City, University of London Featured PhD Programmes
Engineering and Physical Sciences Research Council Featured PhD Programmes
University of Sheffield Featured PhD Programmes
FindA University Ltd Featured PhD Programmes
FindA University Ltd Featured PhD Programmes

Data-driven micro-seismic signal analysis


Project Description

The goal of this PhD is to develop a method for optimal acoustic and microseismic sensor placement for accurately detecting microseismic activity. This project will use an alternative, data-driven, approach which will draw on significant recent developments in the area of machine learning.

Optimal sensor placement, i.e., determining the locations of sensors that maximise information about the monitored dynamic system, is an active research area with numerous applications. The goal of this PhD is to develop a method for optimal acoustic and microseismic sensor placement for accurately detecting microseismic activity. In contrast to commonly used traditional optimisation approaches this project will use an alternative, data-driven, approach which will draw on significant recent developments in the area of machine learning - deep learning and graph signal processing - to develop supervised or semi-supervised regression algorithms. The project will make use of already available microseismic monitoring data from well stimulation projects in North America. The input data with corresponding cost outputs will be used to train a model which will predict a solution for the test data. The main advantage of this approach is avoiding analytical optimisation methods that could be either too complex or inaccurate due to assumptions required to be introduced in order to make the solution computationally tractable.

This PhD will be based at the University of Strathclyde. Optimal sensor placement is an emerging area of research that, to the best of the supervisors’ knowledge, has not been studied in the context of microseismic monitoring. It is unclear whether past methods developed for structural monitoring are transferrable to the problem in hand. These past methods are mainly based on complex optimisation approaches or proper orthogonal decomposition, which both require model simplifications to ensure computational tractability. Some recent attempts to use data-driven methods rely on random forests algorithms and their adoption to the problem in hand is unclear. The novelty of the work is: (1) novel deep learning network architectures for optimal sensor placement; (2) new combined network for iteratively performing sensor placement and event detection; (3) real-time implementation of the developed algorithms.

Funding Notes

Standard eligibility for UKRI studentships.

View Website

Home fee, standard EPSRC stipend plus £5k per year project costs

Non-home students will have to provide the excess fees and may not be eligible for a stipend (see UKRI studentships page)

How good is research at University of Strathclyde in Civil and Construction Engineering?

FTE Category A staff submitted: 20.20

Research output data provided by the Research Excellence Framework (REF)

Click here to see the results for all UK universities

Email Now

Insert previous message below for editing? 
You haven’t included a message. Providing a specific message means universities will take your enquiry more seriously and helps them provide the information you need.
Why not add a message here
* required field
Send a copy to me for my own records.

Your enquiry has been emailed successfully





FindAPhD. Copyright 2005-2019
All rights reserved.