IoT solutions for existing processes: a laser based additive manufacturing (LBAM) use case
01/07/21
< Project overview >
Introduction
This project was to develop a business case for IoT implementation by developing, cataloguing and demonstrating how to integrate IoT sensors and control systems into existing manufacturing processes.
The project uses a laser based additive manufacturing (LBAM) process as a use case for the business case. The work was to be performed by the University of Sheffield, through the Automatic Control and Systems Engineering (ACSE) department and the supporting manufacture using Advanced Powder Processes (MAPP) hub which provides access to LBAM facilities and domain knowledge. Industrial stakeholders include Rolls-Royce.
Project aims
The aim was to develop and document a business case and a supporting demonstrator for an IoT solution on an existing manufacturing process using LBAM as a use case. LBAM is a production process that involves the use of a laser to melt and fuse together powder material to create layers of coating on a substrate. The choice of LBAM as a use case is driven by its potential as a future manufacturing technology and the impact that IoT could have on the LBAM process.
The targeted barriers were (6) business case, (3) educational, and (7) data issues. The challenge lies in evaluating the use of IoT grade sensors and data analytics solutions to aid in process monitoring of LBAM which is lacking in terms of: understanding the methodology to integrate this technology; best practice in the management of data collection; and awareness of the associated costs/knowledge requirements.
The principal beneficiaries are the stakeholders in the LBAM technology (both academic, OEM manufacturers and users, ie SMEs.
What was done?
The research undertaken can be broken down into three main work packages. Firstly investigating how we can monitor the mass flow rate of the LBAM process, secondly thermal characterisation of the melt pool, and thirdly a focus on developing knowledge exchange material and a business case for IoT integration within LBAM.
The actions undertaken were in collaboration with our internal partners, with the University ACSE developing the necessary IoT solutions and supporting software, and the University MAPP Hub providing access and domain knowledge of LBAM.
Results
The backbone of this project is complete, with the development of four hardware platforms for process monitoring (thermal camera platform, hall sensor platform, weight sensor platform, scraper RPM platform), and supporting software (HDF5 [Hierarchical Data Format v.5] data pipeline, thermal characterisation models, historical digital twin of LBAM process).
The results and outcomes of the development, integration and experimentation process has been documented into training material and supporting GitHub repositories. This allows us to see the efficacy of IoT-grade solutions to challenges within LBAM, and their use in process monitoring, along with the associated costs and knowledge requirements to achieve such integration.
Deliverables and other tangible outputs
Publication: a conference paper for the upcoming 54th CIRP Conference on Manufacturing Systems, 2021 has been completed.
Training material: a knowledge exchange report which documents the project aims, objectives and experimental outcomes along with supporting analysis has been made available. Additional presentations and GitHub repositories also provide additional training resources.
Dataset: all data collected (thermal characterisation, mass powder flow, imaging of LBAM processes) has been made available within the project GitHub repository.
Video presentation/demonstration: a number of presentations done over the course of the project have been made available within the project GitHub repository.
Hardware demonstrator: over the course of the project four hardware solutions/demonstrators were developed (IoT thermal camera, hall sensor platform, weight sensor platform, and scraper RPM platform). Each of these platforms was experimentally verified within the LBAM environment and data is available within the project GitHub repository.
Software product/source code: all software demonstrators and source code that support the running and analysis of the hardware demonstrators has been made available within the project repository. This includes thermal characterisation modelling, powder flow rate process monitoring and developments in a historical digital twin for melt pool size and 3D surface reconstruction.
Impact
This project has ensured employment of researchers on this project in the ACSE department and provided valuable experience, knowledge and new skills around integration of IoT solutions into the LBAM process.
The thermal technology and skills developed throughout this project have since been applied in another thermal camera project involving the company Beatson Clark. The project involves using an IoT thermal camera platform to monitor the heat distribution in a glass forge.
Next steps
Engagement between academic researchers and local businesses (Beatson Clark/RR) will hopefully lead to future projects and contribute to the growth of local businesses. There is an intent to publish the project results in an appropriate conference such as the MAPP Conference which has been postponed twice from its original date in 2020, and is tentatively expected to take place later in 2021. There is also an intent to develop one or two journal papers depending on further experimentation.
A workshop is organised to share the IoT experience with Tinsley Bridge. Tinsley Bridge is interested in the digital inspection and measurement methods enabled by IoT technologies.
Lessons learned
The main benefit was the development of strong links with new internal partners (ACSE/MAPP Hub) during this project, and the experiences and skills learnt to achieve the main outcomes. In particular the chance to develop and apply novel IoT solutions to an emerging manufacturing process, ie LBAM, and gain a better understanding of the business costs and challenges.
In hindsight, the challenges of Covid-19 and subsequent restrictions to lab access were not ideal, and even though we were able to shift towards more data analysis and development of machine learning using the data captured early in the project, we could have tried to mitigate this disruption earlier. For example, we had hoped to get more experimental runs and perform specific CT scans of the LBAM material outcomes to align the developed thermal characterisation models with the characteristics of the material.
One of the challenges was in trying to best integrate the develop IoT solutions with the OEM LBAM system. A better engagement with the manufacturer of the LBAM system would have been beneficial to allow faster understanding of the system itself, how and what data is stored locally, and how best to interface with it.
Access to deliverables, resources and media content
Project code, a detailed business case and implementation report and other associated documents can be found in the the Github repository.
In addition, a Kaggle repository hosts the data collected during experimentation
What has Pitch-In done for you?
Pitch-In has been a valuable vehicle for the development of new collaborations and the generation of novel research as it allows for interested parties in the form of industry, academia and consumers to work closely together to generate ideas and develop solutions to challenging problems in their fields.
In our project, the challenge of gathering real-time process data and performing analysis through the use of machine learning to control and understand additive manufacturing processes is of great importance to this emerging industry. This project brought forth many benefits, firstly through the development, integration and testing of novel IoT sensing and process characterisation hardware, along with the supporting software, algorithms and tools for real-time analysis, and importantly the knowledge exchange between all partners involved.
Overall, we feel that through the funding of Pitch-In and its support we as a team have been able to evaluate the role IoT sensing and machine learning can play in the field of additive manufacturing, whilst at the same time it could serve to open up this area of IoT-based process control and analysis to new manufacturers by demonstrating its effectiveness for additive manufacturing.
Project lead
Professor Ashutosh Tiwari, Airbus/RAEng Research Chair in Digitisation for Manufacturing – the University of Sheffield
Partners
The University of Sheffield, including: