EDF – P2P Energy market place

Blockchain based marketplace to favor local P2P energy trading

1. Project summary

Name of company: EDF

Sector: Energy 

Project Period: 2018 – April 2018

Mission of the project: 

The mission of this project is to enable direct energy trading between energy consumers (energy producers/prosumers and end-consumers) based on a local private blockchain. 

Final Result: 

  • A local Blockchain based energy market place has been designed, with a smart contract ruling the trading through an auction mechanism. 
  • An agent-based framework has been developed to simulate local energy market place with realistic consumption/production behavior. This allows to define the technical and economic conditions needed to deploy this solution (i.e. by evaluating the outcomes of the market place in different scenarios characterized by sets of parameters) 
  • The solution has been successfully deployed and tested on IoT devices and is currently being experimented in Ile-de-France (EDF Lab Les Renardières) 

Xdev technologies: Multi-Agent Simulation / ChainDeployer

BC technologies: Ethereum 

Maturity Level: Production 

2. Project details

Summary of the project

The mission of this project is to enable direct energy trading between energy consumers (energy producers/prosumers and end-consumers). First, we have deployed a private blockchain to enable dynamic trading of local energy without trusted third party, with a smart contract setting and enforcing prices with an auction mechanism. The solution has then been tested in a virtual environment, thanks to our proprietary agent-based framework, allowing to simulate multi-agent interactions with realistic consumption / production behaviour. The experimentations have been pushed further, with the construction of a physical model incorporating IoT devices. The success of this Proof of Concept phase has lead to the current deployment of the solution in production phase on a EDF site in Ile-de-France. 


Households equipped with local production means (i.e. prosumers) such as solar panels or wind turbines consume their self-produced energy and sell the remaining energy (i.e. the surplus). Until 2017, French citizens were forced to sell this surplus to EDF, the main energy provider in the country, at a rate fixed by the government. Today, the legislation has evolved to let prosumers sell their surplus to anyone in the vicinity1 at any rate. Thus, the redistribution and pricing of renewable energy produced and exchanged among residences within a community has become a major concern. A classic approach to implement a market place is the use of a centralized system managed by a trust entity. While this kind of architecture is suited for financial trading applications, the cost of trust could be an obstacle for local energy markets. In this context, an increasing number of researchers and energy experts foresee the blockchain as key technology to enable new cost effective local market places. 

However, few have implemented their solutions and moreover, technical feasibility, scalability and economic impact based on realistic data has never been assessed. Indeed, there is no framework available to experiment use cases with multiple, distributed participants interacting with a blockchain. 

Pain Points

The current local energy exchange accountancy system suffers from the following main weaknesses (c.f. Figure 1):

  • Households have to designate a legal representative
  • The billing is centralized: households have to trust the legal representative
  • Payments are done manually with a risk delay and conflicts
  • Lack of openness to other services
Figure 1


Our solution enables dynamic trading of local energy without trusted third party with a simulation tool to evaluate the solution (c.f. Figure 2). The main components of our contribution are:

  • A local Blockchain based energy market place with tokens: households participating to the market place
    support themselves a local Blockchain to propose asks and bids of energy volumes through the use of
    an auctioneer Smart Contract that allocates the offers in a transparent and secured way. Trades are
    accounted using a custom Energy Token, that could be used in other future services.
  • A Blockchain deployment tool: in order to the assess the performance of the infrastructure, we propose a tool to deploy automatically an Ethermint Blockchain through a graphical user interface.
  • Market place simulator: in order to test the market place, we propose a tool that simulates the behavior of the households consisting in producing/consuming energy and making offers to the local market place. Parameters of the simulator can be configured by the user through a graphical user interface. The consumption and production data are provided by EDF.
  • Deployment on IoT devices: the proposed solution will be deployed and tested on hardware with limited computational resources (i.e. Raspberry Pie 3)
Figure 2


In the course of developing the aforementioned system, several challenges have been tackled :

  • Allocation algorithm: while Smart Contract programming languages are limited, we have implemented a complex algorithm for the resource allocation: a strategy proof double auction mechanism2 that will prevent speculative behaviors.
  • Energy consumption of the system: given the context, one requirement of the system is to limit its energy consumption of the marketplace infrastructure to 1% of the total production of the locality. We have tackled this constraint by using a low energy profile Blockchain technology.
  • Assess technical and economic performances: in order to stress-test the Blockchain infrastructure, we have put in place a benchmark tool assessing the transactions throughput of the solution. The economic performance has been studied using the simulation tool through a sensibility analysis. The simulations gave insight on the economic viability of the solution in different geographic locations and locality scales (from 3 to 200 hundred households): the economical gain per household could be assessed in Lille (city in the north of France) and Marseille (south) during winter and summer.


This project led to the publication of a scientific paper in an international conference:

  • Brousmiche, K-L., O. Dib, A. Anaoica, T. Abedllatif.
  • “Blockchain Energy Market Place Evaluation: an Agent-Based Approach” IEEE IEMCON, Information Technology, Electronics and Mobile Communication, Vancouver, 2018.

(1) For now, reselling is limited to the neighbors connected to the same local energy distribution point. However this restriction is likely
to be lifted in the near future.

(2) Pu Huang, Alan Scheller-Wolf, and Katia Sycara, “A Strategy-Proof Multiunit Double Auction Mechanism,” in Proceedings of the First International Joint Conference on Autonomous Agents and Multiagent Systems: Part 1 (ACM, 2002), 166–167.