Stage en Rte-l2s-supergrid

SuperGrid Institute regroupe 180 employés de 28 nationalités différentes dans un environnement dynamique situé à Lyon. En tant que centre de recherche et d'innovation indépendant, nous sommes spécialisés dans le développement des technologies pour le futur réseau de transport d'énergie, le « supergrid », y compris les technologies HVDC et MVDC.

Centre de recherche multidisciplinaire avec des capacités de simulation avancées et de multiples plateformes de test, dont de nombreux laboratoires associés, SuperGrid Institute utilise son expertise pour fournir un large éventail de services et de solutions pour aider ses clients à développer des systèmes électriques, leurs équipements et leurs composants.

**Introduction**:
Lightning is one of the most frequent sources of faults for power system components, in particular for overhead transmission lines (OHL) [1]. Lightning strikes can be classified depending on the stroke component, and the consequence of the strike. A lightning striking the shield wire, or a transmission tower is referred to as an indirect lightning [2]. Depending on the magnitude of the lightning current and other parameters such as the grounding of the tower, it may cause the insulator to flash, which creates a fault, or a disturbance that propagates along the line. A lightning may also bypass the shield wires and may strike a conductor, which is referred to as a direct strike [1]. Such strikes may also cause a fault or only a disturbance. Apart from lightning strikes, faults (or disturbances) may also be due to vegetation or pollution.

Whatever the event causing a fault, protection algorithms [3] must be able to detect and identify the affected line, to isolate the faulty component (and only this one) with a minimum delay. Events causing only a disturbance should not lead to the triggering of protection of affected parts of the grid. Moreover, after the identification and clearing of the fault, the accurate location of the fault is important for line patrol as well as diagnosis and monitoring of the line [4].

**Objectives**:
The aim of this internship is to **develop a single-ended approach to estimate the lightning strike locations in HVAC grids** using travelling waves, without assuming any prior knowledge of the lightning current at the strike location. One of the main challenges associated with single-ended based methods is the ability to detect and interpret correctly the multiple reflected and transmitted waves that travel throughout the grid. This is especially true for complex waveforms resulting from lightning strikes [4].

The proposed approach will be model-driven and data-driven. It will rely on a travelling wave model of the propagation of voltage and current waves throughout the grid after an event (model-driven part of the approach) and will use parts of the current and voltage observations performed at the relay to infer the shape of the voltage and current at the event location (data-driven part of the approach).

Since the proposed algorithm will be implemented within a protection relay monitoring a transmission line. It will have to account for constraints related to the acquisition frequency and to the bandwidth of currently available and future current and voltage measurement devices.

**Main research direction**:
This internship will largely capitalize on previous results obtained by the hosting team on ultra-low latency fault location in HVDC grids [5-7,9]. The work will thus start with a bibliographical overview of recently developed single-ended approaches for fault location and identification [5-8].

Then, simulations of the impact of strikes on HVAC grids will have to be performed to analyze current and voltage waveforms available at various locations of the grid. An ElectroMagnetic Transient (EMT) software will be employed to accurately represent the voltage and current waveform after a lightning strike. The impact of the limited acquisition frequency and bandwidth of measurement devices will have to be taken into account. The approach proposed in [9] will be adapted to HVAC grids.

The ability of the approach to locate faults affecting a part of the line not directly connected to the station acquiring measurements will be evaluated. The impact of the acquisition frequency and of the bandwidth of measurement devices on the fault location accuracy will be analyzed.

**Expected background