Goal
Primeo Netz AG would like to reliably forecast how much solar energy will be produced when and where it will be fed into the grid as a planning basis for grid expansion.
Solution
Development of a reliable forecasting model based on dynamic data and information sources - from the measuring points to the potential study.
Result
Creation of a planning basis for controlling grid expansion and calculating the load profile and voltage tolerance.
Challenge
With over 600 employees, Primeo Energie reliably supplies customers in Switzerland and France with energy. Primeo Netz AG serves the Grids business area - from the transportation and distribution of electricity to the provision of grid-related services and e-mobility.
Primeo Netz AG is strategically positioning itself in the field of data science in order to develop artificial intelligence as a solution component in the energy transition. One specific use case here is predicting how many photovoltaic systems will place demands on the grid in the coming years and what power needs to be available to ensure a stable electricity grid. This should form the basis of knowledge for a needs-based grid expansion and better risk assessment.
Goal
How much solar energy is produced when and where is it fed into the grid? Answering this question is the key for Primeo to be able to better plan structural adjustments to the grid (transformers, distribution nodes, connection reinforcement of individual properties) in line with future requirements.
The aim of Primeo Netz AG is to provide a reliable performance forecast for the next 5 years in the photovoltaic sector. Specifically, this involves forecasting the annual peak for the feed-in of photovoltaic systems in 15-minute resolution per outgoing circuit at transformer stations.
Both the potential production volume of existing PV systems and future PV expansion will be considered.
The inclusion of data science in this process is intended to make the forecasts, which have so far been based heavily on expert knowledge and potential reports, even more reliable, data-driven and dynamic.
Solution
Following the development of the data science strategy, Primeo Netz AG is also relying on eoda's expertise for the implementation of this use case. In order to validate the feasibility and benefits of the use case, the initial implementation - as is usual for data science projects - is carried out as part of a proof of concept. This was based on several sample regions from the Primeo network.
The data basis comprises various sources, which are essentially based on historical production information in the PV sector and potential analyses for the future:
- Measuring point data in 15-minute resolution from the SAP system
- Data on renewable energy systems
- Data from the grid information system
- Geodata with the coordinates of transformer stations, measuring points and house connections
- Building data, territorial authority boundaries and construction zones from the Federal Statistical Office
Primeo also has two potential studies at its disposal, which take into account aspects such as the duration of sunshine in the individual regions or the orientation and inclination of the roofs. A particular challenge: the highly dynamic nature of this environment, for example due to the influence of political decisions. This is another reason why the available data basis for the goal of a 5-year forecast is rather limited.
The data science experts at eoda checked the available data sources for quality and usability for the use case and implemented the necessary measures for data preparation and linking.
For the forecasting model, eoda relies on a combination of survival analysis and probability simulations. This approach enables statements on forecast uncertainty and the easy integration of existing and future studies with PV growth scenarios in order to map the existing dynamics. This approach also allows comparatively unlikely scenarios to be estimated as part of risk-averse planning.
For the interactive visualization, eoda relies on the Python framework Streamlit as a fast and dynamic dashboard solution, which enables the experts at Primeo Netz AG to use the forecast results for their work in a simple and understandable way.
In addition, the interactive visualization offers the opportunity to gain knowledge beyond the actual forecasts.
As part of a training course, the Primeo Netz AG team was enabled to further develop and operate the solution independently.
Result
With the reliable forecast of the future load on the grid caused by PV systems, eoda creates a valuable - because reliable - knowledge base for Primeo Netz AG. For Primeo, this is the planning basis for grid expansion. In addition, the results can be used for calculating the load profile and voltage tolerance as well as for more focused energy consulting in areas with PV expansion.
This project is therefore a successful example of how data science can become a building block for a stable electricity grid in the future.
Get started now:
We look forward to exchanging ideas with you.
Your expert on Data-Science-Projects:
Manfred Menze
projects@eoda.de
Tel. +49 561 87948-370
