Base force and moment based finite element model correlation method
Introduction
Structural analyses are widely utilised to predict the behaviour of a complex system under both dynamic and static loading conditions. This helps to design the structure optimally, verify the design margins prior to the qualification test, and to plan the tests for design qualification. A Finite Element Model (FEM) created using the commercially available software is generally employed for the analyses of complex structures. The accuracy of the analytical results is critical, especially with aerospace structures such as spacecraft to minimize the launch cost. Inaccuracy in the analytical model also leads to improper load estimates that may lead to hardware failures. Conventionally, the accuracy of analytical results is assured by comparing the mode shapes extracted from the modal test with those obtained from the analysis using vector correlation methods such as the modal assurance criterion (MAC) (Allemang, 2003, Allemang and Brown, 1982), the coordinate modal assurance criterion (Lieven et al., 1988) and the normalized cross-orthogonality (NCO) check (Ewins, 2000). The NCO check requires a compatible mass or stiffness matrix that suits to the mode shapes to perform the correlation and it can be done using the system equivalent reduction expansion process (SEREP) (Mercer, 2016, O'callahan et al., 1989) whereas the MAC is performed with two equal size mode shapes. The SEREP uses generalised inverse (Penrose, 1955) to reduce the analytical mass or stiffness matrices to the test location. The frequencies obtained from the test and analysis are directly compared(Siano, 2015). It is accepted that the analytical model passes the quality check if it possesses certain threshold values of MAC or NCO along with the frequency variations within a certain range as specified by the different space agencies (Loads analysis of spacecraft and payloads, 1996, Modal survey assessment, 2008). However, one has to note that even if the analytical model passes all the standard criteria, considerable errors can arise in the predictions of structural responses such as peak acceleration under the forced vibration (Sairajan et al., 2015). The acceleration responses and the stiffness requirements are very important to check and verify the design adequacy of the structural elements, load analysis, and in the design of those subsystems to be attached to the main structure.
Model correlation can also be performed by comparing the frequency response functions obtained from the test and the analysis. Two such correlation criteria are the frequency response assurance criterion (FRAC) (Nefske and Sung, 1996) and the frequency domain assurance criterion (FDAC)(Pascual et al., 1997). In FRAC, the frequency response functions obtained from the test and the analysis are compared against each frequency, whereas the FDAC uses the MAC approach for the response function correlation. In a complex structure, there could be hundreds of response locations that need to be monitored during the test, and hence correlation will be a tedious task. Besides this, the responses highly depend on the locations of measurements, and hence very careful evaluation is needed for the successful correlation. The advancement of dynamic force measurement device (FMD) that can be used during the shaker test or the based fixed shaker driven test, has led to the development of the base force assurance criterion (BFAC) (Sairajan and Aglietti, 2014) to assess the quality of finite element models to predict the dynamic force response characteristics. The BFAC overall shows how well the FEM can predict the forced response characteristics primarily during the major modes of excitation under the base excited conditions. In the case of spacecraft, coupled load analyses (CLA) are performed for the critical load cases that may be encountered during the launch, to determine the critical responses, interface forces, and moments after integrating the reduced model of spacecraft with the launch vehicle. The CLA will be performed by an external launch vehicle team (Lim, 2014) and it is a time consuming and expensive process. The BFAC can also be used to decide whether CLA needs to be performed again after incorporating certain design changes in the spacecraft and FEM by correlating the base forces from the original FEM and the revised FEM. It was also shown that BFAC is more effective than the MAC or NCO check in the prediction of loss factor for visco-elastic systems (Sairajan et al., 2014).
Base force assurance criterion requires interface forces at the boundaries obtained from the mathematical model and the dynamic test. However, complex mathematical models of structures are created based on many assumptions such as rigid boundary conditions and uniform behavior of identical substructures. Besides, there are modeling limitations in the real loading conditions (Abdullah, 2017) and simplified modeling techniques are adopted to represent the joints to limit the overall degrees of freedom (DOF) of the model. The experimental results are also affected by the noise in the measurements, sensitivity of the instruments, non-perfect control system, and signal delays (Marcello et al., 2017). All these lead to inaccuracies in the results. Such inaccuracies are unavoidable in any real system (Remedia et al., 2015). An exact error analysis of a simple beam with a single design variable was performed by Olhoff and Rasmussen (Olhoff and Rasmussen, 1991). Generally, for complex systems, probabilistic methods are adopted to find the sensitivity of inaccuracies (Capiez-Lernout et al., 2006, Chowdhury et al., 2011, Remedia et al., 2015, De Lellis et al., 2020). A probabilistic study to assess the effect of errors in the experimental mode shapes on the test analysis orthogonality check of a spacecraft was performed by Bergman et al. (Bergman et al., 2010).
This work extends an earlier reported work (Sairajan et al., 2015) and here the experimentally determined moments are also included in the computation of BFAC for a flight-worthy spacecraft under the base excitation in two directions. To calculate the base moments using the experimental force data, it is required to separately acquire the three mutually perpendicular components of the forces using tri-axial force gauges. For computing the BFAC, there is no need to perform any additional test like modal testing for MAC or NCO, as the required measurements will be available from the base fixed sine excitation test, done while qualifying aerospace structures such as a spacecraft. Monte Carlo simulations are used to assess the robustness of BFAC to the inaccuracies that may be presented in the FEM and/or in the experimental results using a simplified error model. The procedure is then demonstrated using a mini-satellite model and its experimental results.
Section snippets
Theoretical background
The base excitation analyses of aerospace structures are carried out using the Craig-Bampton reduction method(CRAIG and BAMPTON, 1968, Hurty, 1965, Primer on the Craig-Bampton Method, 2000). Here, this analytical method is summarised. In the Craig-Bampton reduction, the boundary degrees of freedom, R, are retained in the physical coordinates and internal degrees of freedom, L, are retained in the model coordinates. To perform the reduction, the mass, the stiffness, and the force matrices are
Finite element model
A fully integrated flight-worthy mini satellite developed by ISRO is considered in this study. The FEM of the spacecraft along with the axes system is shown in Fig. 1. It is formed by 615,647 grids and 501,675 elements. This figure also shows the locations of seven accelerometers marked 1 to 7. The FEM is generated using MSC Patran and the Craig-Bampton reduced stiffness and mass matrices were extracted using MSC Nastran (2001).
The free vibration characteristics with base fixed boundary
Experimental setup
The spacecraft was tested with sine inputs at the base along the longitudinal and the lateral axes as per the launch vehicle specified sine loads using the shaker system test facility at U R Rao Satellite Centre. The base force was measured during the tests along all three axes for the estimation of the response at the centre of gravity of the spacecraft and to conduct the tests with due precision. However, only lateral axes test results were used in this study. A special setup called FMD was
Robustness of base force assurance criterion
The inaccuracies, either in the mathematical model or in the experiment, can be represented using a simplified multiplicative error model with Monte Carlo simulations and then applied to assess their impact on the dynamic behavior of the complex structures (Sairajan and Aglietti, 2012). A similar model is used in this work and the inaccuracy, E is defined by the equation:where, J is a unit-square matrix with dimensions of the reduced mass or stiffness matrix to simulate
The BFAC and acceleration response
Base fixed response analyses of the spacecraft under sine excitation were carried out in the frequency range of 5.0–100.0 Hz with the help of Matlab (MATLAB, 2011) script using the Craig-Bampton reduced stiffness and mass matrices. The chosen frequency range is based on the specification from the launch vehicle. The amplification of acceleration responses or transmissibility for 1 g excitation at the base was performed for both the lateral axes Y and Z. As stated in the Table 1, there were no
Conclusions
The base force and moment based correlation method known as base force assurance criterion is applied to the model correlation of a real spacecraft. A custom made force measuring device was used during the base sine excitation to measure the base forces, and then the base moments were calculated using the measurements from six Tri-axial force gauges. These forces and moments are then correlated with the FEM predicted force and moments to obtain the BFAC correlation matrix. It is demonstrated,
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors wish to acknowledge Mr. D. Premkumar, UR Rao Satellite Centre, Bangalore for providing the finite element model for this study; Mr. S. Shankar Narayanan, Mr. K.V. Muralidhar and Mr. N. Ravikiran, UR Rao Satellite Centre, Bangalore for their help to finish this study.
References (32)
- et al.
Probabilistic investigation of sensitivities of advanced test-analysis model correlation methods
J. Sound Vib.
(2010) - et al.
Data and model uncertainties in complex aerospace engineering systems
J. Sound Vib.
(2006) - et al.
Probabilistic fracture mechanics by using Monte Carlo simulation and the scaled boundary finite element method
Eng. Fract. Mech.
(2011) - et al.
Structural uncertainty estimation through a craig-bampton stochastic method optimisation in satellites structures
J. Sound Vib.
(2020) Sensitivity analysis and correlation experimental/numerical FEM-BEM for noise deduction assessment of an engine beauty cover
Energy Procedia
(2015)A correlation study of satellite finite element model for coupled load analysis using transmissibility with modified correlation measures
Aerospace Science and Technology
(2014)- et al.
A stochastic methodology for predictions of the environment created by multiple microvibration sources
J. Sound Vib.
(2015) - et al.
Correlation of finite element models of multi-physics systems
Journal of Sound and Vibration
(2014) Correlation of numerical and experimental analysis for dynamic behaviour of a body-in-white(BIW) structure
MATEC Web of Conferences
(2017)The modal assurance criterion - Twenty years of use and abuse
Sound and Vibration
(2003)
Coupling of substructures for dynamic analyses
AIAA J.
Modal Testing Theory, Practice and Application
Reduction of mass and stiffness matrices
AIAA J.
Dynamic analysis of structural systems using component modes
AIAA J.
Cited by (1)
Determining the Sensor Locations for an Effective Modal AnalysisWith an Application to a Car Exhaust
2023, International Journal of Acoustics and Vibrations