PREDICTIVE FAULT DETECTION FOR MACHINE CHARGING IN CONSTRUCTION ENVIRO

Background of thesis project
Volvo Construction Equipment, a division of Volvo Group, is one of the world's largest manufacturers of construction equipment such as wheel loaders, dumpers, excavators, road machinery, and compact machines. The production is distributed in Europe, Asia, North America and South America and Volvo CE employs approximately 15000 people around the globe.
To maintain a leading position, it is important to develop innovative products with high quality, efficiency, reliability and uptime


Suitable background
This MSc thesis is suitable for one or two students that are completing their studies in mechatronics, applied physics, control engineering, or similar.


Experience with mathematical modelling and interest in automotive applications is valuable. Basic knowledge in machine learning is considered a merit. Your own drive, analytic ability and curiosity are important factors. We would like to see you document your work and it is encouraged to take your own initiatives in this thesis work.
The thesis is preferably performed by two master students.

Description of thesis work
Uptime is extremely important in the construction and quarrying industry, and any unexpected breaks in the production cycle must be avoided to reduce cost for a site owner. However, most of the construction equipment’s operating cycle is repetitive, and a fleet of machines are used to perform a certain task. The machine operators involved will not be changed often, and autonomous machines even drive the same exact path. Therefore, the energy usage, efficiency and other machine parameters will be quite similar across a fleet of machines.


One of the possible points of failures for the future fleet of electric Volvo machines is the charging of such machines. For example, during use the charging inlet will get worn out, it is operated in a very harsh and dirty environment, and at some point, it will cease to work and cause a problem during operation. Ideally, we would like to know when a charging inlet is about to fail, and then replace it just before that happens. This is known as predictive maintenance.


This thesis aims to analyze the data generated in a construction environment with its repetitive tasks, and to develop a method for predicting failures ahead of time to reduce the number of unplanned stops in the operation of a site. The requirements on the data are also within the scope of this thesis (which parameters to track, data resolution, data quality).


Thesis Level: Master
Language: English
Starting date: 2023-01-15
Number of students: 2
Tutor Reinier Overmaat, Electromobility Engineer, +46 769409917


Kindly note that due to GDPR, we will not accept applications via mail. Please use our career site.