Physics-Informed Sparse Identification of a Bistable Nonlinear Energy Sink

Bistable nonlinear energy sinks have attracted extensive attention due to their efficient broad-band targeted energy transfer over a wide range of input energy levels. The precise identification of local bistability is of significance for predicting and controlling the system performance of vibration energy harvesting and absorption. This paper proposes a new physics-informed sparse identification method for parameter estimation of a three-degree-of-freedom bistable nonlinear energy sink. The restoring force surface is constructed on the local bistable structure, and the nonlinear elastic force trajectory is intercepted by assuming two quasi-zero velocity planes. Furthermore, the candidate functions of the nonlinear elastic force can be physically informed in the sparse identification algorithm. Numerical simulation shows that the proposed method not only gives sparse identification physics information but also improves accuracy by 1.61% under the noise level of 30 dB. Experimental verifications are performed on a three-story beam-type bistable energy sink. The result shows that the identified nonlinear elastic force has a good agreement with the measured one, and the NMSE is only 0.97%.