"Thanks to this new solution, our employees can control our energy assets in a quarter of the time it took before. By optimizing the process from beginning to end, we can make the most out of our new solar panels, maximize the amount of energy from renewable sources used in our factory, and bring our operational costs down at the same time"
Area: Energy
Problem class: MIP
Technologies: SaaS, GAMS, GAMS MIRO
Alpro is one of Europe’s leaders in the production of plant-based products. Based in Belgium, France, and the UK, they market a wide range of food and beverages made of soy, almonds, hazelnuts, cashews, rice, oats, and coconut. With a big operation in place, their factories handle complex processes that demand substantial energy, and optimizing its management is an essential, yet often quite difficult task.
In this particular project, Alpro partnered with our consulting team at GAMS to maximize the efficiency of one of their plants’ energy systems. Our shared goal was simple: to design a tool that would help streamline their day-to-day energy operations, and facilitate a better and more data-driven decision making.
Working together, we developed two custom optimization models, along with a web-based graphical user interface (GUI) that drastically transformed their workflow. As a result, Alpro not only gained a competitive advantage in their energy management, but also took a significant step towards improving their operational costs and achieving their sustainability goals at the same time.
Managing energy systems in industrial facilities is often a very complicated task. In Alpro’s case, they needed to integrate their own electricity and steam production from generators, boilers, and solar panels, with outside energy sources; all in order to meet the fluctuating demand of their daily operation.
This presented a huge challenge for the team.
Constantly manipulating their energy assets to meet demand was a laborious and complicated task on its own. But at the same time, the option of buying energy from the grid, which sometimes is cheaper than producing their own, or selling their generation surplus back to them for a profit, added a whole new level of complexity.
Now factor in that grid energy has constant price variations, that their solar panels depend on weather conditions, that their facility’s energy demand changes throughout the day, and that they have to follow a maintenance schedule, and what you get is a highly intricate and tedious to manage system.
The path to a simpler and more efficient system was clear: to mathematically optimize the process –to take all variables and constraints as data input, and build a model that could suggest the best energy management solution at any given time.
In close collaboration with Alpro, our team of experts broke the problem down into two parts and developed a custom model to deal with each step. Then, we integrated both models into a single, user friendly, and intuitive GUI.
Here’s how it works.
The first optimization model is in charge of gathering data from multiple sources, and generating conditional bids for the electricity day-ahead market –whether Alpro should buy, or sell energy, and under what conditions. Information from spot price forecasts, solar energy forecasts, factory demand projections, and maintenance schedules is automatically collected and processed; then, Alpro’s operational goals and financial considerations are factored in, and the resulting bids are submitted to the day-ahead auction.
This first step of the process saves Alpro valuable time, all while ensuring that their energy planning is based on the best and most up-to-date information.
When the day-ahead energy market clears, the second model kicks in.
This model is in charge of retrieving the accepted bids from the market, and then using that information to optimize the load of each of Alpro’s energy generating assets. This is done taking into account the factory’s energy consumption prediction, and also the expected production from their own solar plant. By finding the optimal solution, Alpro not only covers their energy needs, but also makes sure to take full advantage of their renewable generation.
The second model’s output is information on how to operate each asset on a 15-minute interval basis. This detailed data is automatically passed onto their control systems, which take over the job for the rest of the day, ensuring constant and optimal management with minimal work.
With the two models working together, there’s a seamless integration between information, energy trading, and operational execution. More than that, while all the complicated math happens in the background, the data and results are clearly displayed in an intuitive dashboard. This new GUI provides Alpro with clear insights into asset operation, day-ahead trades, solar production, and multiple key performance indicators. It’s an easy and user-friendly interface, which allows for informed decisions, and to continually improve every aspect of their energy usage.
“Thanks to this new solution, our employees can control our energy assets in a quarter of the time it took before”, said Dominique Hamerlinck, Energy Manager at Alpro. And it isn’t only about speed and efficiency; Dominique emphasized that one of the main incentives to deploy GAMS models was to minimize costs and reduce their environmental footprint. “By optimizing the process from beginning to end, we can make the most out of our new solar panels, maximize the amount of energy from renewable sources used in our factory, and bring our operational costs down at the same time”.
Since Alpro deployed their optimization models, the energy management process is 3 times faster than before, while costs have been reduced by 25%.