Artificial intelligence based real-time power system stability assessment (AI–ASSIST)

The development of modern society has been based on the use of electricity for decades, so the consequences of a large-scale blackout would be catastrophic. Ensuring the stable electric power system (EPS) operation is therefore necessary. This already formidable technical challenge has been exacerbated in recent years by the rapidly growing number of alternative converter-interfaced sources. Therefore, if we want to reduce the risk of a blackout and at the same time support the implementation of larger amounts of renewable and distributed sources of electricity, there is a need to develop new technological solutions and tools. The project is based on the existing state of technology and available infrastructure in EPS. By using both, we believe that it is possible to achieve the main goal of the project, namely to conceptualize, develop and implement a unique solution for real time EPS dynamic stability assessment (DSA) with the help of artificial intelligence technology. Three key steps are cruical: definition of an innovative concept for establishing cause-and-effect relation between the initial EPS operating state (described by features) and its stability (described by stability indices), the use of artificial intelligence technology for the creation and management of a central database of manageable dimensions containing information on the EPS stability and development of a methodology for a real-time EPS stability assessment, based on fast screening for similar states in the central database. In this way, we avoid the need for performing time-consuming dynamic simulations in real time (as is the practice with existing DSA tools). In Slovenia, transmission or distribution system operators do not yet use such tools, so they have no insight into possible EPS dynamic instability. The project therefore represents a unique activity both in Slovenia and internationally. It will play a key role in managing all the conditions to which the Slovenian EPS will be exposed in the future due to all the changes we are witnessing. The basis for the start of the project will be the past recordings of Slovenian EPS operating conditions, provided by the industrial partner of the project ELES (national transmission system operator in Slovenia). Based on this, we will create and calibrate a static model of the Slovenian EPS, the results of which will be considered as a kind of fingerprint of the operating state (so-called features). Furthermore, a dynamic model of the Slovenian EPS, suitable for the analysis of several types of dynamic stability, will also be developed. The result of each of the dynamic simulations will be evaluated on the basis of characteristic stability indices, for which we will establish an unambiguous cause-and-effect relation with the initial operating conditions. The first role of artificial intelligence is the integration of features and stability indices for an extremely large number of diverse EPS operating conditions into an optimized central database. Here, the emphasis is on the quality and sufficient size of the database and not on the speed of the algorithms. As soon as the database is available, another role of the artificial intelligence appears, which is a subject to real-time operation. We are referring to the fast recognition and identification of sufficient similarity between the current operating state of the actual EPS (ELES measurements) and those in the central database. This design effectively combines the advantages of accurate yet time-consuming dynamic simulations and fast (within milliseconds) recognition of features to which dynamic analyses refer. Project results will be key in supporting ELES in the coming years. The project envisages the implementation of the DSA tool in the Diagnostic and Analytical Centre of ELES, where the operation of the procedure will be thoroughly tested and validated by both project partners (ULFE and IJS) and engineers from ELES.