An Experimental Evaluation of the APR1000 Core Flow Distribution Using a 1/5 Scale Model

Abstract

The experimental data of core flow distribution are indispensable for obtaining licensing and facilitating the design of fluid systems of nuclear reactors. In this study, an Advanced power reactor Core flow and Pressure (ACOP) test facility was established to experimentally simulate the internal flow of the Advanced Power Reactor 1000 (APR1000) on a reduced length scale of 1/5. The core region was simulated by using 177 core simulators representing the fuel assemblies of the APR1000. The APR1000 flow distributions were synthetically identified by accurately measured parameters: the core inlet flow rate and outlet pressure under the four-pump balanced and unbalanced flow conditions. The overall inlet flow rates ranged from 87.7% to 112.0% relative to the averaged flow rate. Here, we scrutinize the flow distributions considering the flow conditions and internal structures and briefly describe the applied scaling method and design concept of the test facility.

1. Introduction

The Advanced Power Reactor 1000 (APR1000) is an advanced Korean nuclear power Reactor improved from the Advanced Power Reactor 1400 (APR1400) and Optimized Power Reactor 1000 (OPR1000) of Korea. Several advanced design features of the Advanced Power Reactor Plus (APR+) and the European Advanced Power Reactor (EU-APR) were also adopted to enhance plant safety with outstanding performance. The reactor core of the APR1000 is composed of 177 fuel assemblies to generate 2825 MW thermal power, and experimental verification of the core design is essential to ensure safety and secure export competitiveness of the APR1000. Core flow distribution data are especially crucial for obtaining licensing and facilitating the design of fluid systems. The core flow distributions are characterized as the core inlet flow and core outlet pressure distributions, which are used as important boundary conditions to determine the Departure from Nucleate Boiling Ratio (DNBR) for evaluating the core thermal margin.

Hydraulic experiments regarding pressure drop and flow distribution were first suggested and performed with a 1/7.5 scale model in the 1960s [1]. These experiments considered all hydraulic parameters and finally suggested the scaling law based on Buckingham’s Pi theorem. They successfully analyzed the effect of various internal geometries on the core inlet flow distribution and Euler numbers for the optimum design of the Connecticut Yankee reactor. Their applied scaling methodology has been widely used to quantify the core distribution over the past several decades in Korea. From the 1970s to the 1980s, an experimental program was conducted on a 1/5.03 scale model of the Yonggwang nuclear unit 3 and 4 to produce experimental data for the core inlet and outlet flow distributions [2]. The obtained data have been used to provide input data for the core thermal margin analysis in Korea, and a small modular reactor and evolutionally Generation III+ reactors were proposed and developed in the 2000s. The core flow distribution data must be obtained from a precise and representative model in strict compliance with the similarity law; therefore, many hydraulic experiments have additionally been performed for various reactor types. Euh et al. [3] carried out core flow distribution experiments for the System-integrated Modular Advanced ReacTor (SMART). They developed core simulators with the venturi meter to measure the differential pressure. The 57 core simulators were precisely calibrated in advance and used to quantify the flow distribution under the nominal condition. Based on the experimental experience, more advanced techniques and improved test procedures were adopted for the flow distribution test of the GENIII+ APR+ reactor [4]. For a successful hydraulic model, core simulators representing the High Performance with Efficiency and Reliability (HIPER) fuel assemblies [5] were developed, and the hydraulic characteristics, such as the axial pressure drop, and cross flow characteristics were verified [6,7]. Euh and Kim [8,9] simultaneously considered two types of similarities, geometric and dynamic, to minimize the distortion of the hydraulic characteristics in reduced-scale test facilities. They conducted various hydraulic tests and identified the core flow distributions under the representative hydraulic conditions of the prototype APR+ reactor. They found that the core inlet flow of the original APR+ design has higher values at the core inlet edge region, which is not profitable in terms of durability, which includes mechanical problems. To reduce the flow rate at the core edge region, an improved design was suggested for the core lower support structure [10,11].

In this study, we produced experimental data on the flow distribution at the core region of the APR1000 with a high degree of accuracy. According to the test requirements, we performed three different experiments under a four-pump running condition as follows: (1) four-pump balanced flow condition, (2) four-pump 5% unbalanced flow condition, and (3) four-pump 15% unbalanced flow condition. The 15% unbalanced flow condition was considered to be the most conservative scenario under the four-pump operating conditions. The Reynolds number dependency on the Eu number was also verified experimentally by varying the flow rate. All experiments were performed under a steady state condition with constant temperature, pressure, and flow rate. The core inlet flow and outlet pressure distributions were quantified at each fuel assembly position, and the results were analyzed and discussed by non-dimensional distribution.

2. Experimental Test Facility and Conditions

2.1. System Configuration and Main Test Section

The schematic of the Advanced power reactor Core flow and Pressure (ACOP) fluid system configuration is shown in Figure 1. For this study, the main test section is the reactor pressure vessel (RPV), including some parts of four cold legs and two hot legs, without considering steam generators. Each extension pipe of the cold legs and hot legs was equipped with a flow meter for measuring the flow rate. The working fluid was continuously heated by waste heat from four 90 kw centrifugal pumps. To maintain a constant operating temperature, the temperature was controlled by adjusting the amount of cooling water entering the heat exchangers through flow control valves, as shown in Figure 2. A vortex flow meter was used, and pressure and temperature were measured simultaneously at the downstream of the flow meter to calculate the density of the working fluid. The accuracy of the instrumentation is summarized in

Table 1. Accuracy of instrumentation.

Instrumentation	Accuracy
Flowmeter (Rosemount 8800 Series)	±0.66% of span
Differential pressure (Rosemount 3051S Series)	±0.15% of span
Pressure (Rosemount 3051S Series)	±0.15% of span
Temperature (K-Type TC Watlow)	±1.5 °C

. The fluid system was designed as a closed loop, and the fluid flow was operated by four hydraulic pumps capable of delivering a head of 50 m and a flow rate of up to 430 m³/hr. The instrumentation and control diagram are shown in Figure 2. A more detailed description of the system has been outlined in previous research [8,9].

The RPV was basically designed to simulate the internal flow based on the principle of similarity. The RPV consists of an assembly of various internal structures, as shown in Figure 3. Every internal structure shape, except for the core region, was manufactured with an exactly 1/5 reduced linear scale referring to the APR 1000 reactor based on the geometrical preservation. The aspect ratio was maintained, and thus the flow area and volume were preserved.

Table 2. Scaling ratio for hydraulic parameters.

Parameter	APR1000	Scaling Ratio	ACOP
Temperature [°C]	310	-	60
Pressure [MPa]	15	-	0.2–0.5
Length ratio [ - ]	1	lR	1/5
Height ratio [ - ]	1	lR	1/5
Area ratio [ - ]	1	lR2	1/25
Volume ratio [ - ]	1	lR3	1/125
Aspect ratio [ - ]	1	1	1.0
Velocity ratio, [ - ]	1	VR	1/2
Mass Flow ratio, [ - ]	1	ρRVR lR2	1/35.6
Density ratio [ - ]	1	ρR	1.40
Viscosity ratio [ - ]	1	μR	5.57
Core Re ratio [ - ]	1	ρRVRDR/μR	1/2.9
DP ratio [ - ]	1	ρRVR2	1/2.85

shows a summary of the scaling ratios related to the main design hydraulic parameters. This ensures that the geometric similarity was preserved along the internal fluid flow path.

The core of the APR1000 consists of 177 HIPER arranged in a 15 × 15 array, as shown in Figure 4. Simulating the actual fuel assembly shape is not feasible with a scaled-down model due to the complex geometry, and also it does not align with the objective of this study. Therefore, as shown in Figure 4, the core region was simulated separately by using core simulators which have the same hydraulic characteristics, such as an axial pressure drop and cross flow characteristics between adjacent fuel assemblies. The core dynamic similarity is entirely preserved when the Eu number is identical to the fuel assembly of the APR1000 reactor and core simulator as follows:

$$\begin{gathered} \frac{E{u}_p}{E{u}_m}=\frac{{\left.\Delta P/\rho V^2\right|}_p}{{\left.\Delta P/\rho V^2\right|}_m}=1 \\ {\left.\Delta P\right|}_m={\left.\Delta P\right|}_p\times \frac{{\rho }_m}{{\rho }_p}\times {\left(\frac{V_m}{V_p}\right)}^2 \\ {\left.\Delta P\right|}_m=161.3\times \frac{983.2}{701.1}\times {\left(\frac{1}{2}\right)}^2 \\ =56.56 \mathrm{kPa} \end{gathered}$$

(1)

The operating temperature and pressure of the test facility were maintained at 60 °C and 2–3 bar, respectively. The velocity ratio was set to a 1/2 scaling ratio, as shown in Table 2. The axial pressure drop of the APR1000 reactor was a given value for normal operating conditions; thus, that of the core simulator was obtained from Equation (1).

The Reynolds number cannot be reduced to the same ratio due to the different hydraulic diameters of the core simulator. As shown in Table 2, even if there is a 1/2.9 Reynolds number ratio, the Reynolds number of the core simulator is about 1.85 × 10⁵, which ensures a fully turbulent region. It is noted that the hydraulic behavior of the core simulator can be preserved with the Euler number’s conservation because the pressure drop is not significantly affected by the change in the Reynolds number in the fully turbulent region [2].

The core inlet flow rate and exit pressure must be measured independently for each fuel assembly; therefore, the core simulator was designed to have the venturi meter at the inlet and three pressure impulse lines: two for the venturi meter and one for the outlet pressure. Therefore, 531 pressure impulse lines were drawn out through the core shroud and upper guide structure regions, which did not affect the fluid flow.

Figure 4 and Figure 5 depict the layout and show a photograph of the core simulators and pressure impulse lines, respectively. The axial hydraulic resistance was simulated by using four orifice plates along the axial direction. The obtained axial pressure drop from Equation (1) was adjusted by controlling the orifice flow area of the core simulator. The core inlet flow rate for each core simulator was calculated by Equation (2). An accurate discharge coefficient was required to calculate the core inlet flow rate by measuring the differential pressure at the venturi throat. Therefore, 177 discharge coefficients were precisely calibrated by varying the flow rate in our previous studies as shown in Figure 6, and the relative errors for the design value were within 0.85% [6,7].

$$\begin{gathered} {\dot{m}}_{cs}=K_i\sqrt{\rho \Delta P_{ven, i}} \left(i=1~177\right) \\ {K}_i=\frac{C_{d,i }{A}_{ven, i}}{\sqrt{1-{\beta }^4}} \end{gathered}$$

(2)

2.2. Test Requirements and Conditions

The experiments were conducted according to the specified test requirements and operated at a steady state by controlling the reactor inlet flow rates, fluid temperature, and system pressure. The main hydraulic parameters are summarized as follows:

• Reactor inlet flow rate (four cold legs);
• Reactor outlet flow rate (two hot legs);
• Fuel assembly core inlet flow rate and distributions (177 points);
• Fuel assembly core outlet pressure and distribution (177 points).

As mentioned earlier, there were three different flow conditions for the four-pump running condition. The balanced flow condition means that the flow ratio of each cold leg was identical with 25% of the total flow rate. The unbalance flow condition tests were classified into two cases based on the difference between the maximum and minimum flow rate ratio, named 5% and 15%. For each experimental condition, independent experiments were conducted to confirm their integrity under identical test conditions. The ensemble averaged cold leg flow rates for the repeatability tests and test matrix are shown together in

Table 3. Test matrix and ensemble average of measured inlet flow rates.

4-Pump Conditions		CL1-A	CL-1B	CL-2A	CL-2B
Balanced flow (BL; 15 Tests)	Flow ratio	0.250	0.250	0.250	0.250
	Inlet flow rates [kg/s]	100.0	100.0	100.0	100.0
5% Unbalanced flow (UB05; 3 Tests)	Flow ratio	0.252	0.252	0.242	0.254
	Inlet flow rates [kg/s]	100.0	99.99	97.60	102.39
15% Unbalanced flow (UB15; 3 Tests)	Flow ratio	0.252	0.252	0.232	0.268
	Inlet flow rates [kg/s]	100.0	99.99	92.80	107.17

.

Two kinds of mass balances were evaluated to check the validity of the test results. The mass flow rates were calculated as the total mass flow rates at three specified locations: four cold legs, two hot logs, and 177 core simulators. The sum of the three flow rates should theoretically be equal with each other, and the total reactor inlet flow rate was set as a reference for comparison with the others. The relative errors for mass balance are defined as follows:

$${ϵ}_{cs}\left(\%\right)=\frac{\sum Q_{cl, i}-\sum Q_{cs,j}}{\sum Q_{cl, i}}\times 100 \left(i=1~4, j=1~177\right),$$

(3)

$${ϵ}_{cl}\left(\%\right)=\frac{\sum Q_{cl, i}-\sum Q_{hl,k}}{\sum Q_{cl, i}}\times 100 \left(k=1~2\right).$$

(4)

The allowable data deviation for the measuring parameter are presented in

Table 4. Criteria of the measured data scatter.

Parameter	Criteria
Each core simulator inlet flow distribution	±1.5%
Each core simulator outlet pressure distribution	±2%
Mass balance at core inlet	±2%
Mass balance at core outlet	±2%

, and the relative errors for all the tests are summarized in

Table 5. Mass balance at the core inlet/outlet for the four-pump tests.

Relative Error/Tests		BL (15 Tests)	UB05 (3 Tests)	UB15 (3 Tests)
${ϵ}_1$ [%]	Mean	0.56	0.54	0.53
	STD	0.010	0.011	0.009
${ϵ}_2$ [%]	Mean	0.91	0.97	0.84
	STD	0.067	0.050	0.052
Total mass flow rate [kg/s]	Mean	399.99	399.98	399.98
	STD	0.0009	0.0009	0.010

. The coefficient of variations of the total mass flow rate were within 0.25%, indicating that steady-state conditions were well maintained during the experiments. The mass balance also satisfied the criteria for all the tests, as defined in Table 4; therefore, the current experiments yielded reliable results with good repeatability.

3. Experimental Results for Core Flow Distributions

The flow rate and pressure were each measured individually using 177 core simulators under the four-pump running conditions. The obtained values represent the fuel assemblies of the APR1000 reactor core. All the measured parameters were normalized by dividing the averaged values (

Table 6. Normalization of flow rate and pressure for the core simulators.

Parameter		Normalized Form
Core simulator inlet flow rate	Qcs,i(i = 1~177)	$Q_{cs,i}/\overline{Q_{cs}}$
Core simulator outlet pressure	Pcs,i(i = 1~177)	$\left(P_{cs,i}-\overline{P_{cs}}\right)/\left(\overline{P_{cs,in}}-\overline{P_{cs,out}}\right)$

) to assess the distributions at the cross-sections of the inlet and outlet.

3.1. Flow Characteristics

As shown in Figure 5, coolant was introduced into the RPV through four cold legs and descended along the downcomer region. The flow direction suddenly changed in the lower plenum region, and then coolant was first mixed as it passed through the flow skirt. The 177 core simulators were mounted on the top of the fuel assembly. Subsequently, the coolant was remixed as it passed through various structures of the lower support structure before reaching the core. In the lower plenum, more complex flow patterns were anticipated due to the complex geometry, which played the role of hydraulic resistance. Figure 7 depicts the schematic of the anticipated flow patterns. If the coolant sufficiently descends to the bottom, it moves towards the center and ascends, encountering significant flow resistance (blue line in Figure 7); however, the flow that directly enters the periphery is relatively less affected by flow resistance (red line in Figure 7).

Figure 6. Flow coefficient distribution from calibration test [6].

Figure 6. Flow coefficient distribution from calibration test [6].

Figure 8 shows a color map of the core inlet flow distribution under the four-pump balanced flow condition. The cross-section drawings of the lower support structure were overlaid on the core inlet flow distribution map to assess the effect of flow resistance. Figure 8 reveals that the effect of ICI nozzles and the flow holes of the ICI nozzle support plates on the flow distribution was not clear compared to the association between the flow resistances and the inlet flow rate. Conversely, the normalized inlet flow rates were more intense near the outer region of the core. As mentioned before, the fluid flow experienced different hydraulic resistance according to the flow pattern. This tendency is likely due to the flow experiencing the least resistance when passing through the raised bottom plate.

The total core inlet flow rates were determined based on the scaling analysis and fixed at 400 kg/s for all of the four-pump running conditions. The effect of the Reynolds number on the Euler number is theoretically reduced for sufficiently high Reynolds numbers, and the Euler number was nearly independent of the Reynolds number at values above 5000 [1]. Considering the capacity of the available fluid system, an additional experiment was conducted with the total flow rates increased by 110% to quantitatively examine the effect of the Reynolds number. Figure 9 shows a comparison of the normalized core inlet flow distribution for the four-pump balanced flow condition. The maximum discrepancy for all core simulators was found to be within 0.45%. Therefore, the flow conditions used in this experiment ensured the preservation of dynamic similarity between the APR1000 reactor and the ACOP test facility.

3.2. Core Flow Distributions

For the four-pump conditions, as shown in Table 3, the normalized core inlet and outlet distributions were shown together in Figure 10 and Figure 11, respectively. The distribution range and minimum values were important data for evaluating the thermal margin to prevent the possibility of Departure of Nucleate Boiling (DNB) occurrence. The core inlet flow rates for both the balanced flow and 5% unbalanced flow conditions were distributed, ranging from 87.9 to 111.8%. For the 15% unbalanced flow condition, the results were slightly different: 87.7% to 112.0%. The core outlet pressure distributions were compared by the relative differential pressure over the averaged core differential pressure, as defined in Table 6. For all three test conditions, the normalized pressure distributions were identical, ranging from −1.7% to 3.2%, resulting in a more uniform distribution at the outlet. This is because the flow imbalance at the core inlet was sufficiently mixed while passing through the core due to the cross flow between the adjacent fuel assemblies.

Figure 9. Averaged normalized core inlet flow distributions under varying flow rate.

Figure 9. Averaged normalized core inlet flow distributions under varying flow rate.

As shown in Figure 10, overall, there were no significant differences in the distribution regardless of the imbalance flow conditions; however, the distributions were not precisely symmetric with respect to the four quadrants. The shape of the lower support structures was not symmetric, and the discrepancy of the flows could be intuitively examined by comparing the local flow values. Figure 12 shows the normalized core inlet flow values along the specified line positions, as defined in Figure 4 under the balanced condition.

The core inlet flow rates across the entire core region confirm the vertical and horizontal symmetry, referred to as the top–bottom and left–right sides, as shown in Figure 12. Each deviation along positions 1 and 2 ranges from approximately 10.8 to 15.5%; however, when integrating the normalized flow rates across the positions and comparing each side, the relative errors were found to be within 0.85%. In the centerlines, the deviations along the horizontal and vertical sides were 18.6% and 19.0%, respectively. However, the relative errors for the total flow rates were 2.83%. The obtained results indicate that the core inlet flow rates were evenly distributed due to the design of LSS, despite the significantly asymmetrical flow conditions, up to a 15% unbalanced flow condition.

4. Conclusions

A core flow test facility, named ACOP, representing a reduced length scale of 1/5 of the APR1000 reactor, was construed to simulate the flow distribution within the core region. The internal structure of the RPV was manufactured while preserving geometric similarity. Core simulators maintaining dynamic similarity were employed to simulate the 177 fuel assemblies. According to the test requirements, three kinds of experiments were successfully carried out for the balanced flow and unbalanced flow conditions, including a repeatability and validity check of the test condition. The core inlet flow rate and outlet pressure were measured, and the synthetical flow distributions were identified. The core inlet flow rates were distributed in a range of 87–112% of the average value for all of the four-pump running conditions. The core outlet pressure distributions exhibited a more uniform distribution within 3% deviation due to the cross flow in the core region. The local core inlet flow was also investigated to examine the effect of the lower support structure asymmetry on the overall flow distribution. In the outer core region, high flow rates were observed due to low flow resistance; however, a consistent inherent core inlet distribution was observed regardless of the flow conditions. The flow was sufficiently mixed passing through the structures in the lower plenum, even with severe flow disturbance on the upstream. These test results can contribute to the analysis and evaluation of flow distributions for obtaining a design license of the APR1000, and three-pump experiments assuming the postulated failure of a single pump are planned as future work.