Research Projects

Cyber-physical interoperability of machines and subsystems 

This work aims to establish a framework for selecting and integrating drives and other subsystems to minimize conflicts, promote standardization and adaptability in machinery design, and reinforces supply chain resilience. Jointly led with the University of Manchester, the project benefits from their extensive expertise in drive systems, which are crucial for machine accuracy and performance. 

The framework will enable real-time optimization of machines based on criteria such as energy efficiency, output quality, and fault mitigation. This will be achieved by processing condition data using novel tools for data fidelity assessment and feeding it into self-learning, cognitive systems for performance optimization. The resulting cyber-physical framework will include APIs for core functions like system optimization, data sharing, and error compensation. These APIs will be validated using standardized methods and tested on partner machinery, enhancing the performance of next-generation manufacturing systems. 

Find out more about the interoperability of machines and subsystems.

Advanced error correction and compensation 

This work aims to improve the accuracy of precision machinery such as CNC machines tools using software-based error compensation. Most machines are complex mechatronic systems with many components and therefore they have positioning errors caused by geometric imperfections inherent in manufacturing and assembly of such complex systems. In addition, there are several heat sources on a machine, including bearings and motors, that cause thermal errors. Compensating for these errors requires a range of mathematical models and performance data therefore two key areas research are the creation of accurate models, some physics based and some using machine learning. The compensation requires high quality data about the current state of the machine in real-time during operation therefore part of the research includes the development and integration of novel strain sensors that can detect the thermal distortion in the machine and this data is useful for the thermal models. Working closely with other work on cyber-physical interoperability, the compensation software can be more easily connected to a variety of different machines. 

Find out more about Advanced error correction and compensation.

Design and Optimisation through Virtual Tools

This research theme focuses on enhancing the mechanical design process for machine tools through the integration of advanced virtual Computer-Aided Engineering (CAE) tools. CAE has become a standard across many engineering disciplines, with digital twins of physical systems increasingly adopted in mainstream design workflows. 

This work package comprises two primary research strands: 

Defeaturing of CAD Models for Downstream CAE Applications 

The first strand investigates the transformation of detailed, CAD-centric models into simplified representations suitable for downstream CAE operations such as Finite Element Analysis (FEA), Computational Fluid Dynamics (CFD), and digital twinning. While CAD models are typically comprehensive, capturing full bills of materials and complex geometries necessary for manufacturing, this level of detail can hinder efficient CAE processing. The defeaturing process, which reduces model fidelity to suit specific CAE requirements, is often time-consuming and may introduce uncertainty into simulation results. This research aims to streamline the defeaturing workflow, enabling rapid model simplification tailored to the needs of subsequent CAE tasks, while maintaining acceptable levels of predictive accuracy and uncertainty. 

Incorporating Machine Tool Uncertainty into Virtual Machining

The second strand addresses the limitations of conventional Computer-Aided Manufacturing (CAM) systems, which typically assume nominal machine performance. This research introduces real-world machine tool uncertainties into the virtual machining process, allowing CAM simulations to reflect the actual performance characteristics of the machine. By embedding these uncertainties into the virtual part generation, operators can assess whether a given machine is capable of producing components to the required quality standards, prior to any physical machining. This approach enhances predictive capability and supports more informed decision-making in manufacturing planning.