Dataiku is GRDF’s handy toolbox for code management, collaboration, and more.Learn More
Solvay is a global leader in Soda Solvay® sodium carbonate and sodium bicarbonate production. These products are present in a wide range of applications, from the glass industry to detergent, metallurgical processes, pulp and paper, and supplements in the pharmaceutical industry. Its production is both an important part of the company’s activity as well as an energy-intensive process with complex implications for energy management.
Several soda ash production plants are distributed across the globe to supply clients around the world, and the production plan for the plants is decided by the central sales and operations planning (S&OP) team. They determine production allocation based on the projection of the variable cost of each plant along with logistic constraints.
Using Dataiku, the energy technology team built a solution that allows the S&OP team to see soda ash production costs and carbon dioxide (CO2) emissions on an hourly basis and compare those costs with an optimized, planned scenario. Ultimately, this solution improves soda ash production by both reducing production costs as well as energy consumption, paving the way for a growing and sustainable business. Here’s a look at how they did it.
Taking a step back, the team did face a variety of challenges in building and implementing a solution.
On the technical front, the chemical process is complex and energy intensive. Not only are energy markets highly volatile, but the contracts and taxonomy differ from one country to another. In addition, asset configuration has a huge impact on CO2 emissions.
But on top of the technical challenges, the energy technology team also came across some people and process challenges. For example, the scope and development of the project had to be done via collaboration with the production team, data science engineering team, and asset management team — that’s a lot of stakeholder management. Plus, given how critical the project is for the company, the solution had to be robust.
The energy technology team at Solvay used Dataiku to build a solution that assesses the variable soda ash production costs and CO2 emissions on a plant-by-plant basis. The project currently runs continuously every hour in six plants (with more to come!). The output of the solution is available via both visual Dataiku dashboards as well as in spreadsheets for easier access to a wider population across the business.
In terms of input data, the Dataiku project has:
- Real-time process data that comes from the plant (steam and electricity generation, soda ash production, etc.).
- Spot energy prices that come from the market — this data is delivered via a shared database from other Dataiku projects in Solvay that retrieve the same data.
- Specific energy contract implementation (gas, power, etc.) and CO2 emission factors, plus other raw material costs such as limestone, brine, ammonia, etc.
To build the solution, the energy technology team first had to identify the meaningful parameters from the plant data that shape the soda ash variable costs and CO2 emissions, then generate the equations. From there, they could build the dashboards with the most important parameters. It took just one month and one person to develop a beta version of the solution for one plant. Once the methodology was validated, it could be easily replicated for the rest of the plants in Dataiku, with some customizations.
Value to the Business
Thanks to the solution the energy technology team developed in Dataiku and in collaboration with the business, there is a better understanding of the influential parameters on the overall cost of soda ash production. This understanding enables the business to reduce operational costs and make faster decisions.
The solution proved so valuable in such a short amount of time that Solvay has already vowed to improve it and onboard more plants in the process. In addition to simplifying and making dashboards more efficient, the team is also looking to build in a real-time comparison between the actual data versus models used for the optimization (based on historical data) and versus a benchmark.
With this replicable project built in Dataiku, Solvay soda ash is setting the foundations to embed AI in processes across the business. In line with the G.R.O.W. strategy, it paves the way for a business that is growing and sustainable.