2022 Strategic - Analysis and optimisation of mix design and properties of low-CO2 concrete using advanced deep learning technologies

Status: Open

Applications open: 19/07/2021
Applications close: 13/08/2021

View printable version [.pdf]
About this scholarship
Description/Applicant information

Over the last few decades, cement has been used throughout the world as a traditional and reliable binder in the construction sector, which is responsible for 5 - 8% of greenhouse gases emission in the world. The production of cement requires intensive energy consumption and a huge volume of virgin raw materials. As a result, the scientific community encourages alternate sources for developing sustainable concrete by using supplementary cementitious material to protect the environment and reduce disposal costs. In this regard, calcium-rich materials, i.e. fly ash, rice husk ash, bottom ash and metakaolin, were activated by using the alkaline solution to produce alkali-activated concrete as an alternate binder to ordinary Portland cement (OPC). These calcium-rich materials are wastes from other industries. Therefore, if they can be used to produce concrete, an enormous amount of wastes can be turned into valuable building materials. Previous studies in the literature have shown that concrete made of these wastes yielded promising results. However, the results are very scattered due to the influences of many factors, such as mixing procedure, molecular of activator, type of activator, replacement level to OPC, activator to binder ratio, chemical composition of binders and interaction of these factors. As a result, there are quite a few experimental investigations on this topic but the findings are not consistent. Especially, the optimal design is barely consistent and it is a case-to-case dependence. Therefore, using the experimental method to examine the optimal mix design and the mechanical properties of concrete is not straight-forward and inconsistent due to many influential factors. Fortunately, advanced computer-based methods, i.e. deep-learning, are good candidates to solve such complex problems.  Therefore, this project intends to develop advanced deep learning technology to address these difficulties in order to find a rigorous solution for an optimal mix design and mechanical properties. Deep neural networks have the capacity of representing any complex functions and thus provide a good framework to model the relationship between the strength (and other mechanical properties) of concrete and the mixture proportions of different types of supplementary cementitious materials. Moreover, deep neural networks are compact representations and have shown superior generalization performances compared to traditional machine learning methods (e.g., support vector machines) in many applications such as computer vision, machine translation, natural language processing and structure damage prediction in civil engineering. The primal aims of this project include
• Develop a database from previous experimental/numerical studies in the literature.
• Develop advanced deep learning technologies to learn a consistent model which can predict the strength (and other mechanical properties) of concretes based on the collected data. 
• Conduct factor analysis using interpretable neural network techniques to analyse the key factors for optimal mixture design.
The developed techniques have great potentials to save cost for the design of new concrete material, in terms of its optimal mix and material properties. They can also be extended for the design of many other types of construction materials. 

Student type

• Future Students

Faculty

• Faculty of Science & Engineering
  • Engineering courses

Course type

• Higher Degree by Research

Citizenship

• Australian Citizen
• Australian Permanent Resident
• New Zealand Citizen
• Permanent Humanitarian Visa
• International Student

Scholarship base

• Merit Based

Value

Total value of the annual scholarships (stipend and fees) is approx. $60,000 - $70,000 p.a. Curtin PhD Stipends are valued at $28,597 p.a. for up to a maximum of 3.5 years.

Successful applicants will receive a 100% Fee offset.

Scholarship Details
Maximum number awarded

1

Eligible courses

All applicable HDR courses

Eligibility criteria

We are looking for a graduate in either Computer Science or Electrical Engineering with strong mathematical background and good programming skills. Experience with deep learning is desirable. 

Application process

If this project excites you, and your research skills and experience are a good fit for this specific project, you should complete the Expression of Interest (EOI) form now. 

You will need to ensure you accurately select the Project lead (listed below) as your nominated supervisor and provide details for at least one referee.

Expression of Interest (EOI) form

Enrolment Requirements

Eligible to enrol in a Higher Degree by Research Course at Curtin University by March 2022

Enquiries

To enquire about this project opportunity contact the Project lead (listed below).

Name: Professor Ling Li 

Email: [email protected]

Contact Number: 9266 7939

Scholarships Email Alert
Sign up now