Unconditionally stable higher-order Newmark methods by sub-stepping procedure

doi:10.1016/S0045-7825(96)01243-1

Computer Methods in Applied Mechanics and Engineering

Volume 147, Issues 1–2, 30 July 1997, Pages 61-84

https://doi.org/10.1016/S0045-7825(96)01243-1 Get rights and content

Abstract

In this paper, a procedure to construct unconditionally stable higher-order step-by-step time integration algorithms based on the Newmark method is presented. The second-order accurate Newmark method is extended to third, fourth and even higher order by linearly combining the numerical results evaluated at some specific sub-step locations. The weighting factors and the sub-step locations are algorithmic parameters. They are chosen to eliminate the leading truncation error terms and to control the stability characteristics. The ranges of parameters for unconditionally stable third- and fourth-order algorithms are given. Besides, by using complex time steps, a third-orderL-stable and a fourth-order A-stable non-dissipative algorithms are derived. Furthermore, the accuracy of the excitation responses is shown to be improved by the present procedure as well. To maintain high-order accuracy, the excitation may need to be modified. The computational procedure is discussed. It is found that there are no changes in the implementation of the underlying Newmark method and no extra computations of high-order gradients. As the numerical results at the sub-step locations are obtained independent of one another, the evaluations can be computed in parallel. Comparisons with other higher-order algorithms are presented.

Reference (27)

HulbertG.M.
A unified set of single-step asymptotic annihilation algorithms for structural dynamics
Comput. Methods Appl. Mech. Engrg.
(1994)
ArgyrisJ.H. et al.
Non-linear oscillations using the finite element technique
Comput. Methods Appl. Mech. Engrg.
(1973)
ArgyrisJ.H. et al.
Higher-order methods for transient diffusion analysis
Comput. Methods Appl. Mech. Engrg.
(1977)
GellertM.
A new algorithm for integration of dynamic systems
Comput. Struct.
(1978)
BorriM. et al.
Dynamic response of mechanical systems by a weak Hamilton formulation
Comput. Struct.
(1985)
TarnowN. et al.
How to render second-order accurate time stepping algorithms fourth-order accurate while retaining the stability and conservation properties
Comput. Methods Appl. Mech. Engrg.
(1994)
HughesT.J.R.
The Finite Element Method: Linear Static and Dynamic Finite Element Analysis
(1987)
WoodW.L.
Practical Time-Stepping Schemes
(1990)
ZienkiewiczO.C. et al.
The Finite Element Method
(1991)
HairerE. et al.
Solving Ordinary Differential Equations I
(1987)

HairerE. et al.

Solving Ordinary Differential Equations II

(1991)

BaileyC.D.

Application of Hamilton's Law of varying action

AIAA J.

(1975)

BaileyC.D.

Hamilton, Ritz and elastodynamics

J. Appl. Mech.

(1976)

Cited by (39)

Directly self-starting higher-order implicit integration algorithms with flexible dissipation control for structural dynamics
2022, Computer Methods in Applied Mechanics and Engineering
An implicit family of composite $s$ -sub-step integration algorithms is developed in this paper. The proposed composite $s$ -sub-step scheme is firstly designed to satisfy two requirements. One is the directly self-starting property, eliminating any starting procedures and avoiding computing the initial acceleration vector. The other is identical effective stiffness matrices within each sub-step, embedding optimal spectral properties. The analysis reveals that the composite $s$ -sub-step implicit schemes with $s \leq 6$ can achieve $s$ th-order of accuracy when embedding the dissipation control and unconditional stability simultaneously, and that in case of $s \geq 7$ , the increase of accuracy requires more sub-steps. Then, only the first six economical composite multi-sub-step schemes are developed. Remarkably, two approaches are also constructed to output accurate accelerations, which is also regarded as another minor superiority. Unlike some published higher-order algorithms, the proposed methods do not suffer from the order reduction and they provide the designed order of accuracy for solving general structures. Linear and nonlinear examples are finally solved to confirm the numerical performance and superiority of the proposed methods.
An improved vibration control method of a flexible non-uniform shaped manipulator
2021, Simulation Modelling Practice and Theory
Citation Excerpt :
Thus, exact dynamic responses of real systems are almost impossible to obtain. Numerical methods such as Newmark and Runge-Kutta can be used to solve mathematical models of dynamic systems [2–5]. Lazarus et al. [6] presents a finite element electromechanical reduced order model for the nonlinear vibrations of piezoelectric layered beams and using von Kârmân nonlinear strain–displacement relationship.
This paper presents an improved vibration control method that controls the amplitudes of residual vibrations of a flexible non-uniform shaped manipulator with higher suppression ratio. The proposed new method suppresses the residual vibrations that occur after motion by shaping the velocity input. By considering the parameters of the experimental system, an analytical model is established according to the natural frequency of the system which is effective in the vibrations that occur in the direction of motion. To demonstrate the accuracy of the established model, transient analysis is performed by the Fast Fourier Transform (FFT) method and the resulting dynamic response is compared with the response of the experimental system. The motivation of this study is to add the shaped exponential-harmonic velocity input to the end of the trapezoidal or triangular velocity input to suppress the velocity response that is in the region of the residual vibrations. The effects of changes in the amplitude, frequency, exponential decay factor, phase angle and application time parameters of the exponential-harmonic input on the control of the residual vibration amplitudes are examined. The obtained results are compared with the experimental results and it is observed that the proposed vibration control method has succeeded in suppressing the residual vibrations by up to 99%. Thanks to the proposed method, the time needed for the manipulator to reach the steady-state time at the end of the motion was reduced from 18 s to the motion time.
A dynamic response analysis method with high-order accuracy for fixed offshore structures based on a normalised expression of external loadings
2021, Ocean Engineering
A new step-by-step dynamic response analysis method is proposed for fixed offshore structures in the time domain with high-order accuracy. The basic principle of the proposed method is to normalise an arbitrary loading into the form of complex exponents with arbitrary-order derivatives, permitting the high-order Taylor expansion of dynamic response components (acceleration, velocity, and displacement) in theory. The proposed method is more accurate than other proposed methods, especially for structures subjected to short-term, high-energy loading with data collected using a limited sampling frequency. The response obtained has higher-order accuracy than that obtained with the widely used Newmark- $β$ method, which only has second-order accuracy because of the assumption of acceleration linearisation. Four examples are used to demonstrate the correctness and accuracy of the proposed method. The first numerical example uses a single-degree-of-freedom (SDOF) system to demonstrate the calculated results obtained using different orders of Taylor expansion with the proposed method. The comparison with the analytical result and the result obtained with the Newmark- $β$ method shows that the proposed method is less sensitive to the sampling frequency and that even the accuracy of the third-order Taylor expansion-based method is higher than that of the Newmark- $β$ method. In the second example, the third-order Taylor expansion-based method is used to calculate the dynamic response of a three-degree-of-freedom (3DOF) system under earthquake excitation sampled at a fixed frequency to show that the proposed method can determine the structural dynamic response under an arbitrary form of loading better accuracy than the Newmark- $β$ method. In the third example, a numerical monopile offshore wind turbine (OWT) is used for verifying the feasibility of the proposed method in calculating the dynamic response of fixed offshore structures. Finally, measurements of a cantilever beam with similar properties of monopile OWT under harmonic excitations and random excitations are used to demonstrate the proposed method and illustrate its potential values for application to dynamic response analysis of fixed offshore structures.
A fast processing method to perform transient analysis for vibration control
2020, Simulation Modelling Practice and Theory
Transient vibration analyses of structures with complex shapes are performed by finite element (FE) methods. Transient vibration analyses take long processing times for the systems having large number of degrees of freedom. A method performing the transient analyses very fast using the Fast Fourier Transform (FFT) is introduced in this work, and it is applied to a curved manipulator. Various analyses are required to determine the input parameters for vibration control, and each analysis takes long FE solution times.
In this work, an FFT method is developed where the samples of the impulse response of the system under study is obtained by the FE software ANSYS first. The samples of the transfer function are then obtained by FFT, and transient responses are found for various vibration control parameters by using FFT. Transient response results for the complex system considered in this study are obtained approximately in 1 s with FFT method, while it takes 36 h with ANSYS.
A one degree of freedom system is considered to verify the results of FFT method. Newmark's numerical, ANSYS, and FFT method results are compared, and it is observed that the results are in good agreement. A curved non-uniform steel manipulator with a complex shape is considered after the verification. The samples of the impulse response of the manipulator is found by ANSYS. Then, FFT method is used to obtain transient responses. A trapezoidal velocity profile is considered to analyze the residual vibration. The effect of the deceleration time on the root-mean-square (rms) values of the residual vibration is studied experimentally and using FFT method. It is observed that the results are in good agreement, and it is concluded that FFT method introduced in this study is very effective to study transient vibration problems in complex systems.
Electromagnetic wave propagation analysis by an explicit adaptive technique based on connected space-time discretizations
2018, Finite Elements in Analysis and Design
Citation Excerpt :
Implicit approaches, on the other hand, are generally unconditionally stable; however, they are usually characterized by higher computational costs. Many procedures can be employed to improve the stability and accuracy of time integration algorithms, such as subcycling techniques [12–14], high-order accurate schemes [15–20], automatic time step control [21–25], etc. As a matter of fact, a lot of research is continuously realized on this field and several time marching techniques are available nowadays for wave propagation analysis [26–39].
In this work, a new explicit time marching technique is considered to analyse 2D electromagnetic wave propagation problems. The technique considers adaptive time integrators, which are spatially and temporally locally computed, providing a connection between the adopted spatial and temporal discretizations. This approach allows the errors produced by both discretization procedures to be counterbalanced, enabling more accurate analyses. A multi time-steps methodology is also considered here, associated to subcycling techniques, enhancing the efficiency and the adaptive features of the method. As it is described along the paper, the new technique is very effective, robust and simple to implement, providing a very suitable numerical approach to analyse complex wave propagation models.
An unusual amplitude growth property and its remedy for structure-dependent integration methods
2018, Computer Methods in Applied Mechanics and Engineering
An unusual amplitude growth in the steady-state response of a high frequency mode for a structure-dependent integration method is numerically and analytically identified. This is a brand new type of amplitude growth and it has never been found in the conventional integration methods. The root cause of this unusual amplitude growth can be revealed by examining the local truncation error constructed from a forced vibration error, where the dominant error term for high frequency modes plays the key role for this unusual amplitude growth. An effective remedy is proposed by introducing a load-dependent term into the difference equation for displacement and/or velocity increment to remove the dominant error term. As a result, this adverse amplitude growth can be automatically removed. Consequently, the inclusion of the load-dependent term in the formulation of the difference equation for displacement increment and/or velocity increment is inevitable for a general structure-dependent integration method.

View all citing articles on Scopus

View full text

Unconditionally stable higher-order Newmark methods by sub-stepping procedure

Abstract

Comput. Methods Appl. Mech. Engrg.

Comput. Methods Appl. Mech. Engrg.

Comput. Methods Appl. Mech. Engrg.

Comput. Struct.

Comput. Struct.

Comput. Methods Appl. Mech. Engrg.

The Finite Element Method: Linear Static and Dynamic Finite Element Analysis

Practical Time-Stepping Schemes

The Finite Element Method

Solving Ordinary Differential Equations I

Solving Ordinary Differential Equations II

Application of Hamilton's Law of varying action

AIAA J.

Hamilton, Ritz and elastodynamics

J. Appl. Mech.