s, hence, the primary circuit was shut off and the possibility of natural circulation in the primary circuit was eliminated. Through these protective actions, the primary pressure boundary (i.e. the reactor pressure vessel, the steam generator pressure vessel, and the hot gas duct pressure vessel) avoided contact with the hot helium, thus ensuring its integrity after the reactor scram. The feedwater pipeline and the main steam pipeline were also cut off by their isolation valves, hence, the isolation of the secondary circuit was implemented within 9 s.
3. Transient temperature analysis
In this study, the THERMIX code was used to investigate the reactor response to the HTR-10 full load rejection test. This code has been widely used for the thermal-hydraulic analysis of pebble bed HTRs. The code can simulate both steady-state operation and transient conditions. For the transient analysis, point reactor model neutron kinetics were adopted by THERMIX, and six different groups of delayed neutrons were considered. The code could simulate the reactor scram process by input of external negative reactivity introduced by the neutron absorbers, and it could evaluate any residual heat generated by the fission products by internally calculating their kinetic reaction equations. Using transient heat conduction equations, the code calculated both the two-dimensional temperature profiles of the reactor itself and the temperature distribution of the fuel balls in a one-dimensional radial direction. The material properties, e.g. heat conductivity and heat capacity, were dependent on temperatures and, therefore, changed with time. Due to the large heat capacity of the core and the ceramic internals in HTRs under transient conditions, the change of the coolant parameters was relatively slow. In a given sufficiently small time step, the gas conditions could be considered stable. The steady-state conservation equations of mass, momentum, and energy of the coolant in the reactor, were, therefore, adopted to simulate the flow conditions in the core and other flow channels. This type of gas convection model is called a quasi-stationary model, where the temperatures of the solid structures were time-dependent and the pebble bed was modeled by a homogeneous porous medium. According to the structure and the geometric parameters of the HTR-10, the model calculations for gas convection were established in two-dimensional axisymmetric (r, z) geometry, as shown in Fig. 3. The radial coordinate started from the core centerline and the axial coordinate started from the top surface of the core with a downward positive direction. The units of the radius and height were in centimeters.
The HTR-10 full load rejection test was analyzed based on the initial conditions and the event sequence described in Section 2. The time duration of the simulation and the test were 6 h and 5.33 h, respectively, and both included the first hour of steady-state operation. The power curve obtained by the code is presented in Fig. 4(a) and shows a sharp drop in the reactor power following the initiation of the test, due to the reactor scram. In a specific transient condition or accident scenario, the maximum fuel temperature was the safety relevant parameter of most significance, and was set as one of the acceptance criteria for accident analysis. The temporal maximum fuel temperature in the full load rejection test was calculated by the code and given in Fig. 4(b). The peak value appeared at the initiation of the test and was 997 oC. The safety limit is 1620 ◦C; there is still a large margin between the maximum fuel temperature and the limit.
After the reactor scram, the self-acting residual heat removal became the only effective means for residual heat dissipation from the core [24]. This is a key safety characteristic of modular HTRs, and is guaranteed merely by natural mechanisms, i.e., heat conduction, natural convection, and heat radiation. The effect of self-acting residual heat removal in the post-scram heat transfer of the HTR-10 in the full load rejection test is briefly explained as follows.
After the test initiation, the helium circulator stopped and the blower baffle closed, hence, the HTR-10 primary circuit was cut off and the coolant in the RPV became stagnant. The system pressure was, however, maintained at 3 MPa, hence, the coolant density was still relatively high. Meanwhile, the temperature gradient between the central region and the peripheral region of the core, which was hundreds of degrees, was still quite large. A significant buoyancy effect was, therefore, induced by the high-density helium and temperature differences. The consequence of this buoyancy effect was the establishment of a natural convection loop of helium in the reactor flow regions, such as the pebble bed, the helium plenums, and various helium flow channels. Arrows in Fig. 5 indicate the coolant flow direction at different stages of the test, and the dotted lines represent the pebble bed configuration. It can be observed that, compared with the helium forced circulation under steady-state operation, the natural convection after the primary circuit cut-off induced a reverse core flow direction. In this convection loop, the heated helium first went upwards to the central part of the pebble bed, and then flowed downwards in the outer region of the core and helium channels. Finally, it returned to the bottom region of the core.
This natural convection, together with heat conduction and heat radiation, provided the natural mechanism for the residual heat dissipation of the reactor. The heat transfer started from the core and moved to its surrounding internals and the outer RPV. From the RPV outer surface, the heat further dissipated, mainly by heat radiation to the reactor cavity cooling system (RCCS), whose function was to actually protect the RPV and the cavity concrete from overheating. During the residual heat removal process, energy was effectively transferred to the following regions: the upper and outer regions of the pebble bed, the ceramic and metallic internals above the core, the side reflector and carbon bricks, and so on. The result of the energy transportation was significant temperature redistribution in the reactor. For example, from Fig. 5, it is obvious that 5 h after the test initiation the high temperature zone of the core moved to the upper part (10-90 cm in Z axis) from the lower part (144-295 cm in Z axis), but the temperature range of the hot zone decreased from 800-1000 oC to 600-700 oC in 5 h. This was a result of the decreasing heat decay production level and the sufficient heat dissipation of the slender reactor.
As mentioned before, there were several thermocouples installed in the reactor internals (top, side, and bottom) and fuel discharging tube. In this study, the temperature transients of these components were calculated. For the purpose of validating the code and models, the calculated temperatures of these measuring points were compared with the test results. Since the two-dimensional code could not provide the azimuthal temperature distribution, one temperature curve obtained from the code was compared with two measured curves, if there were two symmetrical thermocouples that had the same (r, z) coordinates, but were in different circumferential positions.
Figure 6 gives the temperature curves of the top internals. The location of TR1 and TR4 was above the top surface of the pebble bed. Due to natural convection after the test initiation, there was an upward high temperature helium flow stream, causing TR1 and TR4 to be directly heated in the first two hours. The temperatures at these two positions then began to decrease due to decrease in decay heat. The code reproduced this temperature change with a maximum deviation of 100 oC. The shift of the core hot zone demonstrated in Fig. 5 caused a temperature increase for TR2 and TR5, and TR3 and TR6. THERMIX correctly simulated the variation tendency of these four thermocouples, but underestimated of the temperature increments.
From Fig. 5, it can be seen that the radial temperature difference in the side internals was always quite large during the test, hence, the lateral heat transport dominated the transient response of thermocouples. Figure 7 shows the temperature of the thermocouples in the side internals, including the upper and lower parts. In the upper part, SR6 and SR5 were closer to the core, therefore the rapid power drop after the reactor scram caused the temperatures to decrease. However, due to the heating effect of natural helium convection on the upper core region, SR6 and SR5 temperatures rose again to almost stable values. Although SR12 and SR11 were also near the core, their temperatures continued to decline before becoming constant. In addition to the power reduction, the main reason was the cooling effect of the natural helium convection on the lower part of the core. In regards to SR2-SR4 and SR8-10, their temperature increases were different because of the heat storage in the reflector and the carbon brick, where these thermocouples were installed. SR1 and SR7 were installed to monitor the temperature of the metallic core vessel, and were, therefore, close to the inner wall of this component. In these positions, the outward heat dissipation exceeded that transported from the inner ceramic materials, hence, a decrease in temperature from these two thermocouples could be expected. For the side internal thermocouples, the temperatures generated by the code were in good accordance with the test values. The largest discrepancy was 40 oC, and occurred at SR11.
The previously mentioned afterheat reduction and cooling effect on the core bottom can also be seen by the thermocouple temperatures in the bottom reflector, i.e., BR1 and BR3, BR2 and BR4, in Fig. 8(a). The calculated temperatures at these four positions after the test initiation corresponded well with the measured ones, despite a maximum difference of 102 oC, which occurred at BR2 at the beginning of the test.
As for the fuel discharging tube, its upper (FD3 and FD6) and middle part (FD2 and FD5) temperatures lowered quickly, as these regions are near the core bottom, and are, therefore, influenced by core cooling. Due to the sensible heat transferred downwards, the lower part (FD1 and FD4), however, experienced a temperature increase before cooling. Figure 8(b) indicated that the calculated temperature curves of the three pairs of thermocouples got closer and closer to the corresponding test curves, hence, the code results can be considered reasonable and acceptable. The biggest calculation deviation was 72 oC before the test initiation at FD2.
4. Conclusion
As an important commissioning test item, a full load rejection test was performed on the HTR-10 by manually disconnecting the generator from the grid after the reactor continuously ran for 72 hours at the commissioning stage. Following the initiation, all of the protection actions were actuated as planned, including the turbine trip, the reactor scram, and other engineering safety measures, and we were able to demonstrate the interconnection control and interlock protection of the HTR-10 plant.
A modified version of the THERMIX code was adopted for the post-test analysis of the reactor transients. After the reactor shutdown, a natural convection loop of helium formed in the flow regions of the reactor, because of a buoyancy effect caused by the relatively high helium density (under 3 MPa) and the large thermal gradient in the reactor. The natural convection, together with heat conduction and heat radiation, caused self-acting residual heat removal from the core to its adjacent internals. As a result, obvious temperature redistribution occurred in the reactor. Due to the shift of the heat load, the upper region of the reactor heated up, while the bottom part of the core cooled down. The maximum fuel temperature was the most important safety-relevant parameter for the HTR-10, under transient conditions. In the test the temperatures never came close to the safety limit of 1620 °C.
There was a set of thermocouples installed in the reactor internals (top, side, and bottom). The temperatures of these thermocouples was calculated by the code. For the purpose of benchmarking the code itself and the analysis models, the code results were compared with the test data. In general, the code reproduced the correct transient temperature tendencies for the measuring points in the reactor internals. Most of the calculated temperature curves were satisfactory or were in acceptable agreement with the measured data. The maximum relative deviation appeared at TR1 and TR4 (24.8 %, 100 oC), while the largest absolute deviation occurred at BR2 and BR4 (102 oC, 11.6 %). These deviations could be preliminarily attributed to physical property parameters of the materials utilized by the code, e.g., the specific heat capacity of the ceramic materials. The complex geometry of the bottom reflector was another challenge for the simulations, since the real structure was somewhat beyond the capability of the two-dimensional model. Despite these two variables, we believe that the THERMIX code achieved reasonable results from the analysis of the reactor transients during the HTR-10 full load rejection test. Currently the world’s first demonstration project of a modular HTR power plant, i.e., the HTR-PM, is under commissioning in China and will be put into operation soon [25]. Since the HTR-PM design is based on an industrial enlargement of the HTR-10, the THERMIX code, which has been benchmarked here by the HTR-10 full load rejection test data, could be applied for simulating the similar transient conditions for the HTR-PM.
Grid stability and safety issues associated with nuclear power plants
.Nuclear power plant commissioning experience
. Prog. Nucl. Energy. 53, 668-672, (2011). doi: 10.1016/j.pnucene.2011.04.010A design of in-plant electric power system for the HTR-10
. Nucl. Eng. Des. 218, 227-232, (2002). doi: 10.1016/S0029-5493(02)00194-2Thermal hydraulic calculation of the HTR-10 for the initial and equilibrium core
. Nucl. Eng. Des. 218, 51-64, (2002). doi: 10.1016/S0029-5493(02)00198-XThermal hydraulic transient analysis of the HTR-10
. Nucl. Eng. Des. 218, 65-80 (2002). doi: 10.1016/S0029-5493(02)00199-1Air and water ingress accidents in a HTR-Modul of side-by-side concept
. InPreparation, conduct, and experimental results of the AVR loss-of-coolant accident simulation test
. Nucl. Sci. Eng. 107, 99-113, (1991). doi: 10.13182/NSE91-A15725High temperature reactor core physics and reactor dynamics
. Nucl. Eng. Des. 121, 227-240, 1990. doi: 10.1016/0029-5493(90)90108-AThermal response of a modular high temperature reactor during passive cooldown under pressurized and depressurized conditions
. Nucl. Eng. Des. 236, 475-484, 2006. doi: 10.1016/j.nucengdes.2005.10.027Heat transport and afterheat removal for gas cooled reactors under accident conditions. (TECDOC-1163, Vienna, 2000)
. https://www-pub.iaea.org/MTCD/Publications/PDF/te_1163_prn.pdfEvaluation of high temperature gas cooled reactor performance: Benchmark analysis related to initial testing of the HTTR and HTR-10. (TECDOC-1382, Vienna, 2003)
. https://www-pub.iaea.org/MTCD/Publications/PDF/te_1382_web/attention.pdfEvaluation of high temperature gas cooled reactor performance: Benchmark analysis related to the PBMR-400, PBMM, GT-MHR, HTR-10 and the ASTRA critical facility. (TECDOC-1694, Vienna, 2013)
. https://www-pub.iaea.org/MTCD/Publications/PDF/TE-1694_web.pdfBenchmark calculation for the steady-state temperature distribution of the HTR-10 under full-power operation
. J. Nucl. Sci. Technol. 46,572-580 (2009). doi: 10.1080/18811248.2007.9711564Post-test analysis of helium circulator trip without scram at 3 MW power level on the HTR-10
. Nucl. Eng. Des. 239, 1010-1018, (2009). doi: 10.1016/j.nucengdes.2009.02.009Temperature response of the HTR-10 during the power ascension test
. Sci. Technol. Nucl. Ins. 2015, 1-13 (2015). doi: 10.1155/2015/302648Post-test simulation of the HTR-10 reactivity insertion without scram
. Ann. Nucl. Energy. 2016, 36-45 (2016). doi: 10.1016/j.anucene.2016.01.023The design features of the HTR-10
. Nucl. Eng. Des. 218, 25-32, (2002). doi: 10.1016/S0029-5493(02)00182-6Thermal hydraulic instrumentation system of the HTR-10
. Nucl. Eng. Des. 218, 199-208, (2002). doi: 10.1016/S0029-5493(02)00191-7Temperature measuring system of the in-core components for Chinese 10 MW high temperature gas-cooled reactor
. J. Nucl. Sci. Technol. 39, 1086-1093, (2002). doi: 10.1080/18811248.2002.9715297Structural design and two-phase flow stability test for the steam generator
. Nucl. Eng. Des. 218, 179-187, (2002). doi: 10.1016/S0029-5493(02)00189-9Commissioning and operation experience and safety experiments on HTR-10
.Advantages of going modular in HTRs
. Nucl. Eng. Des. 78, 129-136 (1984) doi: 10.1016/0029-5493(84)90298-XThe Shandong Shidao Bay 200 MWe high-temperature gas-cooled reactor pebble-bed module (HTR-PM) demonstration power plant: an engineering and technological innovation
. Engineering. 2, 112-118 (2016) doi: 10.1016/J.ENG.2016.01.020