Validation of Seismic Performance of Stainless Press-to-Connect Piping System under Cyclic Loadings

Abstract

Earthquakes with magnitudes over 5.0 occurred near Gyeongju and Pohang in southern Korea in 2016 and 2017, respectively. These earthquakes had both low- and high-frequency components. Due to earthquakes with high-frequency motions, damage to nonstructural systems has been observed to be relatively more than that in structural systems. Consequently, the seismic design or performance evaluation of nonstructural components in critical facilities has emerged as a key research area in Korea. This study presents the results of experimental and numerical analyses using a high-fidelity finite element (FE) simulation in the ABAQUS platform for a press-to-connect piping system as a nonstructural component. Press-to-connect piping systems based on NFPA-13 with two elbows, a flexible coupling, and a T-joint were used. In addition, a cyclic loading protocol was applied using the KBC 2016 and IBC 2015. Based on the component-level experimental test, an FE model of the press-to-connect elbow was developed, and the high-fidelity large-scale piping system with an elbow was validated in this study. In both the experimental and analytical results, no leakage or plastic deformation of the piping system was observed under cyclic loading conditions. The results of the high-fidelity simulation model of the large-scale piping system were identical to those of the experimental test. More specifically, the error of the von-Mises stress at the upper and lower elbows was less than 9%, and the angle between the elbows was less than 2%, corresponding to the limit state of the drift ratio of the building system. Therefore, the high-fidelity simulation model of a large-scale piping system can have high application value. In addition, the design requirements and engineering demands of the piping system, such as the condition of ASME B and PV section III for service level D, were satisfied.

1. Introduction

Piping systems are essential for providing and distributing energy resources, such as water, gas, and oil, in critical facilities. In addition, they are related to system operation and safety. Failure of piping systems in critical structures can cause system shutdown in extreme cases, which leads to a situation where the plant cannot generate power. The failure of piping systems is mostly associated with leakage caused by corrosion, low-cycle fatigue, construction vibration, and seismic ground motions. More specifically, the failure of piping systems can be classified into two groups: (a) failure of pipes, such as elbows and tee joints, and (b) failure of the support system of the piping system, such as hangers and anchors. For example, during the 1994 Northridge earthquake that had a magnitude of 6.8, three transmission piping systems that supplied the city with water were damaged, and leaks of the gas piping system were also reported because of steel pipe corrosion conditions [1]. In addition, the Kobe earthquake in 1995 that had a magnitude of 6.9 resulted in approximately 1600 leakages of water distribution piping systems, and these leakages mostly occurred at the connections of the piping systems [1]. Pipeline joints can be broadly classified into welding and non-welding connections. The welding joint must be safe against the risk of fires or explosions during construction. However, non-welding connections, such as threaded, grooved, and press-to-connect systems, do not have the same potential risks as welding joints. In addition, many researchers have addressed the seismic performance of threaded and grooved connection systems during and after strong ground motion, considering welding-free conditions. Ryu et al. [2] characterized the failure limit state of various sizes of threaded and grooved piping fit joint systems, and a nonlinear finite element (FE) model of the large scale piping system incorporated with the component level piping was validated based on experimental test results. Ju and Gupta [3] and Ju et al. [4], using a rotation spring on the OpenSees platform, developed a nonlinear FE model that corresponded to the experimental test data of component levels and then developed seismic fragility as a probabilistic seismic risk assessment of a piping system level, which was identified using twice elastic limit states on a threaded T-joint. Furthermore, the fragility estimation of a system-level piping system with four grooved fit elbows and eight grooved fit T-joint components of a 10-story reinforced concrete building was studied by Wang et al. [5]. Soroushian et al. [6] investigated the dynamic characteristics of a 2-story building system with nonstructural components, including partition walls, celling, and piping systems, using schedule 40 steel under multiple seismic ground motions based on ICC-AC 156 parameters; in particular, linear and nonlinear frame building systems were subjected to various ground motions on shaking tables. Subsequently, the difference in frequency between the linear and nonlinear systems was observed, and the damage to the nonstructural components related to the experimental tests was reported. In addition, for other joint systems, the numerical and theoretical models of interference joints of circular and oval piping systems were validated by Cao et al. [7] using a strength analysis, and the contact stress corresponding to the equivalent stress in the quantitatively characterized piping prediction model was adequately assessed. Finally, Kim et al. [8] tested the seismic performance of riser piping, press-to-connect, and groove fit joint systems under in-plane cyclic loads, and then image processing corresponding to the deformation angle at the elbow joints was conducted. In summary, over the past decade, there have been only a few studies on the press-to-connect piping component. Therefore, this study aimed to determine the seismic performance of the press-to-connect joint system in a piping system subjected to strong ground motion using experimental and analytical methods to verify its applicability to sites in nuclear and non-nuclear power plants. More specifically, based on the ASCE 7-16 [9] and NFPA 13 [10] design guidelines, a system-level riser piping system with elbows and T-joints using press-to-connect joints was constructed, and then an in-plane loading condition was applied using actuators in this study. In addition, to identify the dynamic performance of the press-to-connect system, an elbow component experimental test was conducted, and high-fidelity FE simulation models at both the component and system levels with respect to the press-to-connect piping system were validated using the ABAQUS platform [11], corresponding to the experimental test data. Finally, the limit states of the large-scale press-to-connect piping system under cyclic loading conditions was experimentally and analytically estimated.

2. Experimental Test of Component Level

Threaded and grooved systems as non-welding joints in pipelines are commonly used in the transportation of energy resources; however, these joint systems are prone to leaks that can cause a pressure drop and inconsistent pipeline behavior. These issues in the joint system can cause significant damage to critical structures during and after strong earthquakes. To overcome the drawbacks of threaded and grooved joints, press-to-connect systems with O-ring seals at the end of pipes, fitted using a hand-held press, have been developed in recent years. Typically, the failure mode is observed at the elbow of the piping system owing to seismic ground motion. Therefore, the seismic performance of an application of the press-to-connect elbow system must be evaluated. A press-to-connect system that represents an elbow of a piping system was set up for the experiment, as shown in Figure 1.

In this study, the joint system was applied using a double-ring method that uses O-ring seals with a stainless-steel O-ring for precise compression and a rubber O-ring to ensure a permanent leak-free condition. As shown in Figure 1, the elbow is a piping system component made from stainless steel (STS 304) and composed of a 100A (114.3 mm) diameter and SCH 10S (thickness 3.0 mm). To allow for rotation of the elbow, a pin joint connected to both jigs at the ends was applied and then a 90° press-to-connect elbow system was set up. In addition, a sinusoidal loading protocol (60 mm/min) by a universal testing machine (UTM) was applied using the displacement control method. An internal pressure of 2 MPa was applied by filling the elbow joint with water during the experiment to detect the leakage limit state of the piping system.

3. Validation of FE Model of Elbow Component

An FE model of a component level such as the elbow joint of a piping system must be reconciled based on experimental test results to develop a high-fidelity FE simulation model of a large-scale piping system using a press-to-connect system. In this study, to determine the material properties of the STS 304, tensile tests were conducted on three specimens, as shown in Figure 2, and the stress–strain relationships of the STS 304 material are displayed in Figure 3.

Based on the tensile test results, kinematic hardening model for this material was assumed as described in Figure 3. The material parameters were constituted with kinematic hardening modulus (c), equivalent stress in accordance with yield surface $\left({\sigma }^{vM}\right)$, back-stress $\left(\alpha \right)$, and equivalent plastic strain $\left({\dot{\epsilon }}^{pl}\right)$, as defined in Equation (1) [12].

$$\dot{\alpha }=C\frac{1}{{\sigma }^{vM}}\left(\sigma -\alpha \right)\left({\dot{\epsilon }}^{pl}\right)$$

(1)

where, the kinematic hardening modulus was applied in this study regardless of temperature, the equivalent plastic strain can be explained in a similar manner based on the definition of the von Mises constitutive equation [13], and the yield surface can be defined as the stress surface that satisfied the yield condition of the yield function related to the principal stresses. In addition, the equivalent stress in accordance with von Mises stress can be expressed in Equation (2), if ${\sigma }_{12}={\sigma }_{23}={\sigma }_{31}=0$

$${\sigma }^{vM}=\sqrt{\frac{1}{2}\left\{{\left({\sigma }_1-{\sigma }_2\right)}^2+{\left({\sigma }_2-{\sigma }_3\right)}^2+{\left({\sigma }_3-{\sigma }_1\right)}^2\right\}}$$

(2)

More specifically, a bilinear hardening plastic stress–strain curve corresponding to the kinematic hardening model was determined to describe the constitutive model of the elbow system, as shown in Figure 3. Finally, an FE model of the press-to-connect elbow system was constructed, and the jig condition was represented by a 2-node beam element (B31), while 4-node shell elements (S4R) were applied to establish the mechanical joint and straight pipe using the ABAQUS platform [11]. Considering the O-ring seals, the FE model of the elbow joint was simplified based on the spring element because of its extreme complexity and convergence difficulty of the simulation between the O-ring seals at the end of the pipe and mechanical pressing connectors. Figure 4 illustrates the FE model of the elbow joint and related boundary conditions.

To achieve efficiency and accuracy of the FE model, a 2 mm mesh size for the elbow and straight piping system was generated, and the beam element for the jig was set to a 3 mm mesh size in this study. The reconciliation of the experimental and numerical data is shown in Figure 5. As can be seen, the load-displacement relationship of the simulation model was in good agreement with the data of the experimental test obtained using the UTM, and the initial stiffnesses of the results of the experimental test and FE simulation model were almost the same. The maximum forces corresponding to the experimental test result and FE model up to 20 mm were 5.44 kN and 5.59 kN, respectively. Furthermore, the results of the normalized energy dissipation related to the loading protocol is shown in Figure 6a.

Even considering that the difference between the experimental and analytical data slightly increased, the error did not exceed 5%. In addition, the overall trend of the deformation angle of the FE elbow model shown in Figure 6b up to 20 mm is considerably higher than that of the experimental test. This issue may be caused by the springback and Bauschinger effect, due to the difference between tensile strength and compressive strength at the yield position [14], of the constitutive steel material model on the press-to-connect elbow system. However, the error for the validation of the FE model of the press-to-connect elbow system was less than 7% for both the force–displacement relation and deformation angle. Consequently, the three-dimensional FE model of the press-to-connect elbow system was reliably verified as a high-fidelity simulation component model. In addition, during the experimental test of the component level, there was no failure or leak on the press-to-connect elbow system. Subsequently, additional testing was conducted, and a leakage was detected at $\pm 11°$ of the deformation angle of the elbow. Considering the test and FE simulation model, it can be concluded that the press-to-connect elbow system presented sustainable seismic performance.

4. Experimental Test of Large-Scale Piping System

To evaluate the seismic performance of a large-scale piping system with a press-to-connect elbow system, an experimental test must be conducted. The common failure mode related to low-cycle fatigue or ground motion is concentrated at the elbow joints or T-joint systems in the piping system [8]. Therefore, in this study, a large-scale piping system with two press-to-connect elbow systems and one press-to-connect T-joint system with riser pipes based on NFPA 13 [10] were constructed in this study. In addition, according to guideline NFPA 13 [10], two flexible couplings, that is, sockets in the riser piping system, to absorb the energy of the vibration from the ground motion or other mechanical vibrations were installed within 610 mm from the upper bound (ceiling) and 305 mm from the lower bound (bottom floor), as shown in Figure 7.

The piping equipment was anchored by bolts at all ends. To measure the load–displacement relationship of the system, three-axis load cells (LC1, LC2, and LC3) were installed at the ends of the piping system, as illustrated in Figure 8. Furthermore, three-axis rosette strain gauges were applied to the crown at the elbow joint, and to analyze the effect of the performance of the press-to-connect system on the piping, a strain gauge was attached to the middle of the socket. As shown in Figure 8, the actuator was connected to the upper slab of the steel frame, in order to apply the loading protocol. To simulate the allowable side sway during and after an earthquake, the hinge system condition was applied to the steel frame as a geometric boundary condition.

Moreover, the cyclic loading protocol based on displacement control was repeated 10 times or more to satisfy the maximum drift ratio of the building system corresponding to design codes KBC 2016 [15] and IBC 2015 [16], as listed in

Table 1. Allowable inter-story drift ratio for KBC 2016.

	Level
	Special	I	II
Drift ratio	0.010 h	0.015 h	0.020 h

and

Table 2. Allowable inter-story drift ratio for IBC 2015.

	Risk Category
	IV	III	I or II
Drift ratio	0.010 hxx	0.015 hxx	0.020 hxx

, where h and hxx denote the story heights for design codes KBC 2016 and IBC 2015, respectively.

In addition, during the experiment the press-to-connect piping system was filled with water to detect the damage state with respect to the leakage and pressurized with an internal pressure of 2 MPa using an air-water booster. Finally, the 20 mm/min rate loading protocol was executed in the test using the displacement control.

5. Validation and Verification of FE Large Scale Piping System Model

Experimental tests of large-scale piping systems are typically time-consuming and costly. To overcome these issues, it is necessary to develop a high-fidelity simulation model based on an experimental test to understand the in- or out-of-plane behavior of the piping system. Considering the material properties of the component-level test data, the bilinear kinematic hardening plasticity model as a nonlinear behavior material model was used to construct a high-fidelity FE simulation model of the large-scale piping system with a press-to-connect system. This large-scale piping system consisted of an upper elbow, a low elbow, a T-joint, and two flexible couplings to absorb energy from an earthquake. Based on the schematic design and experimental test of the piping system, a high-fidelity simulation FE model with the components was developed, and to allow seismic sway, the displacement and rotation in the z-direction were unfixed, as demonstrated in Figure 9.

Additionally, to describe the characteristics of the T-joint in the piping system as a press-to-connect system, both S3 and S4R shell elements were utilized, and other piping components such as the straight pipe, elbows, and flexible couplings were constructed using the S4R shell element in the ABAQUS platform [11]. Thus, a total of 31,751 and 32,360 elements and nodes were generated in the high-fidelity simulation model, respectively. Next, the results of the analytical model of the piping system were validated using the experimental test results. The characteristics of the seismic performance of the elbow joints with a press-to-connect pipe were extensively focused on in this study because the elbow is the weakest component in the piping system during and after a strong earthquake. Therefore, reconciliation of the experimental and numerical results at the elbow was the main priority in this study. First, the von Mises stress of the large-scale piping system was verified, as shown in Figure 10. The results of the high-fidelity simulation FE model are in good agreement with the data observed from the experimental test. The maximum and minimum stresses of the large-scale press-to-connect piping system is described in Figure 10c.

The maximum von Mises stress of the FE model in the piping system and the stress from the experimental test at the lower press-to-connect elbow were 187 and 205 MPa, respectively, at the 10th cycle during the loading condition. In addition, the observations from the upper elbow system, that is, the stresses of the FE model and experimental test, were 371 MPa and 403 MPa, respectively. In the case of the stress analysis, the error of the stress for the validation of the piping system model was less than 9%, which indicated that the FE model of the press-to-connect large-scale piping system using the ABAQUS platform [11] was a high-principled model. In particular, during and after a strong earthquake, the seismic sway to validate the FE model with the more critical condition was also an important phenomenon in the structural and nonstructural components. Figure 11 illustrates the deformation angle of the two-elbow system obtained from experimental and analytical data. The deformation angle results of the experimental and analytical models were not significantly different. It can be observed that the high-fidelity simulation FE model was identical to the experimental test.

Figure 11 presents a validation of the deformation angles of the large-scale press-to-connect piping system. However, because the interference slip issue between the opening and closing modes during the cyclic test was not considered in the analytical FE model of the large-scale piping system, an insignificant error in the validation between the analytical and experimental results was observed in this study. This issue could also be caused by the springback phenomenon, which is not reflected in the analytical model. Even with this small issue, it was revealed that the characterization of the parameters and material properties of the high-fidelity FE model were extremely reliable. Additionally, it is interesting to note that the seismic sway standard (0.02 h) corresponding to KBC 2016 [15] and IBC 2015 [16] was satisfied in the press-to-connect elbow system of the large-scale piping system in both the experimental and analytical results. In addition, to verify the seismic performance of the piping system, the following equations of the ASME boiler and pressure vessel code section III, subsection NB [17] for service level D, can be applied.

$$B_1\frac{P_E{D}_o}{2t}+B_2\frac{D_o}{2I}{M}_E\le 3S_m$$

(3)

where B₁ and B₂ are the primary stress indices for the product under investigation, D_o is the outside diameter of the pipe, I is the moment of inertia, M_E is the amplitude of the resultant moment owing to the weight and inertial loading obtained from reversing the dynamic loads, P_E is the pressure that is coincident with the reversing dynamic load, and S_m is the allowable design stress intensity value at a temperature consistent with the loading under consideration. The allowable design stress (S_m) for the stainless-steel material (SUS304) was 137 MPa, and 3S_m was considered to be 410 MPa. Finally, considering ASME B and PV section III [17] for service level D of the press-to-connect piping system based on the NFPA 13 [10] condition, the stress level mentioned above can be experimentally and analytically satisfied. Therefore, the high-fidelity FE model of the piping system related to the data from the experimental test was verified according to seismic demands, and the FE model has a high application value for further research.

6. Conclusions

The piping system is an essential nonstructural component that provides or maintains sustainable energy distribution in critical/noncritical facilities. The joint system, reported as one of the most vulnerable parts during and after earthquakes, consists of various non-welding connections, such as threaded, grooved, and press-to-connect joint systems, in the piping system. This study focused on experimentally and analytically evaluating the seismic performance of an elbow joint and a large-scale piping system with a press-to-connect joint system, an O-ring seal, and a compressed seal for leak-free operating conditions. More specifically, based on an experimental test of the piping system, a high-fidelity FE model of the press-to-connect piping system was validated and verified. The conclusions corresponding to the seismic requirements and engineering demands of the piping system in this study are as follows:

To evaluate the dynamic characteristics of the press-to-connect elbow system, an in-plane cyclic loading test was performed, and a force–displacement curve of the elbow system was obtained. Subsequently, a high-fidelity FE model of the elbow system was constructed using the ABAQUS platform [11]. However, because the O-ring seal and compressed seal for permanent leak-free press-to-connect were extremely complicated in the FE model, a simplified 3D elbow joint model using a spring element in the ABAQUS platform [11] was generated in this study. In comparison with the results of the experimental test and analytical model, the high-fidelity FE model was slightly more conservative than the data obtained from the test, which is attributed to errors by the springback phenomenon and Bauschinger effect of the constitutive material model on the press-to-connect elbow system. Consequently, notwithstanding the insignificant error, the press-to-connect elbow FE model developed in this study was reliable.

Next, to understand the field application value and seismic performance of a large-scale piping system containing a press-to-connect elbow, T-joint, and flexible coupling associated with NFPA 13 [10], a cyclic loading test using a sinusoidal wave was applied to the large-scale piping system. The velocity and displacement for the tests were 20 mm/min and ±60 mm, respectively. Regarding the elbow component level FE model, a high-fidelity simulation FE model of the large-scale piping system was constructed in this study using the ABAQUS platform [11]. The von Mises stress was relatively small at the lower and upper elbows of the large-scale piping FE model in comparison to the experimental test results. Despite this issue, the interference error was not considerable because it was less than 9%. In addition, the deformation angles at the lower and upper elbows were in good agreement with the experimental test results. The error and change in the stress and deformation in the high-fidelity simulation FE model may be caused by the effect of the slip and spring-back phenomenon in the piping system. However, because the interference error at the elbow and large-scale piping system applied to the press-to-connect type condition was relatively trivial, the application value of the large-scale piping FE model may be extremely high. In addition, the press-to-connect large-scale piping system may satisfy the condition of ASME B and PV section III [17] for service level D, both experimentally and numerically.

Finally, this study explored the dynamic characteristics and seismic performance of the press-to-connect elbow and large-scale piping system under in-plane cyclic loading conditions, and the design requirements and engineering demands of the piping system were mostly satisfied. Therefore, although the initial costs to provide or distribute sustainable energy using the press-to-connect piping system instead of a steel piping system in a critical facility are higher, the press-to-connect piping system with an O-ring seals and compressed seals can reduce the long-term costs in terms of maintenance, as there was no leakage or plastic deformation from the experimental and numerical data.

Furthermore, the interaction between the building and press-to-connect piping system must be evaluated, and the system-level seismic fragility must be constructed to understand the risks and drawbacks of the piping system.