Energy harvesting IoT systems for predictive maintenance in marine diesel engines

14/06/20

< Project overview >

Introduction

This project investigated options for Royston Power Generation Ltd (Royston) to develop a predictive maintenance service for their UK and international marine diesel engine customers.

The project originated when Newcastle University suggested to Royston that Pitch-In might be an effective vehicle for investigating the adoption of IoT to monitor remotely, the performance of marine diesel engines as a means of reducing both operational costs and carbon emissions. This led to discussions on the development of a sensor system that was powered through energy harvesting. A specific area of innovation that Royston has expressed interest in is applying vibration sensors and harvesters to enable battery-free operation.

Project aims

This Pitch-In project focused on using IoT in conjunction with vibration energy harvesting to develop a battery-less wireless predictive maintenance system as part of Royston’s product and service development activities. While Royston were aware of the benefits of IoT, they were not confident in committing to this type of project without technical support from Pitch-In. Additionally, there was also a corresponding financial risk at the implementation stage.

As an engineering SME, Royston gives priority to ‘safer’, less risky projects with a clearer pathway to deployment when allocated funding from their RD&D budgets. This Pitch-In project facilitated the de-risking of the introduction of IoT solutions into Royston’s product and service development for a predictive maintenance system.

What was done?

An experienced senior research fellow was employed on this project. This helped significantly to achieve overall good outcomes despite the project’s short duration and the limitations imposed by the COVID-19 pandemic. The latter significantly limited the interaction with Royston, such as accessing their facilities to train models or test the vibration harvesters. Despite this, we were able to hold meetings with the company, which helped us to progress the project.

The project conducted focused on three main activities (which will continue after the end of the project) which involved exploration of the following areas:

  1. Energy harvesting, where different vibration harvesters were tested.

  2. Energy management solutions.

  3. Different techniques for data analysis.

Results

For energy harvesting we tested several different vibration harvesters starting with piezoelectric transducers which appeared to be good candidates due to the small factor, ease of installation , and most importantly, large power output (tens of mW of power). However, after testing a particular model, S128-H5FR-1107YB, we realised it was unsuitable for use with a diesel engine because it was more responsive to bending rather than to pure vibrations. Therefore, we decided to test some magnetic/kinetic energy harvesters which preliminary analysis indicated were more suitable.

For the energy management unit we tested different commercially available solutions such as the ultra-low-power boost charger TI BQ25505. These were tested because of their ability to convert the input voltage to a level where the energy can be stored in a storage device and their ability to track maximum power point tracking. For example, for the TI BQ25505, a boost converter adjusts its input impedance and in this way, the power source always operates at its optimal point to maximize the harvested energy.

For data analysis we are still exploring different options for data analysis ranging from signal processing techniques, such as time-domain analysis, frequency-domain analysis, time-frequency domain analysis, to more sophisticated techniques, such as by using classification to predict whether there is a possibility of failure in next n number of observations, or regression to predict how much time is left before the subsequent failure could happen. The outcomes from such analysis could form a key element of conditioning monitoring and predictive maintenance regimes for diesel engines deployed in remote, mission-critical marine environments.

Due to the short duration of this project, all the activities above are still ongoing; our aim is to consolidate our outcomes and publish within the next year.

Impact

As stated previously, this was a short project which nevertheless laid the foundations for further investigations. In particular, it provided a framework for further engagement with Royston which came about directly because of Pitch-In. This also led to support for Royston from the ERDF-funded Arrow SME regional support programme and the award of a PhD studentship from the ERDF-funded Industrial Innovation Programme.

Next steps

It was intended that this Pitch-In project would have led to further collaboration with Royston via an ERDF-funded Intensive Industrial Innovation Programme PhD Studentship and/or a further larger and detailed grant proposal to EPSRC.

Unfortunately due to COVID-19, Royston have been unable to progress the studentship (where the student would have been up to 50% company-based). We have however submitted a £60k proposal (supported by Royston) for a feasibility study under the Industrial IoT theme of the EPSRC Connected Everything (CEII) fund.

We also plan to produce two publications based on this project (ideally one conference and one journal publication).

Lessons learned

I believe that knowledge sharing between my group and Royston was the main achievement of this project and in this sense both parties benefited from our collaboration.

Additionally, being a new lecturer, this project helped me understand better the needs of local SMEs such as Royston where collaboration with universities can contribute significantly to consolidating and strengthening their position in highly competitive international markets. Finally, the project helped me to create a foundation for my future research work and industrial collaborative projects.

Without COVID-related issues, we would have had a more active collaboration with Royston and a more immediate and deeper follow-up collaboration. I would have probably asked for a longer project because the initial plans were very ambitious for a six-month project. I would have probably waited for a less-complicated period to start a collaboration and for a longer time.

I believe it would have been better to have two research associates rather than one on this project; one focused on the architectural aspects and the other on the data analysis. Having only one person for six months didn’t enable me to achieve all the ambitious outcomes I had planned when the project was conceived.

Project lead

Dr Domenico Balsamo – Newcastle University

Partners

Royston Power Generation Ltd

< Theme >