Hi! I’m Lawrence, an environmental data scientist for Quanterra Systems.
I use a combination of Python, R, MATLAB, and LaTeX in my work, which covers data analysis/visualisation, signal processing, machine learning, finite element modelling, and technical writing. I also have an interest in working with geospatial data.
I hold a PhD in Electronic and Electrical Engineering, MSc in “Applied Acoustics”, the Institute of Acoustics “Diploma in Acoustics and Noise Control”, and a BSA (British Society of Audiology) Certificate in aural impression-taking. Please check out some of my publications if you are interested in my academic work.
Many of the older posts on this blog come from my time working for the now defunct Solent Acoustics, a subsidiary of Solent University, Southampton, so you may see a few links back to the use of their equipment, and the collaborators on some of the posts that also worked at the university.
I graduated from Southampton Solent University in 2013 with a BSc in “Live and Studio Sound”. I then joined the university as a Graduate Associate in “Broadcast and Cinema Sound”, and subsequently became a Research Assistant in Acoustics as part of Solent Acoustics.
As part of that role I undertook consultancy tasks such as the noise and vibration measurement of car parts, medical devices, and headsets. I also delivered practical sessions to both undergraduate, masters, and short-course students in the areas of acoustics and electro-acoustics as an Associate Lecturer.
I left Solent University in 2019 to start a PhD at the University of Southampton, which focussed on the development of a novel temperature monitoring system for aero-engine nozzle guide vanes (NGVs), using ultrasonic guided waves. I successfully completed the PhD in 2022. In my last year I was awarded the EPSRC Doctoral Prize, which allowed me to continue at the university as a research fellow, carrying out a feasibility study into the application of an ultrasonic guided wave based temperature monitoring system for printed circuit boards (PCBs). I then secured a position as a “New Frontiers Fellow”, in which I continued my PCB related research experimentally.
In mid-2024 I left my role at the university and joined Quanterra Systems as an Environmental Data Scientist, where I am working to improve their data processing pipeline for eddy covariance measurement.
I have a growing interest in cartography, which stems from a fascination for visiting new places, as well as my love for GeoGuessr! I’m eager to learn more, but for now you can check out a few of my blog posts on using R/QGIS for geospatial stuff, as well as the range of sources that got me started.
Saurabh
19 April 2022 at 12:15 pm
Hi Yule
Excited by your work , i have a query on how to plot two wave files in one window. I hope you can help me.
Lawrence Yule
19 April 2022 at 11:24 pm
I’m sure I can help, what software are you using?
Allen
17 July 2023 at 9:23 am
Hi Yule, I learned a lot from the 2D-FFT code you shared and thanks a lot!
I am a new beginner in COMSOL, I have made a simple model of an acoustic wave time-domain study and a frequency-domain study and simulated them using the same sources and same boundary conditions, I want to match the results of the time-domain study using the FFT with the frequency-domain study, however, I still can’t do it even though I’ve been spending almost a month on it, can you help me with this please?
Lawrence Yule
23 August 2023 at 10:48 am
Hi Allen, apologies for the slow reply! I am happy to help, perhaps you could share your models with me and I’ll take a look?
ali kadkhodaei
12 September 2024 at 2:20 pm
hi yule
I have a question
if in your example we change dimension of plate to 610mm*455mm the model simulation takes a lot of time about 5 hours at least. what’s your suggestion to reduce time of calculation ?
thanks a lot
Lawrence Yule
12 September 2024 at 4:16 pm
Hi Ali, there are a few things that you can tweak, but ultimately you will have to accept that these simulations can take a long time. The excitation frequency has the strongest link to the sim time, so try to reduce it if possible. You can see this from the mesh calculation (velocity[m/s]/N/f_0[Hz]), where N should be around 5, but you can try reducing this to 4. The time step is also linked to the excitation frequency (CFL/(N*f_max)), try changing f_max to f_0, or reducing f_max. Also if there is any symmetry in your geometry you can try the techniques described here: https://www.comsol.com/support/learning-center/article/Using-Symmetry-to-Reduce-Model-Size-35921 I hope that helps!