Toward fitting structured nonlinear systems by means of dynamic mode decomposition

Ion Victor Gosea,Igor Pontes Duff
DOI: https://doi.org/10.48550/arXiv.2003.06484
2020-11-23
Abstract:The dynamic mode decomposition (DMD) is a data-driven method used for identifying the dynamics of complex nonlinear systems. It extracts important characteristics of the underlying dynamics using measured time-domain data produced either by means of experiments or by numerical simulations. In the original methodology, the measurements are assumed to be approximately related by a linear operator. Hence, a linear discrete-time system is fitted to the given data. However, often, nonlinear systems modeling physical phenomena have a particular known structure. In this contribution, we propose an identification and reduction method based on the classical DMD approach allowing to fit a structured nonlinear system to the measured data. We mainly focus on two types of nonlinearities: bilinear and quadratic-bilinear. By enforcing this additional structure, more insight into extracting the nonlinear behavior of the original process is gained. Finally, we demonstrate the proposed methodology for different examples, such as the Burgers' equation and the coupled van der Pol oscillators.
Numerical Analysis
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: how to fit nonlinear systems with specific structures in the Dynamic Mode Decomposition (DMD) framework. Specifically, the author proposes an identification and order - reduction method based on the classical DMD method, which allows fitting nonlinear systems with bilinear and quadratic - bilinear structures to measurement data. This method pays special attention to two types of nonlinear terms: bilinear and quadratic - bilinear terms. By introducing these additional structures, the characteristics of the nonlinear behavior of the original process can be more deeply extracted. ### Main contributions of the paper 1. **Extension of the DMD method**: - The author extends the classical DMD method so that it can handle nonlinear systems with specific structures, especially bilinear and quadratic - bilinear systems. - These extended methods include DMDc (DMD with control input) and ioDMD (input - output DMD). 2. **Fitting of bilinear systems**: - For bilinear systems, the author proposes a method to fit nonlinear terms that contain the product of state variables and control inputs. - Specifically, for given measurement data, by constructing the matrix equation \( \mathbf{X}_{k + 1}=A\mathbf{X}_k+N(\mathbf{X}_k\odot\mathbf{U}_k)+B\mathbf{U}_k \), where \( \odot \) represents the Hadamard product, and \( \mathbf{X}_k \) and \( \mathbf{U}_k \) represent the snapshot matrices of the state and control input respectively. 3. **Fitting of quadratic - bilinear systems**: - For quadratic - bilinear systems, the author further extends the method to handle quadratic terms that contain the product of the state variables themselves. - Specifically, for given measurement data, by constructing the matrix equation \( \mathbf{X}_{k + 1}=A\mathbf{X}_k+Q(\mathbf{X}_k\otimes\mathbf{X}_k)+N(\mathbf{X}_k\odot\mathbf{U}_k)+B\mathbf{U}_k \), where \( \otimes \) represents the Kronecker product. 4. **Practical applications**: - The author verifies the proposed method through different examples, such as the Burgers equation and the coupled van der Pol oscillator system. - These examples demonstrate the effectiveness and applicability of the method, especially when dealing with complex nonlinear dynamic systems. ### Mathematical formulas - **Bilinear systems**: \[ \dot{\mathbf{x}}(t)=A\mathbf{x}(t)+N\mathbf{x}(t)\mathbf{u}(t)+B\mathbf{u}(t) \] \[ \mathbf{y}(t)=C\mathbf{x}(t)+F\mathbf{x}(t)\mathbf{u}(t)+D\mathbf{u}(t) \] - **Quadratic - bilinear systems**: \[ \dot{\mathbf{x}}(t)=A\mathbf{x}(t)+Q(\mathbf{x}(t)\otimes\mathbf{x}(t))+N\mathbf{x}(t)\mathbf{u}(t)+B\mathbf{u}(t) \] ### Conclusion This paper successfully fits nonlinear systems with specific structures to measurement data by extending the DMD method. This not only improves the understanding of complex nonlinear dynamic systems but also provides an effective tool for practical applications.