Qualification test campaign of RFA one fairing engineering model

The fairing is an important element of the rocket as it protects the payload from the external environment during the most aggressive phases of the flight. A test campaign was developed with the aim of qualifying the fairing structure of the RFA One. This test is intended to simulate the quasi-static loads of the most critical phase during flight. A FEA model was developed to identify the most critical phase of flight and to identify the quasi-static reactions that replicate the flight loads. Based on these reactions, the loads and test levels were defined. The test results were compared with the FEA data to correlate and improve the model. The static test was successfully performed at acceptance and qualification Level 2. The load and strain results indicate that the DLL was achieved and exceeded by a factor of 1.11, while the fairing maintained its functionality and key performance. This paper presents the most relevant results of the test, discusses the results compared with the FEA model and summarizes some lessons drawn from the test campaign.


Introduction
The main objective of the static test campaign was to qualify the payload fairing (FRG) engineering model structure for flight load conditions, identify the worst load case for flight conditions, and evaluate its capacity to endure and maintain performance.Throughout the test campaign, several measurements were performed to characterise the structural behaviour of the FRG.The measurements also provided valuable information for the FEA model correlation and future improvements to the structures.

Success Criteria
The static test campaign will be declared successful if the following test objectives are met: 1. FRG maintains its functionality and key performance after executing static tests at acceptance and qualification levels, or if remedial actions are needed.2. Specified test loads are properly applied.

Test item
The test items are shown in Figure 1 and consist of the FRG assembly and the 2nd stage Top Separation Flange (TSF).The FRG is divided into two main assemblies, designated as Y POS and Y NEG, with the separation plane located at 135 degrees and 315 degrees, as shown in Figure 2.Each FRG half consists of three main structural elements: the composite shell structure, the metal flange structure connecting the two halves, and the composite ring connecting the two halves to the 2nd stage TSF.

Test Setup
Figure 3 shows the test set-up with the FRG attached to the TSF by four separation locks locate at 0, 90, 180 and 270 degrees, illustrated in Figure 2. The TSF was connected to a rigid interface, thus introducing a fixed boundary condition at the root of the FRG, simulating the connection to the EASN-2023 Journal of Physics: Conference Series 2716 (2024) 012090 IOP Publishing doi:10.1088/1742-6596/2716/1/012090 2 interstage.The FRG was positioned on the fixture with two of the locks, lock 0 and 180 degrees, in line with the applied lateral load, as shown in Figure 2. The static load case consists of two lateral loads (LL) at two different points and one axial load (AxL), illustrated in Figure 3.The LL were applied using slings wrapped around the FRG between 90 and 270 degrees, which were then connected to the end of a hydraulic actuator located on the load tower.The AxL was applied with an actuator positioned on the ground and aligned with the FRG centre axis.A steel cable was fixed to a JIG placed on the nose of the FRG, and the opposite end was connected to a hydraulic actuator.For each hydraulic actuator there was a load cell measuring the applied load.

Test matrix
The test matrix for the static test campaign is presented in Table 1.Two types of tests were defined for the test campaign: Calibration and Static.The Static tests were performed in three levels, a functional level and two qualification levels (DYL and DUL), based on the qualification verification stage [1].
The static tests objective was to simulate the quasi-static loads induced by inertial, aerodynamic and pressure loads for the most critical phase of flight, corresponding to the phase of maximum aerodynamic pressure, Max Q.
The Calibration test objective was to determine the sensibility of the Wheatstone bridges installed on the FRG structure.The calibration was performed with a single load applied in turn for each bridge type.

FEA model
A global FEM (GFEM) of the fairing was developed, with the data used for the manufacturing of the EM, as well as the interstage that interfaces with the fairing.The goal was to predict the behaviour for the worst-case scenarios of flight.The GFEM is nonlinear static with contacts and pretension to provide the adequate load paths with the variation of load directions, angles of attack and oscillations in acceleration.Relevant stages of flight such as Lift-Off, MaxQ and S1-MECO were analysed for a multitude of directions, illustrated in Figure 2, since the structure is not entirely axisymmetric, the angles of attack, combined with different inertial loads, in the GFEM analysis.From the results envelope two worst-case orientations of the aerodynamic loads at MaxQ were selected, based on highest stresses on the connection points (metallic parts) and highest failure index in the composite shell and flange.The directions are aligned with locks and aligned with the longitudinal opening (0º and 315º in Figure 2).For this paper only the test results with 0º alignment are presented since it presented the absolute worst case for the majority of parts.
The GFEM, was reduced into a Test FEM (TFEM), using the same nonlinear and pretension effects, but using the boundary condition defined by the test setup which mean interfacing with the test rig and with the slings, shown in Figure 3.The purpose of the TFEM was to replicate the physical test, hence validate the test procedure in terms of the behaviour of the fairing structure when subject to qualification verification loads and compared to the GFEM MaxQ loads.Tension, compression, bending, shear and density properties were obtained from sample testing and used to define the composite properties in the FEM.The use of this type of solution allowed for the extraction of resulting forces, displacement, and strain data for intermediate load steps.The calibration test provided data for a correlation between Test FEM and manufactured prototype.This data was useful besides the correlation effort between FEM and physical testing, for an assessment of how well the test setup replicated flight loads.
Figure 4 to Figure 5 show the longitudinal strain shown in each of the halves of the fairing comparing the GFEM and TFEM.The skin strains show some differences in distribution, from GFEM to TFEM, but caused a similar critical spot (location and intensity) leading to the assumption that the interface loading was going to be realistic when the test loads, described in chapter 7 were applied.

Test levels
The test levels for acceptance and qualification tests were defined according to dedicated standards ( [2], [3]), utilizing test factors KA (acceptance test factor) and KQ (qualification test factor) applied to the design limit loads (DLL).The objectives for the three qualification levels are as follows: • First qualification level: load factor equal to the design limit load (DLL), to verify functional requirements and manufacturing adequacy, corresponding to acceptance tests level LCC001.
• Second qualification level: load factor equal to the design yield load (DYL=1.11xDLL), to verify the occurrence of permanent deformation, equal to qualification tests level LCC002.
• Third qualification level: loads factor up to design ultimate load (DUL=1.26xDLL), to verify rupture/fracture requirements, equal to qualification test level LCC003.Figure 9. FBD: bending moment.

Static test load cases
Figure 7 to Figure 9 shows a free body diagram (FBD) along the height of the FRG.The plots show the load case results obtained from the GFEM (MaxQ for 0 degrees), represented as a solid line with circles.
The FBD results were used to define the DLL from which the test load cases were derived.The aim was to achieve a similar behaviour along the x-axis, and to achieve the resultant force at the root, a critical part of the structure.The test load cases, indicated by LCC, are also shown in the plots.
For the axial direction, the load magnitude was defined to encompass any resultant force occurring in flight load cases.The small peak observed on the Max Q curve was attributed to a numerical phenomenon and not considered in this evaluation.So, the applied load was defined as the maximum value of the axial force from the FBD, which can be seen in Figure 7.
For the lateral load, the magnitude and application point were defined to match the resultant lateral load and bending moment at the root of the FRG.To achieve both, the LL1 was set at x=70% and the LL2 was set at x=40%. Figure 8 shows the distribution of lateral load along the x-axis of the FRG and how it compares to the flight load case, surpassing the flight load curve at the FRG root. Figure 9 shows the FBD bending moment between test setup and flight conditions, with the test setup surpassing the flight load case at the FRG root.

Calibration load cases
For the calibration test, a single load was applied in turn for each Wheatstone Bridge.Four bridges were instrumented on the FRG: two full bridges in a bending configuration (BB), with BB1 and BB2 located at x=48% and x=24%, and two half bridges in an axial configuration (Ax), with Ax1 and Ax2 located at x=73% and x=21%.The BB bridges were calibrated with a load applied in a horizontal direction, and the Ax bridges were calibrated with a compression load applied in a vertical direction.The magnitude of the loads was defined as a fraction of DLL, to avoid damaging the structure.
The root of the FRG (x=0%) was identified as the most critical region for the resultant lateral and bending moment loads.The calibration lateral load, applied at x=100%, was defined as 32% of DLL.This load would induce a bending moment of 62% of the DLL for x=0%.
The axial load was reduced to 62% in respect to the DLL.The magnitude of the applied load was sufficient to compute the linear correlation of the bridge output and the induced load, while also providing test results to correlate the measured deformation values and the FEA predictions.

FRG EM calibration
The aim of the calibration test is to establish a linear relationship between the independent variable and the dependent variable.In this context, the independent variable is the bridge results and the dependent variable is the induced axial load for the axial bridge and the bending moment for the bending bridge.This method yields a regression coefficient that can be applied to the bridge to measure the local bending moment or axial load value of the instrumented region.
The electrical signals from the strain gauge bridges and the data from the load cells were analysed by plotting the bridge response (mV/V) as a function of applied loads.Figure 10 shows the linear regression plot for the bending bridges (BB1, BB2) and axial bridges (Ax1, Ax2).Table 2 lists the R2 and standard deviation for each bridge regression.The test was performed with steps of 20% of DLL until reaching 60% of DLL, followed by 10% increments.After reaching the functional level, the loads were further increased by an additional step of 11% of DLL to achieve the first qualification level, DYL (LCC002).
The test was successful, meeting and exceeding the targets for applied lateral load and bending moment.Slippage was observed at the interface between the FRG main assembly and the TSF during each load step, but no permanent damage was identified.
The load cases applied during the test campaign were validated with the test data recorded by the loads cells measuring the applied load of each hydraulic actuator (load cells LC1, LC2, and LCAx were used to measure LL1, LL2, and AxL, respectively) and the Wheatstone bridges located on the outer surface of the FRG (at different heights, as indicated in section 8.1).The applied loads and bending moment derived from the test sensors were compared with the results of the GFEM (MaxQ) and the TFEM (LCC002), verifying that the functional level and DYL were met.

Moment Vs test load.
The FRG structure was subjected to lateral loads, LL1 and LL2, with the purpose of inducing a bending moment and measuring its responses.The bridges BB1 and BB2 responses were recorded and compared to those obtained from the GFEM and TFEM.
Figure 11 shows the test results of the applied moment as a function of the total lateral load, defined as the sum of LL1 and LL2, with the representation of functional level and DYL.The two lateral loads were measured with load cells.The bending moment was measured with the two bending bridges (BB1 (SG) and BB2 (SG)).An analytical bending moment was calculated with the value of the lateral loads and the distance from the application points.The analytical bending results were denoted as BB1 (LC) and BB2 (LC).The BB2 (SG) response shows a similar progression to the analytical prediction of BB2 (LC), with both exceeding the DYL level.BB1 (SG) exceeded the predicted BB1 (LC), and reached the target acceptance level at 65% of the total lateral load.This may indicate that the FRG was more heavily loaded at this location.

FBD flight case Vs test case
Figure 12 shows the lateral load FBD along the FRG x-axis.The dash-point line represents the GFEM (MaxQ).The dashed line represents the TFEM (LCC002).The solid line represents the lateral load distribution applied by the actuators along the x-axis, represented by LL(x).The plot in Figure 12 shows that the LL test results were achieved and exceeded the DYL level.Additionally, it shows that the test load distribution follows a similar behavior as the TFEM.The LL for the test was 2.5% higher than the TFEM, and the maximum LL value exceeded the GFEM values by 12%, as shown in Table 3.
The excess load was a result of the identified slippage.With each instance of slippage, there was a displacement of the "fixed" end of the FRG, leading to a simultaneous relief of stress and a reduction in the measured BB values.The lateral load values were increased to achieve the target bending value.Table 3.Comparison of lateral load test values vs TFEM (LCC002) and GFEM (Max Q).
Figure 13 shows the axial load FBD along the FRG x-axis.The dash-point line represents the GFEM (MaxQ), while the dashed line represents the TFEM (LCC002).The solid line represents the axial load distribution applied by the actuators along the x-axis, denoted as Ax(x).The Ax(x) was analytically calculated based on the geometric characteristics of the test setup (height of actuators, inclination of EASN-2023 Journal of Physics: Conference Series 2716 (2024) 012090 slings, etc.).The plot also includes the local bridge results of Ax1 and Ax2 at the moment of the maximum applied load, represented as single dots in their respective x-axis positions.
The GFEM curve was compared with the TFEM curve, and a new axial load value was defined for the test.The analysis showed that a lower axial test load was necessary to achieve an identical resulting load for TFEM and GFEM for x=0%.Therefore, the applied axial load during the test (Ax(x)) would be lower than previously defined.
The Ax bridge results show a difference between the applied loads and bridge results.For 100%≤x<-70%, the bridge Ax1 shows a value that is 19% higher than the applied load and 21% higher than the TFEM.Ax2 shows a measured result close to the TFEM (<2%), but more distant to the analytically calculated Ax(x) (53.4%), as shown in Table 4.
Figure 14 shows the bending moment FBD along the FRG x-axis.The solid line represents the bending moment distribution applied by the actuators along the x-axis, denoted as Mx(x).The Mx(x) was analytically calculated based on the geometric characteristics of the test setup and the distance of the load application point to x=0%.The plot also includes the local bridge results of BB1 and BB2 at the moment of the maximum applied load, represented as single dots in their respective x-axis positions.
The plot demonstrates that the bending moment M(x) (solid line) and the BB values follow the same tendency.This indicates that the bridges were able to measure the bending as intended.Figure 14 also shows that the bending values were achieved and surpassed the GFEM values.The bending bridges also exceeded the TFEM model prediction by 8% on BB1 and 1% on BB2, as shown in Table 4.

Model correlation
Taking into consideration that the lateral loads and bending moments met the specified test targets and were in line with the TFEM, a comparison of strain results and measurements was conducted.This analysis provided additional insights, contributing to the validation of the TFEM model.The strain state on any point of the payload FRG under load was the result of two distinct cases applied in sequence, pre-tension and load application.While FEA allowed the extraction of strain values at the beginning or end of each of these cases, these results were not easily obtained during static testing due to practical difficulties of obtaining data when the FRG was being assembled and fastened.In order to compare virtual and testing data, the TFEM results were adjusted to best fit the initial measured strain state.Figure 14 gives an overview of the single strain gauge (SG) location which will be used for this comparison.The SGs, referred to as CSF, were instrumented near separation locks of the FRG, previously identified as the most strained location, as shown in Figure 5.
As a first assessment of the overall behaviour of the structure when subjected to loads, the axial strain values occurring on the FRG skin (SA3) were evaluated (Figure 16), as this area was not expected to be influenced by localized effects, based on the TFEM results.The good correlation observed showed that the general strain distribution should be close to that predicted by FEA and that further comparisons on other areas with more complex distributions could proceed.
The strain measurement points closest to the separation locks of the FRG YNEG (CSF AY, CSF AZ, CSF TY, CSF TZ) provided indirect information on the stress state of the entire lock area.A more detailed approach was taken on this point and the strain variation was examined in its axial (A) and tangential (T) components.Figure 17 and Figure 18 show the similarity between the results from TFEM and test result, that indicate that the TFEM model provided a good characterization of lock area.

Conclusion
The FRG structure was tested until DYL, maintaining its functionality and structural integrity throughout the test campaign, although uncertainties associated to the applied axial load mean that it cannot be stated that acceptance and qualification levels were met.
The bending bridges proved to be an effective and accurate method of measuring applied moment.That allowed to validate the application of the lateral loads.
The idealized test setup provided a satisfactory replication of the flight loads on the FRG structure, as it generated a good representation of the resulting load in critical areas (separation locks).The free body diagram obtained from the FEA provided the necessary information to compare the resulting forces between flight and test conditions.
The successful application of the lateral loads and bending moment enabled for strain results to be compared between the test setup and the finite element analysis.The good correlation observed at the selected measurement points allowed the FEA model to be validated as a development tool.
The next phase is planned to perform the EM_S_STR_Q_2 test, qualifying the structure at DUL level with revised instrumentation and more accurate measurement of the axial load.

Figure 1 .
Figure 1.Overview of the FRG and 2nd stage TSF.

Figure 3 .
Figure 3. Set up of static test jig.Figure 4. GFEM vs Test FEM: skin strain on tension side.

Figure 4 .
Figure 3. Set up of static test jig.Figure 4. GFEM vs Test FEM: skin strain on tension side.

Figure 5 .
Figure 5. GFEM vs Test FEM: highlighted area of most strained location.

Figure 10 .
Figure 10.Bending and axial Bridges calibration: linear regression of the mV/V to % of maximum load

Figure 11 .
Figure 11.Total lateral load Vs bending moment in percentage of maximum DLL.

Figure 16 .
Figure 16.Comparison of SA3 test result Vs FEA model.

Figure 17 .
Figure 17.Comparison of CSF axial test result Vs FEA model.

Figure 18 .
Figure 18.Comparison of CSF tangential test result Vs FEA model.

Table 1 .
Static test matrix for EM test campaign.

Table 2 .
Calibration statistics for bending and axial bridges

Table 4 .
Bridge test values comparison with flight FEA (Max Q) and FEA model (LCC002)