The Center for Data Innovation recently spoke with Martín Nogueira, co-founder of Humara, a Spain-based company developing AI-powered software and digital twins for waste-treatment facilities. Nogueira explained how the company’s platform combines engineering models and real-time operational data to help operators design facilities, simulate upgrades, and improve plant performance.
David Kertai: What problem is Humara solving?
Martín Nogueira: Waste-treatment plants are highly complex because they process different types and quantities of waste every day. Yet most are still designed and operated using spreadsheets, static engineering drawings, and disconnected software tools. Operators have no reliable way to predict how changes, such as reconfiguring a processing line, installing new equipment, or receiving a different mix of waste, will affect plant performance. As a result, they often make major operational and investment decisions without testing their assumptions first.
Humara addresses this challenge through two connected software products built around a digital twin—a virtual replica of a facility that mirrors its equipment and operations. Engineering teams use Humara Design to build a virtual plant with equipment from our catalog. The platform automatically performs mass-balance calculations that track how materials move through each stage of the facility, generates technical documentation, and creates a 3D model. Humara Operate then connects directly to a plant’s existing control systems so the digital twin continuously receives live operational data. Together, these tools help operators design facilities more efficiently and optimize day-to-day performance.
Kertai: What types of operational data does your AI system analyze to support decision-making?
Nogueira: To understand plant performance, the platform analyzes operational data from across the facility. This includes the type and quantity of incoming waste, equipment operating conditions, processing throughput, equipment uptime, recovery rates that measure how much recyclable material is captured, and the purity of each recovered material stream.
Unlike traditional monitoring systems, our platform connects this data through engineering models that represent how materials and equipment behave throughout the facility. Rather than simply reporting performance metrics, it explains why performance changes and shows how one adjustment can affect the entire operation.
Kertai: How does the platform use virtual facility models to test change before implementation?
Nogueira: We create a virtual copy of each facility that accurately represents its equipment, configuration, and material flows. Before making changes to the real plant, operators can test different scenarios inside the digital twin. They can simulate new waste streams, adjust equipment settings, modify processing lines, or evaluate major upgrades and immediately see how those changes affect recovery, throughput, and product quality.
Kertai: How do you ensure the accuracy of your AI-driven recommendations?
Nogueira: We ensure accuracy by grounding our AI models in engineering principles instead of relying only on statistical patterns. Our models combine mass-balance calculations with the known behavior of real equipment, then continuously refine their predictions using live operational data from each facility.
Since every recommendation is tied to the plant’s underlying engineering model, operators can understand why the system reached a conclusion instead of treating it as a black box. That combination of engineering expertise and continuous calibration produces recommendations that are reliable, transparent, and practical for day-to-day operations.
Kertai: Could you share an example of your system in use?
Nogueira: We currently work with leading waste operators in Spain, including facilities owned by major international companies. In one materials-recovery facility, the digital twin monitors performance in real time and compares actual output with the plant’s theoretical optimum. The system identified recyclable materials that should have been recovered but were instead being lost during processing, helping operators pinpoint opportunities to improve recovery rates.
Engineering teams have also used Humara Design to evaluate retrofit options and generate complete technical documentation much faster than traditional design tools. By combining facility design, simulation, and daily operations in a single platform, we help operators make better-informed decisions before making physical changes.
