Robert Gordon University
Fully-funded PhD Studentship
Deep Learning-based Wind Speed and Solar Irradiance Forecasting to Enhance Green Energy Generation
Closing Date – Midnight, Sunday, 6 June 2021
Applications are sought for a fully-funded) Research Studentship (PhD) to carry out research at Robert Gordon University, Aberdeen, United Kingdom, under the supervision of Professor Li Zhang.
Duration and Funding
The duration of project will be up to 36 months, commencing in October 2021. The studentship covers both tuition fees (UK or International level) and a tax-free stipend (living allowance) of £15,000 per annum.)
The successful candidate will be required to relocate to Aberdeen as soon as possible to study, although studies may start remotely initially.
The proposed research is aimed at undertaking semantic cloud segmentation and remote sensing and surveillance to inform wind speed and solar irradiance forecasting, as well as unlock the subtle variations from air, ocean and soil signals to identify long-term impact of our social activities and carbon footprint on the nature.
The world’s energy future has been shifting from traditional leading sources of electric power generation such as coal, petroleum, natural gas, and other fossil fuels, to renewable energy sources such as wind and solar power, to mitigate the impact of global warming and climate change. According to Global Energy Review 2020, renewable energy has contributed nearly half of the Britain’s electricity generation and nearly 28% global electricity generation in the first quarter of 2020. Offshore wind farms and solar panels are the biggest contributors to the aforementioned green energy generation. However, owing to the intermittent nature of such renewable energy sources, accurate prediction of generation capacity is essential to maintain system reliability and maximize renewable energy integration. Moreover, existing studies revealed that wind speed and solar irradiance are crucial factors pertaining to precise wind and solar power forecasting.
Therefore, in this research, we leverage the newly developed AI methods such as deep neural networks (DNNs) for undertaking short-term wind speed and near-real-time solar irradiance prediction to enhance energy generation forecasting. Because of the direct impact of cloud position and coverage on the surface irradiance (e.g. Direct Normal Irradiance (DNI) and Diffuse Horizontal Irradiance (DHI)), which is the vital factor for wind speed and solar irradiance forecasting, real-time cloud coverage information from multi-spectral satellite imagery will be extracted. Since the current modern geostationary weather satellite samples a total of 16 bands every 10 mins over the full disk of the earth, this empowers near-real-time analysis of cloud characteristics (e.g. the percentage of cloud cover as well as cloud and aerosol particles). Novel evolving DNN models and quantum-inspired or graph convolutional networks will be explored for cloud instance and semantic segmentation. Furthermore, the aforementioned cloud detection method may also help with the identification of other climate
parameters, e.g. hurricanes and storms, to aid energy generation forecasting. Finally, the wind speed and solar irradiance prediction will be conducted by incorporating the cloud assessment output with other weather variables.
The proposed project based on wide-spectrum high-resolution satellite images will also undertake remote sensing and surveillance to unlock the subtle variations from air, ocean and soil signals, to identify fundamental and long-term impact of our social activities and traditional and renewable energy generation processes on the nature, e.g. global warming, marine ecosystems, deforestation and microplastic pollution at large scales such as cities, oceans or beyond. The resulting AI-based models will be extensively evaluated using multi-sensor optical remote sensing and energy generation data sets. The project will foster wider multi-disciplinary national, international, academic and industrial research collaborations.
Training and Skills
The PhD candidate will work with a multi-disciplinary supervision team with world-class researchers in the field and receive specific research training on machine learning, deep learning, remote sensing, evolutionary computation, subsea engineering, computer vision and image processing techniques.
Applicants should have a very good BSc (Honours) (First or Upper Second class) degree or a Master degree (with Distinction or Merit) in Computing Science or related discipline.
Essential Knowledge and Experience
- IELTS (or equivalent) at 6.5 overall and 6.5 in each component/element is required for Research based degrees – the successful candidate will be required to pass the IELTS or equivalent English qualification before starting their studies. For more information, please visit https://www.rgu.ac.uk/study/international-students/english-language-requirements
- Strong academic background in computer science and/or engineering disciplines.
- Strong programming skills in Python, MATLAB, Java, C++ and other programming languages
- Expertise in machine learning and data analytics
- Experience of developing decision support systems
- Initial knowledge and expertise in image classification, computer vision and/or signal processing
- Good academic writing and communication skills
- Expertise in deep learning and/or computer vision
- Experience of using/developing a variety of deep neural networks for image segmentation and generation tasks
- Experience of remote sensing and time-series forecasting using deep networks and/or hybrid models
- Successful completion of industrial projects
- Strong academic writing and communication skills
- Research publications in top journals and conferences
Applicants should have good personal and communication skills, strong professionalism and integrity, and be capable of working on their own initiative.
Enquiries should be submitted through https://www.findaphd.com/phds/project/deep-learning-based-wind-speed-and-solar-irradiance-forecasting-to-enhance-green-energy-generation/?p130922 or be emailed direct to Kate Lines at firstname.lastname@example.org
Applications should be submitted through https://www.findaphd.com/phds/project/deep-learning-based-wind-speed-and-solar-irradiance-forecasting-to-enhance-green-energy-generation/?p130922 or be emailed direct to Kate Lines at email@example.com by midnight, Sunday, 6 June.
The application should consist of:
- A covering letter or personal statement of interest
- IELTS (or equivalent) certificate
- Two references (at least one academic or professional)
Further information such as passport details or transcripts may be requested during the short-listing stage. Interviews (which may include a short practical test) are expected to take place week commencing 5 July 2021.