On semi-active inerters for improving machining productivity

The inerter is a mechanical element, synthesised in 2002 as an analogue to the electrical capacitor. Originally used in Formula 1 racing as the `J-damper', its potential has since been explored in other vehicles, as well as for vibration control of civil structures. In very recent years, some study has been given to the design and control of semi-active inerters. Such devices would be capable of varying their inertance in response to a control signal.

To date, no study has been made of the semi-active inerter in the context of machining chatter. This undesirable form of vibration, leading to poor surface finish on machined parts, is a major issue in machining. The growing requirements of high speed machining of lightweight, flexible parts mean that the need to develop new strategies to tackle chatter will only increase. This thesis seeks to fill this gap in the literature.

As a feasibility study, two chatter suppression strategies are developed using a simplified single degree of freedom chatter model. Both strategies assume the existence of an ideal semi-active inerter placed between the vibrating element and ground, allowing the natural frequency to be adjusted on-line. The first of these strategies, discrete inertance variation, is analogous to an existing lobe seeking strategy conducted by changing the spindle speed. It is shown that, with relatively modest ranges of inertance, this is an achievable strategy for high speed machining. The second strategy relies on cyclically adjusting the natural frequency to disrupt self-excited vibration. It is found that the amplitude of this variation is the important characteristic, rather than the ratio of the frequency of inertance variation to the tooth passing frequency. In both cases, the need to be able to rapidly control inertance is noted.

The design needs of a semi-active helical inerter are considered, with magnetorheological fluid providing the semi-active control. Three different layouts are studied using quasi-static models. The bypass valve type layout is selected as the most promising for future study. The design of the valve is considered and a new optimisation scheme is developed which better suits the need of the bypass valve than previous schemes. The inerter model is extended into a quasi-dynamic model, which allows the varying inertance to be considered. This model would be key for developing any practical control scheme.

Prototype inerters were designed and tested. Initially an oil-based designed is built, followed by a design using magnetorheological fluid. The prototype was tested using a servo-hydraulic actuator, with the goal of validating the models developed in the previous chapter. Unfortunately, trapped air in both systems led to these results being inconclusive in both cases. The use of magnetorheological fluid for flow directional control in this way is unusual at this scale and this work is important for any future researchers who wish to work with the fluid in this way.

With this in mind, the issues encountered with the experimental rig are further analysed. Improvements to the design and filling method are proposed. Some more substantial design changes are also presented. Finally, some focus is given to the practical issues of implementing semi-active inerters in machining. The need to miniaturise the design to fit into modern machine tools is highlighted. Two areas in which this would be less of an issue -- fixturing and robotic machining -- are discussed. Notably, key challenges for robotic machining include the number and placement of the inerters, and whether new strategies would be needed to tackle mode-coupling chatter.