Total

Optimizing the return rate of a steam network at Total with the creation of an optimized flow propagation model and real-time identification of anomalies.

Context

The petroleum refining process involves a power grid that powers the industrial asset and heats the water to steam.
Any loss in the steam network has a direct impact on the energy efficiency of the installation, and on the costs of production.
Its optimization is therefore crucial for the profitability of Total's refining activities.

DCbrain intervention

After creating a digital double of the steam network, DCbrain analyzed the plant's building management system data.
The analysis showed that to optimize the steam system as a whole, it was necessary to optimize each unit of steam production separately.
DCbrain has calculated, through Machine Learning, the efficiency level of each unit before re-aggregating these data into streams.
Our ability to propagate the output of each unit in the global network explains the accuracy of our results.