Energy and logistics optimization_
Or how to optimize the consumption of a storage site!
Energy optimization in an agricultural context_
Our client offers a range of digital tools, intended for agricultural cooperatives, to facilitate and optimize the production, storage and sale of cereals.
In order to expand its offer, in a tense energy context, our client wishes to propose a tool for optimizing the electricity consumption of grain storage sites.
More specifically, the grain storage silos are equipped with fans, which allow them to preserve the quality of the cereals by maintaining an adequate temperature inside the silos. These fans are particularly energy-intensive and can be completely ineffective if not used correctly (for example if it is too hot outside). It therefore appears very important to be able to control and optimize the operating periods of the fans.
The objective of this project is to develop a tool to automatically detect the hours of fan ignition on a grain storage site, based on a simple reading of the site's total consumption.
Energy optimization: the OSE mission_
With this objective of energy optimization, OSE has developed a web service that automatically detects the hours of operation of the fans of a storage site. This service interfaces with our client's application, which sends the calculation requests and retrieves the results for display to the user.
To work, the algorithm developed by OSE only needs to know:
The electrical power of the fans installed on the storage site
The curve for the total electricity consumption of the storage site (fans and all other equipment)
The algorithm automatically (and instantly) detects fan on times and returns the results. The algorithm has been put into production and is successfully used on a hundred storage sites with the aim of saving energy.
Implementation of advanced signal processing techniques as well as expert rules to ensure high robustness and explainability of automatic ignition hour detection.
Energy optimization: the benefit of the OSE solution_
The automatic detection of fan on times based on the site's total electricity consumption reading is a technical challenge that OSE has taken up. Indeed, the consumption of the fan(s) is drowned in the consumption of all the other equipment that operates in parallel and energy optimization is a major issue.
It was therefore necessary to finely characterize the patterns of the consumption curve of the fans to be able to detect their ignition even when other equipment is activated or extinguished. For this, the scientific approach used is based on two pillars:
🟠Advanced analysis and processing of the electrical consumption time signal (sampled in 10min time steps) in order to extract and recognize characteristic
🟠The use of expert rules obtained through significant exchanges with business engineers.
This combination provided a high level of robustness, explainability and accuracy of results across a wide variety of sites with different combinations of installed fans and equipment.
The results were evaluated by comparison with timestamp readings of the automatic ventilation systems extracted manually from a few storage sites.
2 développeurs OSE mobilisés
Durant 6 mois