Last September, GE Digital, the GE entity dedicated to the development of software solutions and data analysis, launched in Europe and Israel, a competition open to data-processing start-ups in partnership with the Paris-based NUMA accelerator. Four of these 5 challenges were won by French start-ups.
The Eiffel Tower – Photo credit : Robin – CC
The evaluation and selection of startups was based on three criteria: the innovative nature of the solution, the compatibility of the solution with the architecture of the Predix platform and confirmation of its attractiveness for GE customers.
In order to separate the candidate start-ups, five challenges related to the analysis and exploitation of energy data were conceived:
Challenge 1: Forecasts on a local grid
The challenge was to improve production and power forecasting solutions at local levels and in real-time. The challenge is to improve forecasting capacities by combining wind, solar, meteorological and energy consumption data from customers, both private and business.
“Short-term forecasting is a critical process for grid operators, as they integrate a growing proportion of intermittent renewable resources. Today, at the national level, forecasting techniques can reach a level of confidence of 95%, while at the local level it can only reach 30%. The objective here is to make small-scale prediction as good as predictive performance on a larger scale.”
This challenge was won by Predictive Layer.
Challenge 2: Local and blockchain-based system of peer-to-peer electricity exchange
“The challenge is to explore the use of blockchains to enable consumer-producers that have invested in the production of renewable energy (such as solar roofs) to resell their excess capacity to their neighbours or to a neighborhood district, without any intermediary. To ensure the security of transactions despite the absence of centralization, a reliable and transparent history of transactions must be maintained “.
This challenge was won by Evolution Energie.
Challenge 3: Indoor Location Analytics
The challenge was to develop software that provides information on the use of a 5,000 m² space equipped with intelligent lighting.
“The objective of this challenge is to characterize the way in which this space is occupied (in order to optimize it) as well as user trajectories (in order to improve their experience). The location is equipped with 440 light sources, each capable of accommodating up to 4 different types of sensor. The frequency of the sensor data can be adjusted to the point where it reaches real time.”
This challenge was won by Irlynx.
Challenge 4: Optimizing Hydroelectric Production
“Develop an optimization algorithm for the planning of hydroelectric production through a network of dams and rivers. The challenge is to develop a solution that enables planners to design and update production plans, estimate profitability and long-term sustainability of operations, while allowing operators to make decisions Immediate.”
« Développer un algorithme d’optimisation pour la planification de la production hydroélectrique à travers un réseau de barrages et rivières. Le défi porte sur la mise au point d’une solution permettant aux planificateurs de concevoir et mettre à jour des plans de production, d’estimer la rentabilité et la durabilité des opérations sur le long terme, tout en permettant aux opérateurs de prendre des décisions immédiates ». This is a new development compared to the current situation: if planning is centralized, operational managers do not have the tools to assist them at the production site level.
This challenge was won by Cosling.
Challenge 5: Management of massive space-time series
The challenge is to design a data infrastructure and analysis software capable of collecting, storing, processing and visualizing the massive amount of data produced by the new generation of ultra-high resolution surveillance equipment. Distributors responsible for monitoring and controlling the power grid for stability and reliability have traditionally used SCADA information systems.
The challenge here is the capacity of the distributor to detect potential problems as quickly as possible and with the greatest possible relevance. The new generation of Wide Area Monitoring Systems (WAMS) provides knowledge of the state of the electrical network which is more precise than SCADA systems, with high resolution both in time and space. These WAMS rely on a large number of Phasor Measurement Units (PMUs). Geolocalised by GPS, these units make high frequency measurements of the characteristics of waveforms from electrical signals. The processing of the volumes of data thus created requires tools with unprecedented performance.
This challenge was won by Cityzen data.
ITEMS International for Think Smartgrids
Source : GE Digital