Joint project "Software-based real-time grid control"

The massive penetration of distributed renewable energy sources combined with the electrification of heating and transportation poses congestion problems in some parts of distribution networks and microgrids. Furthermore, the large variability of renewables, combined with the absence or inertia of inverters, threatens the overall stability of the national grid. The traditional methods of grid reinforcement and an increase in spinning reserves come at a very high cost and have the drawback of using fuel-based generators.

The goal was to provide a method for controlling power flows in a grid in real-time (sub-second timescale) in order to be able to follow the fluctuations of the fast changing resources, such as photovoltaic (PV) solar panels. The method should optimise the use of local resources, such as thermal loads (heat pumps), e-car charging stations, PV panels and storage units, while respecting all the security constraints (voltages, line ampacities, transformer ratings) and providing frequency support to the main grid. The method should be easily implementable and simple to maintain.

• A composable method for the real-time management of power flows using software agents was developed. The method was evaluated, patented and prototyped on a real-scale operating grid. It integrates heterogeneous electric resources in the distribution grid: PVs, batteries, fuel cells, heat pumps, Electric Vehicle (EV) charging stations while maintaining quality of service and providing support to the main grid.
• A method to manage energy consumption in a building and connected devices in order to provide energy storage services simultaneously at a broad range of time-scales was developed and demonstrated. A study recommending optimal designs for the Swiss building stock to maximise the resulting benefit to the grid has been published.
• A novel distributed energy storage solution, the Multiport Energy Gateway (MEG), was developed, which enables the distribution of electrical energy storage elements among several smaller units. Compared to large-scale bulk storage solutions this simplifies maintenance of the system and increases its flexibility, while simultaneously making it easier to recycle, upgrade and modify.

Implications for research

It is the first of its kind, as other agent-based methods are not able to operate in real-time. The method required the solution of a number of grid-modelling and technology issues, such as: the fast solution of state estimation and of inverse problems, the real-time identification of storage state, the ultra-short term prediction of loads and PV generation, the aggregation and translation of heterogeneous resources, the dynamic choice of slack resources. The results show how to integrate all resources in a device-independent way and obtain

• Quality of service in distribution grids and microgrids without grid reinforcement,
• Integration of a charging station with PVs and limited grid access,
• Participation of low voltage grids for voltage stability in medium voltage,
• Frequency support by local distribution grids,
• Dispatchable feeders,
• Islanding / de-islanding operation,
• Black start operation,
• Integration of real-time demand response.

Implications for practice

For the practice, in addition to the research implications mentioned above, the support of distribution networks and microgrids for higher frequency and voltage stability in the nationwide network is relevant.

Integration of Intermittent Widespread Energy Sources in Distribution Networks

The joint project consists of two research projects

Integration of Intermittent Widespread Energy Sources in Distribution Networks: Scalable and Reliable Real Time Control of Power Flows

• Prof. Jean-Yves Le Boudec, Laboratoire pour les communications informatiques et leurs applications, EPF Lausanne; Dr. Alexandre Oudalov, Prof. Joseph Sifakis, Prof. Mario Paolone

Integration of Intermittent Widespread Energy Sources in Distribution Networks: Storage and Demand Response

• Dr. Colin Jones, Laboratoire d'automatique 3, EPF Lausanne; Prof. François Marechal, Prof. Drazen Dujic