Mudit Dubey (md598@cam.ac.uk)

Using a social network of industrial assets to enable smarter decision-making in factories

Supervisors: Dr Bhupesh Kumar Lad and Dr Ajith Parlikad

AVTEC is a large independent manufacturer of powertrains and precision-engineered products in India. It is part of the CK Birla Group, a large conglomerate with subsidiaries in the technology and automotive, home and building and the healthcare and education sectors.

In this project AVTEC in co-operation with IIT Indore are looking at optimising job sequencing, batch sizing and maintenance scheduling at AVTEC’s Pithampur site, which is characterised by multi-stage processing of several sub-products followed by the final assembly in a flow shop environment. The project will look to identify the information communication required and developing cyber-twins, or simulators that contain the constraints and characteristics of its real-life counterpart. The information communication systems will be established within this cyber-twin network, which will contain the machines, as well as the production department (responsible for making scheduling decisions) and the maintenance department. This method uses a distributed operations planning approach as each part of the network acts according to its individual constraints first before a system-level approach is taken, rather than a centralised approach which would be significantly more time-consuming as the solution space is very large.

Mateusz Pniewski (msp39@cam.ac.uk)

Social platform for machines

Supervisor: Dr Ajith Parlikad

The concept of Social Internet of Industrial Things (SIoIT) has been first used about a decade ago by Julian Bleecker and his works on Near-Field Interaction and the Internet of Things. The concept is based on social networks’ ability to efficiently reach users with information relevant to them and is hence an interesting framework for the problem of connecting elements of the IoT network.

This project will be primarily concerned with development of an architecture of a social platform that will enable the aforementioned communication between machines. The architecture will draw from social platforms commonly used in real life and use their key features most suitable for efficient machine communication. One of the key challenges that will be addressed by the project is developing an efficient way of storing information fed into the platform by machines for other machines to access it quickly (in real time) and easily to make suitable adjustments. To prevent machines from browsing through all the historic data an intelligent way of filtering out information will need to be employed. Another part of the project will aim to develop a template of a ‘profile’ for any machine to be able to join the social network, be compatible with other machines and be able to understand information fed onto the platform instantly. The highlight of the development of the architecture will be implementing it in a demonstration in the Robot Lab. A scenario of a situation relevant to many manufacturing processes will be selected to best demonstrate practical implementation and power of the developed social platform.

Will Harborne (wh279@cam.ac.uk)

Modelling and optimisation of the grinding circuit in a large mining operation

Supervisor: Alan Thorne

Intellisense.io is a Cambridge based start-up. It develops Internet of Things applications in partnership with large mining companies to optimise performance and efficiency. Mining is a particularly asset intensive industry, and the unpredictable commodity prices and increasing operations cost mean that there is a significant drive for improvements.

The project will focus on optimising grinding in a SAG mill in Chile. The SAG mill is a bottleneck process and is only stopped once every 4-6 months for maintenance. If the rotation speed of the mill is too low the ore is not broken up efficiently. This causes a build up in the machine and further efficiency and capacity reductions. If the rotation speed is too high the ore can cause heavy wear on the machine and cause it to be stopped more often for expensive maintenance. Optimising the operating parameters of this bottleneck process therefore has the potential to be very valuable in terms of energy savings, capacity increases, and maintenance cost savings.

Finding the optimal operating parameters will require several data sets from the mine, including: material flows, motor data, pressure sensors, and acoustic sensors listening to the noise in the spinning mill. Using this data a statistical model will be created for how the SAG mill works and is affected by different parameters. This will then be used to optimise the operation of this particular mill. The project outcome will be to integrate the model with Intellisense.io software, and create a demo video. The model should be such that it can be customised to fit other mine sites with different sized SAG mills.