This paper presents the design and the initial experimental results of a novel impact oscillator rig developed by the Centre of Applied Dynamic Research at the University of Aberdeen. In this rig, the excitation force is generated electromagnetically and it acts directly on the mass in contrast to the most of the experimental set-ups where the excitation passes through the structure. This significantly enhances flexibility of the system allowing to observe subtle phenomena. The evolution of the design from an initial concept to the finalised rig is discussed in details where a special attention is paid to the instrumentation and parameter identification which are important for the mathematical modelling. The initial experimental results demonstrate potentials of this rig to study fundamental impact phenomena, which have been observed in various engineering systems. They also indicate that this new rig can be a good platform for investigating nonlinear control methods.

This work explores complex dynamics of a new mass excited impact oscillator reported in Wiercigroch et al. (Nonlinear Dyn 99:323–339, 2020) both experimentally and numerically in the context of development of chaos theory and its applications. The parameters of the rig were characterised and are presented in the paper. To improve quality of the recorded phase portraits, a new technique for processing of the experimental data allowing to reduce the influence of noise and to obtain clear orbits especially for higher periods is proposed. A comparison with the previous studies on the base excited impact oscillator confirms that the rig is much more accurate as well as it has capability to generate a wide range of excitation patterns. It is demonstrated that a precise control of the excitation is achieved by changing the coil current. It is also shown that the rig is able to capture co-existent attractors and multi-stability by reproducing various predicted numerical responses, which has not been possible before. The results obtained using a simple impact oscillator model are in a good agreement with the experimental results, which indicates that the rig can be used for further fundamental studies of impact phenomena including grazing. It can also serve as a tool to study nonlinear control including bifurcation control and control of co-existing orbits.

Vibro-impact phenomena in engineering systems, considered an adverse effect in some settings, are an intrinsic part of the mechanism in others. In energy harvesting, a vibro-impact component is often intentionally introduced to increase the power output or the system’s bandwidth. The impacts can be treated as “hard” for instantaneous impacts or “soft” for compliant materials. Since both types of models exhibit complex dynamics, a comparison is non-trivial. We develop a soft impact model for a vibro-impact energy harvester, calibrating it with the relevant hard impact model for large stiffness, and systematically compare the different phenomena and dynamics in various compliant regimes. Numerical results are used in two different parametric analyses, considering the bifurcation diagrams in terms of device size and external forcing parameters. Varying the natural frequency of the membranes that form the impact boundaries, we observe shifts in the bifurcation structure that promote period-1 orbits for increased softness parameters, often generating higher power output, but also introducing parameter sensitivities for increased softness. Complementary analytical results reveal unstable periodic orbits and co-existing behaviors, potentially missed by computational methods, that can influence the bifurcation structure and in turn the energy output. A non-dimensional formulation highlights the significance of ratios of external and natural frequencies in delineating soft and hard impact scenarios parametrically. The soft impact model exhibits new symmetry breaking bifurcations related to key quantities that characterize the soft impact dynamics, such as the effective restitution coefficients, the impact phase, and the contact time interval, not captured by hard impact models.