Learnings

Outcomes

1.       New PMU trilateration method that reduced the number of required PMUs in oscillation source localisation by more than 70% in the considered cases.

2.       New energy method that calculated the mode of interest and its arrival and departure times in the wavelet domain to enhance the accuracy of the conventional energy method from 10% to 67% in the considered cases.

3.       New wavelet method that used the wavelet energy to enhance the accuracy of the conventional energy method from 10% to 95% in the considered cases. 

4.       Five deliverable reports and associated MATLAB programming codes detailing the designs and software implementation of these new methods.

5.       Three publications to increase the impact of the innovation.

Lessons Learnt

Managerial:

1.       It was useful to engage with the stakeholder as soon and as frequently as possible to keep the direction of innovation on the right track, for example, to make the simulation settings close to reality.

2.       It would be useful to have the university research admin support and other relevant staff in the project meetings to understand their admin procedures better for reduced delays.

Technical:

1.       The impedance-based PMU trilateration method requires power system topology information but in some situations this may be difficult to acquire, as the topology may dynamically change. For future projects, this method could be improved by using machine learning algorithms, such as reinforcement learning, to adapt the method to different environments.

2.       Some of the data from the ISO New England test case library may have issues, such as Case 5. For future projects, data pre-processing may be adopted to check the quality of data before applying them to the oscillation source localisation methods.