Use this URL to cite or link to this record in EThOS:
Title: Shape error modelling, simulation and detection for free-form compliant surfaces
Author: Babu, Manoj Kumar
ISNI:       0000 0004 9358 0146
Awarding Body: University of Warwick
Current Institution: University of Warwick
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Thesis embargoed until 29 Jan 2022
Access from Institution:
Manufacturing processes should meet competing requirements of being efficient, environmentally friendly and deliver high-quality products, whilst minimising cost and shortening time to market. These requirements necessitate reducing the time involved in product development and production ramp-up, which can then consequently shorten New Product Introduction (NPI) time. Moreover, the shortening of NPI time must be accomplished without compromising the product quality while simultaneously meeting the requirements of frequent quality improvement. A major factor affecting the NPI time of a new assembly system is the challenge of achieving the required product quality right- first-time. This often requires expensive changes to the design of products and processes after the freezing of design. The inability to prevent these design changes arises due to simplistic assumptions related to part and process variations during the design process. In the automotive and aerospace industries, a large part of these changes (60-70%) is conducted to correct quality problems caused by product geometric and dimensional variations. This is especially critical for free-form surfaces, which have high aesthetical and functional requirements as de ned by tight geometric dimensioning and tolerancing (GD&T). Dimensional and geometric deviations of free-form surfaces can lead to numerous quality related problems including (i) high rate of re-work or scrap, (ii) inferior product functional performance, (iii) tooling failures, and, (iv) unexpected production downtime which in turn, reduces both product quality and production throughput. These factors together with the extensive use of free-form surfaces in automotive and aerospace industries have prompted a critical need for advancements in (i) modelling and simulation of 3D free-form surfaces shape errors during the early design phase; and, (ii) inspection of geometric and dimensional characteristics by using 3D scanners with capabilities for short mean-time-to-detection (MTTD) of quality faults. This thesis addresses the needs for (i) modelling and simulation of 3D free-form surface shape errors at the early design phase by developing a novel morphing Gaussian Random Fields (mGRF) methodology; and, (ii) inspection of 3D free-form surface shape errors with short MTTD by developing an innovative Spatio-Temporal Adaptive Sampling (STAS) methodology for deviation pattern prediction using 3D-surface scanners. The developed mGRF methodology, during the early design phase, generates non-ideal part instances, i.e., parts with shape (form) error, which exhibit spatial correlation pattern similar to the true manufactured part and conform to GD&T form tolerance specifications. The methodology first, models the spatial correlation in the deviations of the part from its design nominal using Gaussian processes and next, utilises the modelled spatial correlations to generate non-ideal part instances by mGRF through conditional simulations. The methodology has the capability to: (i) model the spatial correlation in part deviations with a single manufactured part measurement data, or by using historical measurement data of a similar part; (ii) model shape variations in a batch of parts; and, (iii) simulate various design intents by generating specific global and local deviations. The aforementioned capabilities enable to accurately model and predict the effect of part shape variations on product quality during the early design phase, and therefore prevent unnecessary design changes during later stages of production. The developed STAS methodology for free-form surface deviation pattern prediction estimates the whole part deviation based on partial measurement of a free-form surface. This is done by spatio-temporal sampling of a part surface, by first modelling spatio-temporal correlations in a batch of training parts using space-time Kalman filter Then, applying this model during the inspection process to (i) iteratively select the next region to be measured according to a predefined coverage criterion; and, (ii) determine the total number of regions necessary to be measured in order to estimate the whole part deviation within a predefined accuracy. As an outcome, the proposed methodology minimises necessary part inspection coverage, thereby, reducing inspection time needed for a given part produced using a specific manufacturing process. The developed methodologies are demonstrated, verified and validated with industrial case studies using an automotive door inner part. The ability of the developed mGRF methodology to simulate non-ideal part instances utilising a single part measurement, and with historical measurement data is demonstrated using an automotive door inner part. Additionally, the simulation of parts conforming to given GD&T form specifications, and various design intents are also demonstrated. The developed STAS methodology for freeform surface deviation pattern prediction is first, compared with state-of-art methodologies and then verified and validated by measuring the automotive door inner part with a robotic 3D-surface scanner. The test study demonstrated an average cycle time (CT) reduction of 42.18% from 510.5 seconds required to measure the whole part, to 295.18 seconds required to partially measure (approximately 33%) the part, with an associated 3-Sigma prediction error of 0.27 mm. This reduction in CT enables the effective in-line use of 3D-surface scanners leading to shorter MTTD of quality faults related to free-form surface shape errors. Results contribute towards creating a robust design and an adaptive assembly system that taken together mitigate the effects of geometric and dimensional variations and act as a critical enabler for near zero-defects manufacturing strategy.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available
Keywords: TL Motor vehicles. Aeronautics. Astronautics