Intelligent Maintenance of Transmission Grid Assets
Contact: Dr. Blazhe Gjorgiev and Dr. Laya Das
In cooperation with the Swiss transmission system operator Swissgrid, we aim to develop tools for fault detections, diagnostics, and prognostics of power grid assets.
This project aims at efficiently address two challenges: (i) determine the maintenance windows of grid assets through predictive maintenance methods; (ii) optimally schedule the maintenance of grid assets with respect to grid operations. The maintenance plan is the determination of the period at which an asset needs to be maintained. The maintenance window is based on the expected time to failure, which is obtained by a predictive maintenance algorithm. The maintenance scheduling is defined as the determination of exact dates for the maintenance of all assets across the grid. Maintenance scheduling is overlooking the maintenance of all assets, considering the maintenances window of each individual component (determined by the maintenance planning process), the cost of maintenance, and the operations of the grid.
The proposed research tackles two main research goals:
- Developing data-driven predictive maintenance algorithms at component level enabling early fault detection with a sufficiently long foresight period, i.e. a maintenance window. It is expected that a common framework is set up and implemented for power lines and transformers of components and/or sub-systems.
- Developing a maintenance scheduling support system at grid level that (a) integrates updated information on the system condition from individual components; (b) achieves a decision on the optimal point in time for maintenance and the optimal maintenance resource allocation; and (c) evaluates the impact on the risk at network level and the total maintenance costs.
The partnership with Swissgrid ensures relevance to current industrial practices for maintenance planning and scheduling as well as the implementation of the developed tools.
