Abstract:The identification of time-varying physical parameters of nonlinear systems is still a challenging task. Limited studies based on the wavelet multiresolution analysis (WMA) have been attempted, which requires full measurements of structural displacement, velocity and acceleration responses of all degrees of freedom and exact information of external excitations. This limits the engineering application of these methods. This paper proposes approaches to identify the time-varying physical parameters of nonlinear structures in three cases using only partially measured structural responses. Firstly, the identification of time-varying nonlinear structures with a small number of elements under known excitations is discussed. The fading-factor unscented Kalman filter (FUKF) method is applied to locate the time-varying parameters, and the WMA integrated with UKF method is employed using partially measured acceleration responses. Secondly, it is further extended to the identification of time-varying nonlinear structures with a small number of elements but under unknown excitations. An improved fading-factor unscented Kalman filter under unknown input (FUKF-UI) method is proposed to locate the time-varying parameters, and WMA integrated with UKF-UI method is utilized with partially observed acceleration and displacement responses. Thirdly, for practical engineering applications, the identification of time-varying nonlinear structure with more elements under unknown excitations is conducted. The proposed FUKF-UI method is employed to locate the time-varying parameters of the whole structure. Then, the whole structure is divided into several substructures and the unknown interaction forces are regarded as the fictitious unknown inputs to the substructure. Thus, physical parameters of each substructure can be identified in parallel by the combination of WMA and UKF-UI. Three numerical studies corresponding to these three cases are conducted, respectively, to demonstrate the effectiveness and accuracy of the proposed methods.
What problem does this paper attempt to address?
The main problem that this paper attempts to solve is: how to identify the time - varying physical parameters of nonlinear structures with only partial measurement data, especially under unknown external excitation conditions. Specifically, the paper proposes new methods for the following three situations:
1. **Small - scale nonlinear structures with known external excitation**:
- Use the Fading - factor Unscented Kalman Filter (FUKF) method to locate the time - varying physical parameters.
- Adopt the method of combining Wavelet Multiresolution Analysis (WMA) and Unscented Kalman Filter (UKF) to identify the time - varying physical parameters by using the partially measured acceleration responses.
2. **Small - scale nonlinear structures with unknown external excitation**:
- Propose an improved Fading - factor Unscented Kalman Filter under Unknown Input (FUKF - UI) method to locate the time - varying physical parameters.
- Use the method of combining WMA and UKF - UI to identify the physical parameters based on the partially measured acceleration and displacement responses.
3. **Large - scale nonlinear structures with unknown external excitation**:
- Use the proposed FUKF - UI method to locate the time - varying physical parameters of the entire structure.
- Divide the entire structure into multiple sub - structures and regard the unknown interaction forces as "virtual unknown inputs", so that the physical parameters of each sub - structure can be identified in parallel.
Through these methods, the paper aims to overcome the limitations of existing methods that require complete measurement data and accurate external excitation information, and improve the feasibility and accuracy in practical engineering applications. The paper also verifies the effectiveness and accuracy of the proposed methods through numerical research.
### Formula Summary
1. **Equation of Motion**:
\[
M\ddot{x}(t)+R(h(t), x(t), \dot{x}(t)) = g f(t)
\]
where \(M\) is the mass matrix, \(f(t)\) is the known excitation, \(g\) is the influence matrix, \(x(t)\), \(\dot{x}(t)\), \(\ddot{x}(t)\) are the displacement, velocity and acceleration vectors respectively, \(h(t)\) is the physical parameter vector containing stiffness, damping and nonlinear model parameters, and \(R\) is the total restoring force of the structural system.
2. **State - Space Equation**:
\[
\dot{Z}=\begin{pmatrix}\dot{x}\\\ddot{x}\\\dot{h}\end{pmatrix}=\begin{pmatrix}\dot{x}\\M^{- 1}g f - M^{-1}R(h, x, \dot{x})\\0\end{pmatrix}+w
\]
3. **Discrete Observation Equation**:
\[
y_{k + 1}=\ddot{x}_{m,k + 1}=L_a\ddot{x}_{k + 1}+v_{k + 1}
\]
where \(y_{k + 1}\) and \(\ddot{x}_{m,k + 1}\) are the observed value and the measured acceleration respectively, \(L_a\) is the accelerometer position matrix, and \(v_{k + 1}\) is the measurement noise.
4. **Sigma - Point Calculation**:
\[
v_i^k=\begin{cases}\hat{Z}^k&i = 0\\\hat{Z}^k+\sqrt{(N+\lambda)\hat{P}_Z^k}(:, i)&i = 1,\ldots,N\\\hat{Z}^k-\sqrt{(N+\lambda)\hat{P}_Z^k}(
\]