Thesis worker: 30 hp - Evaluation of deep learning methods for powertrain sound source characterizat

Scania Nederland B.V.; Södertälje kommun; On-site; SEK 100 000 - 400 000; Full time

Job summary

A leading automotive company in Sweden is offering a thesis position focused on evaluating deep learning methods for powertrain sound source characterization. This role involves building models based on simulations and measurements, requiring a Master’s degree in relevant fields. Candidates should have experience in machine learning and Python. Applications should include a CV, personal letter, and transcript of grades.

Qualifications

• Experience with ML and Python packages is preferred.
• Knowledge of acoustics and signal processing is a bonus.

Responsibilities

• Explore deep learning approaches for sound source characterization.
• Build surrogate models trained on simulation and measurement data.
• Evaluate monopole distributions for accurate modeling.

Skills

Machine Learning

Python

Acoustics

Signal Processing

Education

Master’s degree in Mechanical/Civil/Aerospace engineering, Technical Physics, Applied Mathematics, Computer Science

Job description

This position is within one of TRATON’s companies.

Thesis worker: 30 hp - Evaluation of deep learning methods for powertrain sound source characterizat

Introduction

Acoustic characterization of powertrains is essential to design better vehicles for driver comfort, fulfilment of legal requirements and overall customer perception.

Background

In order to build reliable acoustic simulation models of the complete vehicle, powertrain noise sources must be accurately represented. At the same time, the representation has to be sufficiently simple so as to enable the simulation of multiple vehicle configurations in a short time. One representation that potentially satisfies these contrary requirements uses monopole clusters as equivalent sources. The source data used to generate these monopole clusters can be obtained either using simulations or far-field measurements in a component test rig using a sparse microphone array.

Objective

The goal of the thesis is to explore the potential and advantages of deep learning approaches using Convolutional Neural Networks (CNN) – as an alternative to classical signal processing and matrix inversion based approaches – to understand optimal monopole distributions for accurate sound source characterization.

The student is expected to build surrogate models trained on both simulation and measurement data. These measurements are performed in a semi-anechoic powertrain test-bench. Some important features of the model that will be explored in detail are the resolution of the velocity/pressure distribution in the near-field of the test object, frequency range of validity, neural network architecture and training/validation dataset generation. Experience with ML and Python packages is preferred and knowledge of acoustics and signal processing is a bonus.

Education/program/focus

Education: Master’s degree in any of the following – Mechanical/Civil/Aerospace engineering, Technical Physics, Applied Mathematics, Computer Science

Start date for the thesis work: January 2026

Estimated time required: 20-25 weeks

Contact persons and supervisors

Dayasagar Srinivasan, Senior Engineer, [e-mail]

Application

Your application must include a CV, personal letter and transcript of grades

Background check

A background check might be conducted for this position. We are conducting interviews continuously and may close the recruitment earlier than the date specified.