OBS! Ansökningsperioden för denna annonsen har
Information about the the project
Among available renewable energy sources, wind energy has probably the largest potential. The main challenge when operating the power system is to keep the system in balance, i.e. to keep the energy supplied in balance with electricity demand. On short term time scales (minutes to an hour) the challenges relate to power quality issues, such as stability of the voltage and frequency. On medium time scales (minutes to hours), the scheduled production must meet the planned demand and the power produced must balance the load.
The variability and unpredictability of wind power over short time scales remains a problem which limits its integration into the electric grid. One way to smooth the fluctuation is battery storage or power reserve capacity, but they are both expensive.
A better way to smooth wind production fluctuations is to forecast wind speed/power by foreseeing the stochastic nature of the wind. Furthermore, precise forecasting of wind speed/power forecasting can improve the efficiency of energy system and minimize the risks caused by system overloading and extreme weather conditions.
Unsurprisingly, forecasting wind speed is an active research area. A number of methodologies have recently been developed for forecasting which can be divided into three catogories: traditional mathematical statistics, numerical weather forecasting and machine learning. Regression analysis and time series analysis are some of the examples of traditional mathematical statistics methods.
Recently, efforts to improve forecasting methodologies have also included the use of Empirical Mode Decomposition (EMD) in many areas from solar and wind energy to financial time series. EMD divides data into its frequency components, which represent a number of high to low frequency components. The high frequency component corresponds to short-term changes, and low frequency component corresponds to long-term changes. By using different combination of the frequency components of the data we can predict both short and long-term predictions much more accurately compared to using the entire data set.
The Division of fluid dynamics carries out research in a wide range of applications such as aerospace (which is largest), automotive, windpower, hydropower, medicine, process industry, vehicle aerodynamics, shipping, and bio-mechanics. Almost every single research project is carried out in cooperation with a company. We do both numerical simulations and experiments. We develop our own in-house CFD codes for both incompressible and compressible flow, aero and vibro-acoustics as well as system analysis of gas turbines and we are very active in the OpenFOAM community. We have a fully equipped laboratory (Chalmers laboratory of fluids and thermal sciences) with three large windtunnels, a full set of PIV and IR imaging equipment. We do research on turbulent incompressible and compressible single and multiphase. The reseachers at the Division carry out both numerical and experimental research and develop new and improved turbulence and multiphase models and experimental techniques for both fundamental and real-world flows. We frequently build dedicated experimental facilities in collaboration with industry in order to better understand fluid flow details.
-The details of the departmental research activities may be found here.
Information about the position
In this post-doc project we will improve the accuracy of the short-term wind speed forecasting by using a hybrid EMD-SVR method (SVR=Support Vector Regression). In the post-doc project we have access to long-time measurements of the local wind conditions (including measurments from Chalmers own windpower station at Björkö). This will be used as training data for our machine-learning method. The research question is to develop an efficient hybrid EMD-SVR method which can be used for short-time (minutes to an hour) wind forecasting.
Research to be carried out by the succesful applicant
• Assemble raw data of wind conditions
• Identify which frequency components are important for difference forecasting scenarious (minutes to on hour)
• Develop a new machine learning method (most likeluy EMD-SVR) based on these data
• Analyze and identify weakness/strength of the method
• Write scientific journal paper(s)
Your major responsibility as postdoc is to perform your own research in a research group. If the succesful candidate is interested, the position may also include supervising master's and bachelor's thesis projects. Another important aspect involves collaboration within academia and with society at large. The position is meritorious for future research duties within academia as well as industry/the public sector.
A two-year postdoctoral position starting in early Autumn 2021. The successful candidate will be offered a full-time temporary employment of two year position.
To qualify for the position of postdoc, you must have a doctoral degree in a relevant field; in mechanical engng, Physics, Math. Engng, Aeronautics or any corresponding PhD. The degree should generally not be older than three years. You are expected to be somewhat accustomed to teaching, and to demonstrate good potential within research and education.
The position requires sound verbal and written communication skills in English. Swedish is not a requirement but Chalmers offers Swedish courses.
You are a creative person and have problem-solving ability. Good cooperation and social abilities are also appreciated.
Chalmers continuously strives to be an attractive employer. Equality and diversity are substantial foundations in all activities at Chalmers.
Our offer to you
Chalmers offers a cultivating and inspiring working environment in the dynamic city of Gothenburg.
Read more about working at Chalmers and our benefits for employees.
About the application procedure and to apply