Characterising the amplitude and mechanisms responsible for the pushout load and torque for a staked self-lubricating spherical plain bearing

Hatherell, Jacob

Abstract

Staking is a cold metal joining process widely used in the manufacture of self-lubricating spherical plain aerospace bearings. These are used to form components such as tie-rod links or threaded rod ends as often seen on control surfaces for rotary winged aircraft. The staking of a bearing is achieved by plastically deforming the bearing’s outer race to precisely retain the bearing assembly. Traditional analysis methods of these bearings include analytical methods, computational methods, or through manufacturing trials. Each of these methods have their own strengths and weaknesses with the choice of which method to use being a trade-off between time, cost, accuracy, understanding of the influencing variables, and generalising that understanding across a wide range of bearing geometries.

The work presented in this thesis details the development of a virtual design of experiments (Virtual-DoE) methodology to model the staking of self-lubricating spherical plain bearings by combining the two disciplines of computational modelling and the applied statistical methods of a design of experiments. The Virtual-DoE methodology allows for the creation of a rigorous and thorough test programme to be rapidly analysed within a computational modelling environment. This analysis method demonstrated its capability to identify all the relevant parameters that impact the staking process and to characterise their influence on the pushout strength and post-stake torque through a series of closed-form solutions.

The accuracy of these solutions was validated against manufacturing trails data over a period of 18 months with their performance far exceeding the accuracy of the traditional bearing analysis methods. The understanding of the fundamental mechanisms that control the staking process, enabled by the Virtual-DoE methodology, has allowed for both the optimisation of new bearing designs and a “first time right” capability that significantly reduces the likelihood of scrapped bearings during manufacturing or the need for costly and time-consuming manufacturing trials. The broad applicability of a Virtual-DoE provides an inexpensive, methodical, and scalable solution that could be applied to the majority of complex cold metal joining processes.

To support the computational model developed for the Virtual-DoE, ring compression tests (RCT) were undertaken to characterise the dynamic friction behaviour of the stainless steel that aerospace bearings are made from. A fundamental flaw with RCT was observed whereby the standard RCT analysis method cannot accurately model friction behaviour that dynamically changes with contact pressure. To address this flaw, a new iterative analysis method was proposed that compared to the standard method saw a five times reduction in modelling error.

• Thesis Type: Thesis
• Deposit Date: Oct 25, 2023
• Publicly Available Date: Jul 25, 2024
• Keywords: Design of Experiments, Computational simulation, Spherical Bearing
• Public URL: https://uwe-repository.worktribe.com/output/11392727
• Award Date: Jul 25, 2024