Thesis worker - Will My Vehicle Be Ready?

Scania; Stjärna; Presencial; SEK 167.000 - 257.000; Tempo parcial

Resumo da vaga

Scania seeks a Master's student to work on Predictive Vehicle Readiness and Smart Charging for Electric Heavy Vehicles. The role focuses on predicting departure readiness from charging data and energy needs, building probabilistic forecasts, and testing strategies against historical charging sessions.

You will extend the model toward bidirectional charging and evaluate its value in depot peak shaving and electricity price optimization, delivering a prototype view, final report, and presentation.

Qualificações

• Master's student in Data Science, Applied Mathematics, Statistics, Computer Science, Electrical Engineering or related field.
• Strong Python & SQL skills for data analysis and solid knowledge of statistics and ML.
• Fluent English is required.

Responsabilidades

• Review literature on departure and energy prediction, probabilistic forecasting and smart charging.
• Build probabilistic models for departure time and mission energy and combine with charging progress.
• Define how a not-ready outcome is identified and evaluate models.
• Design charging/preconditioning strategy and test by replaying historical charging sessions.
• Extend strategy to bidirectional power transfer and estimate energy/power while maintaining readiness.
• Simulate value of flexibility for scenarios like depot peak shaving or price optimization.

Conhecimentos

Python

SQL

Statistics

Machine Learning

English

Formação

Master's student in Data Science / related field

Descrição da vaga

This position is within one of TRATON's companies.

Thesis worker - Will My Vehicle Be Ready?

Thesis Worker 30 hp Charging and Power Distribution

Will My Vehicle Be Ready?

Predictive Vehicle Readiness and Smart Charging for Electric Heavy Vehicles

Introduction

Thesis work is an excellent way to get closer to Scania and build relationships for the future. Many of today's employees began their Scania career with their degree project.

Background

An electric truck or bus has to be ready when its next mission starts. Smart charging and preconditioning can lower electricity cost and battery ageing, but only if the vehicle knows when it will leave and how much energy the mission needs. Today, the driver or fleet operator has to enter this manually, and in practice it is often missing or inaccurate.

Connected vehicles log every charging session and every drive. Each departure shows when the vehicle actually left and how much energy the mission used, which gives a large, naturally labelled dataset for learning these patterns.

Bidirectional power transfer, defined in ISO 15118-20, could also let a parked vehicle send energy back to the depot or the grid, for example to cut depot peak load or to take part in electricity and balancing markets. Heavy vehicles have large batteries but energy sent back may have to be recharged before departure, and the extra cycling adds battery wear. Knowing whether a vehicle will be ready is therefore also the key to knowing how much flexibility it can safely offer.

Objective

Develop and evaluate a method that estimates, while a vehicle is charging, the probability that it will be ready for its next mission, and uses that estimate to decide when to charge and precondition, and how much energy the vehicle can safely offer through bidirectional power transfer.

• How accurately can departure time and next-mission energy be predicted from a vehicle's own history, compared with simple baselines such as same time as last week?
• Can the predictions be combined into a readiness probability that can be trusted, so that vehicles given 90% are ready about 90% of the time?
• Which charging and preconditioning strategy best balances readiness, electricity cost and battery ageing, compared with charging immediately at plug-in?
• How much energy and power could a vehicle offer through bidirectional power transfer without putting its next mission at risk, and what is that flexibility worth after battery-wear cost?

As thesis worker, you will:

• Review research on departure and energy prediction, probabilistic forecasting and smart charging.
• Build probabilistic models for departure time and mission energy, and combine them with charging progress into a readiness probability.
• Define how a not ready outcome is identified in the data, and use it to evaluate the models.
• Design a charging and preconditioning strategy and test it by replaying historical charging sessions against baselines.
• Extend the strategy to bidirectional power transfer, estimating how much energy and power each vehicle can offer while keeping its readiness probability above a set level
• Estimate the value of this flexibility by simulation on historical sessions for 1 or 2 use cases, such as depot peak shaving or electricity price optimization, including battery degradation cost
• Deliver a prototype view, a final report and a presentation.

Education/program/focus

• Master's student in Data Science, Applied Mathematics, Statistics, Computer Science, Electrical Engineering, Vehicle Engineering or a related field.
• Required: good Python & SQL skills for data analysis, solid knowledge of statistics and machine learning, and fluent English.
• Meritorious: Time-series or probabilistic modelling (e.g., survival analysis), and knowledge of electric vehicles, batteries, power system or charging.

Number of students: 1

Start date for the thesis work: January 2027

Estimated time required: 20 weeks

Contact persons and supervisors:

Aravind Marimuthu, Charging Function Owner, [e-mail]

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