ARRS projects

Local-flexibility market platforms for distribution networks (DN-FLEX)

The EU long-term vision is foreseeing a climate neutral economy by 2050, which will require deep emissions reduction. As a consequence, also the powersystem is facing very big changes. On the supply side, the end target is the creation of a new energy scenario widely dominated by renewable energy sources (RES) and mostly based on distributed energy generation. On the demand side, at least two transformations are occurring. The first one is the electrification of transport. Electric mobility is expected to play a major role in the decarbonisation of the transport in Europe and worldwide in the next decades. The second one is the electrification of heating. As heating and cooling in the built environment accounts for almost 40 % of the total final energy demand in Europe, also the electrification of heat is one of the main pathways to decarbonisation. Therefore, the transmission and distribution networks are facing the challenges of large shares of renewable generation, electric vehicles(EVs)and heat pumps (HPs), all affecting network operation to a large extent.

The present project aims at addressing the above challenges with the development of local-flexibility market platforms for distribution networks. The proposed system will enable the provision of network services with distribution-network operational constraints in mind. On the DSO side, there are three calculation modules, which together form the distribution-network constraints platform. On the aggregator’s side, the flexibility aggregation platform includes the flexibility scheduling module that aggregates and activates the flexibility potential of active customers in an optimal way, while taking into account distribution network constraints. On the market side, the market operator platform contains the services scheduling module for the provision of services to different power system stakeholders based on the service demand and flexibility supply offer.T hese platforms will be tested in a pilot project within an actual LV network.

The authors acknowledge the project (Local-flexibility market platforms for distribution networks (DN-FLEX), L2-3162) was financially supported by the Slovenian Research Agency. Project will last from 1.10.2021 – 30.9.2024.

Composition of the research group

Faculty of Electrical Engineering (UL FE):
Prof. dr. Boštjan Blažič (SICRIS), leader of the project, Janja Dolenc (SICRIS) in Marjan Ilkovski (SICRIS).

GEN-I:
dr. Rok Lacko (SICRIS),  Luka Nagode, Jan Tršinar in Matej Malenšek

Elektro Gorenjska (EG):
Anže Vilman (SICRIS) in Nejc Petrovič (SICRIS)

Project phases

There will be 6 phases (work packages) in this project.

Work Package 1: Development of the distribution network constraints platform
Main objective of this work package is to develop and test forecasting algorithms suitable for forecasts of consumption of highly variable LV consumers. Development of a state-forecast algorithm, capable of robust and accurate state-estimation ofday-ahead operation will also be done. We will aim to develop the distribution network node capacity module for calculation of maximum nodal generation and consumption, which will not exceed network operational limits. All algorithms will be tested with simulations.

Work Package 2: Development of the market operator platform
Upgrade of the existing flexibility aggregation platform will be performed, so that it will takeinto account distribution network constraints.  Market operator platform, aimed at the provision of services for the DSO, will be  developed. Proposed algorithms will be tested with simulations.

Work Package 3: Development of the advanced network planning tool
We will aim to statistically model RES generation and load consumption, with the focus on PV, EVs and HPs. We will propose a methodology for a network planning approach. Network planning tool enabling the analysis of highly variable distribution network operation and the simulation of smart grids solutions (including flexibility provision) will be developed.

Work Package 4: Demonstration of the service platformsand results evaluation
The main objective is to demonstrate the operation of the developed services platforms in the field, i.e., in a LV network, where active customers will participate in network control by demand response. To select the LV network for the demo, install and connect metering equipment, and analyse the LV network will be done. The use cases for services at the DSO and TSO levels will be defined. Demonstration will provide feedback on the algorithm operation and generate practical results.

Work Package 5: Scalability and replicability study, regulatory recommendations
Based on the demo results, a study will be carried out to assess the impact of a wide scale (national level) implementation of the local flexibility market. Technical, social and regulatory barriers for local flexibility market will be identified, and changes will be proposed to enable the use of a services market atnational distribution level.

Work Package 6: Communication, dissemination and exploitation of the results
Communication with stakeholders, including network customers and the Energy Agency will be ensured. A subcontractor will be hired for professional communication with network customers. Project results will be disseminated in the form of newsletters, conference publications and journal papers. Exploitation plan for the project results will be built.

Other sofinancers

Deep reinforcement learning for optimisation of LV distribution network operation with integrated flexibility in real-time (DRIFT)

High variability of distribution network operating conditions due to stochastic generation from RES is already a cause for concern. The circumstances are expected to worsen with a rising share of new loads, mainly heat-pumps (HPs) and electric vehicles (EVs), as these are loads linked to users’ habits and perception of comfort, which is still poorly analysed. The increasing generation and increasing consumption, not necessarily at the same time, are leading to overloading of network elements (usually transformers) and the deviation of voltage profiles from the required levels (high voltage levels due to generation and low voltages due to consumption).

On the positive side of the network impact of the green transition, renewables, new loads (EVs, HPs) and storage represent also a source of flexibility, which can provide value to the power system both through generation assets management and the flexibility in loads operation. RES can usually use their reactive power or active power curtailment, loads can provide a certain degree of shifting in their consumption without significantly impact the user’s comfort and storage is the most versatile asset in terms of flexibility. The increasing volume of flexibility provided by active users is an important feature of the future power system. Flexibility services provide indeed clear benefits for distribution system operators (DSOs) in terms of voltage control and congestion management, maximizing RES, EVs and HPs integration. Moreover, flexibility can be offered as a service also to other power system stakeholders such as balance responsible parties (BRPs) for portfolio balancing and by transmission system operators (TSOs) as reserve provision. The active role of the network users is clearly expressed also in the policy package “Clean Energy for all Europeans” (proposed by the EC on 30 November 2016), stating that the process of power system transformation brings a completely new role for the electricity consumers: from peripheral users to key players of the power system.

The authors acknowledge the project (Deep reinforcement learning for optimisation of LV distribution network operation with integrated flexibility in real-time (DRIFT, L2-4436)) was financially supported by the Slovenian Research Agency. Project will last from 1.10.2022 – 30.9.2025.

Composition of the research group

Faculty of Electrical Engineering (UL FE):
Prof. dr. Boštjan Blažič (SICRIS), leader of the project, and Janja Dolenc (SICRIS)

Faculty of Computer and Information Science (UL FRI):
dr. Jure Žabkar (SICRIS),

Elektro Gorenjska (EG):
Nejc Petrovič (SICRIS) in Blaž Dobravec (SICRIS)

Project phases

There will be 6 phases (work packages) in this project.

Work Package 1: Project starting point definition
The main objective of this work package is to review the state-of-the-art in the field of AI applications for power network operations and define the use cases along with KPIs.

Work Package 2: Development of the DRL-based network control algorithm
The methodology that uses deep reinforcement learning (DRL) to build grid-operating agents for real-time network control will be developed. Data for the learning process will be generated. Proposed algorithms will be tested and validated through simulations.

Work Package 3: Upgrade of the Distribution network simulation tool
A digital twin of the demo site, i.e., secondary substation TP Srakovlje, will be built. Along with the digital twin, reference simulation models of the Slovenian LV network will be selected and generated. The main objective will be to upgrade the Distribution network simulation tool, which will enable the testing of the DRL network control algorithm.

Work Package 4: Field implementation and testing of the developed network control algorithms
Network control algorithm operation with real-time controller-in-the-loop simulations will be performed. Demonstration of the operation of the developed network control algorithm in the field, i.e., in an LV network, where consumers will participate in network control by demand response will be performed.

Work Package 5: Scalability and replicability study
Based on the demo results, a study will be carried out to assess the impact of a wide-scale implementation of the DRL-based network control algorithm.

Work Package 6: Communication, dissemination and exploitation of the results
Communication with stakeholders, including network customers and the Energy Agency will be ensured. Project results will be disseminated in the form of newsletters, conference publications and journal papers. Exploitation plan for the project results will be built.

Other sofinancers