Researchers propose a system based on predictive control and energy storage to minimize the impact of seawater transport in arid zones, anticipating the surge in saltwater use projected for 2032.

The growing utilization of seawater in Chilean mining, driven by water scarcity in arid zones and stricter environmental regulations, is significantly raising the energy requirements associated with the transport of this resource. Faced with this scenario, researchers from the Department of Electrical Engineering at the University of Concepción developed a real-time energy management model for pumping stations, which incorporates economic model predictive control (EMPC) along with solar photovoltaic generation and battery storage systems.

The study, recently published in IEEE Transactions on Industry Applications, addresses the problem from a multivariable optimization perspective, considering both the hourly variability of electricity tariffs—including real-time pricing (RTP) and maximum demand (MD) charges—and the operational constraints of the hydraulic and electrical systems.

The application case corresponds to a pumping system located in northern Chile, responsible for supplying a reverse osmosis plant and a pumping station consisting of seven 1343 kW pumps each, used to drive water 120 km away with an elevation difference exceeding 1000 meters. Based on data, different operating scenarios were evaluated: from conventional operation based solely on water demand to hybrid configurations with solar generation and Megapack-type battery storage.

The simple use of an intermediate reservoir for load shifting allowed for a 29% reduction in operating costs. The incorporation of solar generation not only enhanced that reduction but also allowed for tackling maximum demand charges, which represent a significant portion of monthly expenses. In parallel, the exclusive use of batteries proved to be less efficient on its own, but its inclusion in a hybrid model with solar generation was key to achieving the best system performance.

This combined configuration achieved a reduction of nearly 50% in total operating costs and a 75% reduction in costs associated with RTP tariffs, without compromising the continuity of the water supply. The model is based on a receding horizon approach, with hourly control intervals, and employs a SARIMA prediction system to estimate future tariffs, given their high volatility. The mathematical formulation of the problem was posed as a mixed-integer linear programming (MILP) model, solved using the HiGHS optimizer programmed in Python. Simulations confirmed the stability of the system under all operational constraints, including tank volume limits, hourly solar availability, and battery charging and discharging capacity.

As explained by Daniel Sbarbaro, SERC Chile researcher and author of the paper, “the study was motivated by the sustained increase in electrical energy consumption associated with pumping seawater to mineral concentration processes, a practice increasingly common in areas with water scarcity.”

Sustainable solution to saltwater demand

According to Cochilco projections, the use of seawater in mining will increase by 167% between 2021 and 2032, while continental water consumption will fall by 45%. This structural change will imply an increase in energy demand associated with water transport, making the implementation of efficient and economically viable solutions—such as those proposed by this model—urgent.

In this context, Sbarbaro mentioned that “seawater pumping can represent up to 20% of the total electrical energy consumption of a concentrator plant. Given this level of impact, there is growing interest from the mining sector in developing solutions that allow for more efficient management of this consumption, from both an economic and environmental perspective.”

“From a technical standpoint, the greatest challenge was the formulation and calibration of a model capable of robustly representing the temporal evolution of energy prices,” he added. On the other hand, he clarified that the implementation does not require significant investments in infrastructure: “The model can be integrated into existing control systems using standard computing platforms. The proposed strategy is designed to be operational with the SCADA systems already present in most mining operations.”

In order to continue developing this initiative, the SERC Chile researcher indicated that “future research in this field could explore the optimization potential of the system using optimal sizing and global optimization techniques. Optimal sizing would consist of determining system components through an economic analysis of the investment. On the other hand, global optimization would focus on the simultaneous optimization of multiple pumping stations, leveraging previous studies to identify the most efficient design for the system at a larger scale.