Light Controlled Factory
Call Us: +44 (0)1225 385138

Research Theme 1

Measurement Assisted Assembly Technology with integrated processing machines (Years 0-4)

Future AIT factories will require product specific customisation of assembly, ultimately adapting the condition of assembly for each part, whilst ensuring assembly integrity and high process yield. For this to be realised, the AIT factories would need to include appropriate semi-finishing processes that can be deployed to create and verify the condition of assembly of every major part. The two WPs are as follows:

WP1.1 Novel metrology assisted assembly with integrated processing technologies (Maropoulos, Mullineux, Keogh, Huntley). This WP will:

(i) Define and develop novel concepts for measurement assisted AIT factories. The research will define a theoretical framework for structuring measurement enabled automation in AIT environments using quantitative criteria such as assembly capability, scale, positional accuracy, product tolerances and production rates.

(ii) Investigate and define technologies to achieve a positional accuracy of 10 to 250 μm over large assemblies (10 to 30 m). LVM techniques such as laser trackers and iGPS will be considered in both static and dynamic modes. As there is little information and no standards on the dynamic tracking capability of such systems, this WP will investigate the dynamic performance characterisation of laser trackers, working closely with NPL. There will also be links with WP3.5 and 3.6 for parts pose identification and assembly re-configurability. Technical issues such as update rates, latency, jitter and noise will be overcome using estimation and filtering algorithms, and optimally fusing data streams from multiple sensors.

(iii) Define and develop methods for integrating in AIT environments processing machines for machining and material deposition on large metallic, composite and hybrid components, commonly encountered in the aerospace, renewable energy, civil nuclear and marine sectors. Such processing machines will be light-weight, modular and of low cost, relying on external guidance for precision and mounted on the assembly structure or be end-effectors of robots; replacing heavy, inflexible and capital intensive machines. Designs of Cartesian machines developed in Bath will be evaluated, together with commercial machines e.g. PKMs from Renishaw.

WP 1.2 Investigate the dynamic coupling of measurement and actuation for correcting errors (Keogh, Mullineux, Maropoulos). This key WP will investigate the closed loop, real-time integration of the measurement sensors and the control systems of the actuators, to correct errors and minimise the dynamic impact of processing machines on the spatial fidelity of the assembly. Centralised control algorithms will be designed to determine dynamic compensation from any multi-input (sensing) system through to a multi-output (actuation) system. H optimisation techniques will be evaluated, which embed appropriate filters. Parameter dependence (e.g., temperature, operating speed) may also be embedded using “linear matrix inequality” techniques to provide even higher order fidelity into the controller performance. Positioning errors will be eliminated by updating the AIT’s actuators using correction vectors computed from the external LVM sensors with target positional accuracy of between 10 and 250 μm over large assemblies (10 to 30 m). Dynamic positioning errors (above 5Hz) caused by vibrations from milling spindles (50 to 500 Hz), chattering, cutting forces (up to 300 N) and fast accelerations can be significant, especially for light-weight structures. Controlling these errors to 250 μm will require localized, fast actuators. Vibrations can be tracked using laser vibrometers with 1 kHz data acquisition. The research will use piezoelectric actuators to apply dynamic compensating forces up to 300 N and 500 Hz. The design of the mounting of multiple actuators will apply compensating forces in multi-degrees of freedom.