# Recherche

# Sossan Fabrizio

### Professeur-e HES Associé-e

#### Compétences principales

### Professeur-e HES Associé-e

Bureau: ENP.23.N417

Rue de l'Industrie 23, 1950 Sion, CH

- Réseaux électriques
- Réseaux électriques avancés

2023

2022

2018

**Résumé:**

Multiple research works and power systems operational practices have qualitatively associated the progressive connection of stochastic renewable energy resources with the increase of power systems reserve requirements. At the same time, the price and technology of MW-class Battery Energy Storage Systems (BESSs) have considerably improved, which opens up the possibility to make electric distribution networks dispatchable. In this paper, we investigate the impact on the bulk power system of dispatchable electric distribution networks that host a large share of stochastic resources. The essential questions inspiring this research are: (1) Assuming that BESSs are deployed to achieve dispatchability of distribution grids embedding stochastic resources, what is the impact on the bulk power system reserve requirement? (2) Is this large-scale integration of BESSs economically viable compared to centralized reserve procurement from traditional power plants? To address these questions, we consider the case of the Danish transmission grid and the associated fleet of conventional power plants and compare it against locally dispatched distribution grids. We perform stochastic simulations to quantify and validate the amount of reserve necessary to operate these power systems with a desired reliability level. We establish a numerical equivalence between saved conventional reserve capacity and amount of BESS storage deployed in distribution networks. Then, we quantify the economic pay-back times of BESSs capital expenditure (CAPEX). The results show that: (1) large scale deployment of BESSs with dispatchable distribution networks is a viable technical solution to address flexibility requirements for the bulk power system and (2) this solution is economically viable with a pay-back time in the range of 11–14 years compared to providing flexibilities from conventional power plants.

2019

**Résumé:**

Distributed energy resources (DERs) installed in active distribution networks (ADNs) can be exploited to provide both active and reactive power reserves to the upper-layer grid (i.e., sub-transmission and transmission systems) at their connection point. This paper introduces a method to determine the capability area of an ADN for the provision of both active and reactive power reserves while considering the forecast errors of loads and stochastic generation, as well as the operational constraints of the grid and DERs. The method leverages a linearized load flow model and introduces a set of linear scenario-based robust optimization problems to estimate the reserve provision capability (RPC) area of the ADN. It is proved that, under certain assumptions, the RPC area is convex. The performance of the proposed method is tested on a modified version of the IEEE 33-bus distribution test system.

*Proceedings of the 13th IEEE PowerTech (POWERTECH) 2019*

**Résumé:**

The flexibility of distributed energy resources (DERs) accommodated in active distribution networks (ADNs) can be aggregated and then used to provide ancillary services to the transmission system. In this context, this paper presents a linear optimization method for the transmission system operator (TSO) to allocate its required active power reserve from aggregated resources installed in active distribution systems (ARADSs) as well as dispatchable bulk power plants (DBPPs). It consists in a linear optimization problem that minimizes the sum of the expected cost of active power reserve allocated from all possible providers (including ARADSs and DBPPs) and the expected cost of load not served over a desired time horizon. The value of lost load (VOLL) index is used as a criterion to realize an economical balance between the expected cost of allocated reserve and expected cost of load not served. The method leverages scenarios of power system contingencies and forecast errors of loads and renewable generation to represent typical operational uncertainties. A simulation proofof-concept using real-data from the transmission system operator of Switzerland, Swissgrid, is provided to illustrate the performance of the method.

2018

*Proceedings of the 2018 IEEE PES Innovative Smart Grid Technologies Conference Europe (ISGT-Europe)*

**Résumé:**

Uncertainty levels in forecasting of renewable generation and demand are known to affect the amount of reserve required to operate the power grid with a given level of reliability. In this paper, we quantify the effects on the system reserve and reliability, due to the local dispatch of stochastic demand and renewable generation. The analysis is performed considering the model of the IEEE 39-bus system, with detailed dynamic models of conventional generation, wind generation, demand and an under-frequency load shedding mechanism. The analysis compares to cases: the base case, where renewable generation and demand power are stochastic and the power reserve is provided by conventional generation, against the case where the operation of traditionally stochastic resources is dispatched according to pre-established dispatch plans thanks to controlling local flexibility. Simulations reproduce the post-contingency dynamic behavior of the grid due to outages of generators. The contingencies are selected to trigger under frequency load shedding mechanisms, hence to demonstrate the different levels of system operation reliability for the two case studies. Simulation results show that dispatching traditionally stochastic generation scores better regarding to expected energy not served, producing an increase of the system reliability.

**Résumé:**

This paper addresses the allocation of frequency control services (FCSs) from aggregated resources of active distribution systems (ARADSs) for balancing the transmission system while considering credible contingencies as well as forecast error of loads and generations through scenarios. First the paper introduces a generic framework for modeling ARADS including distributed energy resources (DERs) as seen from the transmission system operator (TSO) perspective. Afterwards, based on the proposed modeling framework and relying on a DC power flow model, the problem is formulated as a linear optimization problem consisting in minimizing the cost of FCSs provision and deployment from all possible providers including ARADSs and GENCOs. Nodal and total Expected Load Not Served (ELNS) indices are used to measure the security of the transmission system for the different scenarios. Finally, a proofof- concept of the proposed planning strategy is proposed by considering the IEEE 24-bus system.

*Proceedings of the 15th International Conference on the European Energy Market (EEM) 2018*

**Résumé:**

Transmission System Operators (TSOs) deploy frequency control reserves and regulating power to maintain the load-generation balance in real-time operation of power systems. In the Nordic countries, the TSOs buy regulating power from the Nord Pool regulating power market. In this paper, we developed a tool to quantify the price of regulating power as a function of both economic parameters such as spot (day-ahead) market price, and technical factors representing the current state of the system. First, the Nord Pool is considered as a single bidding area and an aggregated regulating power price is obtained, proving the validity of a simple non-linear algebraic model, when there is no influence of interconnections with neighboring areas. Then, we developed a case study for the West Denmark area, to demonstrate that for complex systems, where there is possibility of trade with other areas and there is high penetration of intermittent generation (e.g., wind power), this simple formulation is no longer valid. Finally, to solve this inconsistency, an improved model is here proposed by considering the effect of interconnections through two scenarios: one for unconstrained trade through the interconnections with neighbouring areas, and the second one where at least one of the interconnecting lines is congested. In addition, the wind penetration level is included as a parameter the non-linear algebraic model.

Réalisations