| Description |
There is a large number of use cases for digital forecasting, monitoring, balancing, maintenance, and operation of energy systems in smart cities and smart industries.
A smart power distribution system manages the equipment, such as transformers, intelligently, based on monitoring and analyzing the system, consumption, and power quality data. One of the promising approaches is the use of digital twins. This solution involves the software modeling of different equipment, e.g., power generators, transformers and storage units. The model is a replica of the physical system, i.e., a network of power plants, transformers and other energy sources and distribution facilities. This software is the so called digital twin of the physical system. By running this computer model and using prediction algorithms use Machine Learning (ML), Deep Learning (DL), etc. the potential faults in the actual (physical) system can be detected and corrected before they occur. The objective of this project is to make a prototype of a AI-driven fault predicting power network.
Tasks: The students need to perform the following activities:
a. Studying of the predictive maintenance (fault prediction/correction) problem and performing thorough survey of different algorithms used for the problem, particularly, digital twins. The solution selected should be compared with the alternative (s) and its use be justified.
b. Selecting a set of technologies (algorithms) to be used and justify their choice.
c. Deciding on the hardware and software required so that the prototype while being simple be representative of a system with a reasonable practical dimension. The prototype should also be scalable. The route from prototyping to full scale implementation needs to be addressed. The output of this activity will be a brief report and a parts list with cost information.
d. A design consisting of the procedure for programming of node processor and the central unit hosting the prediction algorithms, e.g., digital twin, ML, DL, etc. The design should also have a test plan.
e. Implementation of the above system.
f. Test and validation and fine tuning.
|