Master Thesis: Hierarchical multi objective/ multi step ahead forecasting.

Currently in Siemens Energy, almost all the machine learning use cases are based on time-series data from various sensors on turbines across the world. There are hundreds of models running for sales operation planning, forecasting power demand, unit performance, efficiency and so on to increase the business value of our customers. Multiple models are maintained for different horizon forecasts. One of the major steps forward in this offering is the reduction in number of models and hence the cost of maintaining these models. Further improvement in forecast accuracy of the existing time series estimators by embedding more information (temporal/non-temporal hierarchies) from the sensors into the models is also in the focus.

Your role in this project is to conduct an empirical research to build and benchmark hierarchical time series forecasting models that can predict multi step ahead. Your job will also focus on investigating about the business impact of your new proposed solution, and how it is useful for the company in comparison with the current forecasting models.

 

LET’S TALK ABOUT YOU

 

Your profile
* Academic background in energy systems, engineering, computer science, statistics, machine learning, mathematics, or another relevant field.
* Curiosity about data analytics (especially in time series analysis) and you must have programming skills (preferable: Python, R).
* Knowledge of machine learning/deep learning/time series-oriented libraries (scikit-learn, caret, theif, hts, mlr, keras, tensorflow, pytorch etc…), data handling libraries (Pandas or tidyverse) is meritorious.
* We consider meritorious a GitHub record with univariate, multivariate time series forecasting projects and deep learning.

 

Your responsibilities
* Literature Review of existing methodologies used for embedding hierarchical information in the time series forecasts and time series/deep learning models for multiple horizon predictions.
* Perform a comparative study on the most used time series forecasting models and investigate their relative performance using real sensor data from gas turbines.
* Compare hierarchical forecasting approaches for different base forecasting models and reconciliation methods.
* Benchmark the accuracy of Hierarchical Time Series (HTS) models and deep learning models for multi forecast horizons against existing traditional baselines (prophet)
* Understand and present the business impact of your solution in comparison to company’s current forecasting models- in terms of performance, resources, cost, etc.

 

Your opportunities for personal growth:
* A chance to put your specialist knowledge to practical use and take an application-oriented approach to work.
* A conducive workplace and supervision to learn and understand the practical challenges of large-scale machine learning, time-series forecasting and come up with innovative ways to tackle them.
* A chance to create an impact in current processes in Siemens Energy by means of actively contributing to research on improvement of existing forecasting models.

 

LET’S TALK ABOUT US

"Let’s make tomorrow different today" is our genuine commitment at Siemens Energy to all customers and employees on the way to a sustainable future.

Siemens Energy is one of the world’s leading energy technology companies. The company works with its customers and partners on energy systems for the future, thus supporting the transition to a more sustainable world. With its portfolio of products, solutions and services, Siemens Energy covers almost the entire energy value chain – from power generation and transmission to storage. The portfolio includes conventional and renewable energy technology, such as gas and steam turbines, hybrid power plants operated with hydrogen, and power generators and transformers. More than 50 percent of the portfolio has already been decarbonized.

A majority stake in the listed company Siemens Gamesa Renewable Energy (SGRE) makes Siemens Energy a global market leader for renewable energies. An estimated one-sixth of the electricity generated worldwide is based on technologies from Siemens Energy. Siemens Energy employs 91,000 people worldwide in more than 90 countries and generated revenue of around €29 billion in fiscal year 2019. In Sweden Siemens Energy has 2600 employees in 10 locations.

Read more about Siemens Energy here:

www.siemens-energy.com

 

MORE INSIGHTS

Be Energized. Be you.

Lucky for us, we are not all the same. Through diversity we generate power. We run on inclusion and compassion. Our combined creative energy is fueled by at least 130 nationalities. Siemens Energy celebrates character – no matter what ethnic background, gender, age, religion, identity, or disability. We energize society. All of society. 

 

Job & Career: https://www.siemens-energy.com/global/en/company/jobs.html

 

Application

Don’t hesitate – apply via https://jobs.siemens-energy.com/jobs id nr 227964 not later than 2021-12-05.

For questions about the role, please contact Namita Sharma on namita.sharma@siemens-energy.com.

For questions about the recruitment process please contact the responsible recruiter Nelly Johansson on nelly.johansson@siemens-energy.com.

Location: Finspång

 

Trade Union Representatives:
Veronica Andersson, Unionen, 0122-840 21
Simon Von Eckardstein, Sveriges Ingenjörer, 0122-842 24
Jan Lundgren, Ledarna, 0122-812 33
Jonny Persson, IF Metall, 0122-817 69