Water treatment operators face a critical trust problem: when an AI system recommends "increase methanol dosing by 18%," they have no way to understand why. A new approach called Physics-Informed AI is solving this by embedding deterministic process knowledge directly into machine learning models, making AI recommendations transparent and auditable from the ground up. Why Black-Box AI Fails in Water Treatment? The water sector has long relied on deterministic, mechanistic models rooted in physics. Engineers use frameworks like the Activated Sludge Model family (ASM1, ASM2d, ASM3) to simulate wastewater treatment, and the Anaerobic Digestion Model No.1 (ADM1) for sludge digestion. These "white box" models are fully interpretable; every parameter has physical meaning, and every equation can be questioned and understood. But pure machine learning models sit at the opposite extreme. They are powerful and adaptive, yet structurally opaque. When a data-driven tool ingests sensor data and SCADA streams, performs statistical operations, and returns a number or alert with no explanation, operators understandably hesitate to act on it. A conversational AI agent that translates a black box's output into readable sentences is not the same as a model whose reasoning is inherently transparent. One translates the black box; the other dismantles it. What Is Physics-Informed AI and How Does It Work? The most promising path forward lies in the middle: hybrid "grey box" models that embed deterministic process knowledge within a data-driven framework. In a parallel hybrid architecture, a mechanistic model provides the structural backbone and mass-balance coherence, while a machine learning component learns the residuals, capturing variability, sensor drift, and process dynamics that equations alone cannot represent. The result is explainability by design. The mechanistic layer gives operators something they can reason about; biological and chemical knowledge is preserved and visible. The machine learning layer acts as an adaptive corrective lens rather than a mysterious predictor. When such a Physics-Informed AI model is embedded in a digital twin with real-time data connectivity, a recommendation like "increase aeration" can be traced back to specific process signals and model states. Physical grounding is what makes the model inherently explainable. How to Make AI Predictions Fully Traceable in Water Systems - Sensitivity Analysis: A classical technique familiar to process engineers, this method asks a simple but powerful question: if input variable X changes by a small amount, how does model output Y respond? Applied to hybrid models, it reveals which biological rate constants, hydraulic retention times, or influent load parameters most strongly drive predictions, focusing operator attention precisely where it matters most. - SHAP Values: Drawn from cooperative game theory, SHAP (SHapley Additive exPlanations) values quantify the contribution of each input variable to any individual prediction. Instead of seeing "increase methanol dosing by 18%," the operator sees: "Nitrate spike detected at reactor inlet: +12%, drop in water temperature: +4%, low dissolved oxygen: +2%. Recommended to increase methanol dosing by 18%." The recommendation is the same; however, the reasoning is now auditable. - Combined Audit Trail: Applied together to a Physics-Informed hybrid model, sensitivity analysis interrogates the deterministic backbone while SHAP illuminates the adaptive machine learning layer. The combination delivers a full audit trail from raw process data to actionable recommendations, which is precisely what operators and managers need to move from skepticism to confidence. What Real-World Impact Does Explainable AI Have on Adoption? For utility managers, the business case is direct. Explainable model outputs reduce the frequency with which valid AI recommendations are overridden by skeptical operators, which is a well-documented barrier to digital-tool adoption. Research shows that hybrid modelling can improve wastewater treatment model accuracy significantly, reducing prediction errors under challenging validation conditions and increasing confidence for use as a decision-making tool for managers, operators, and process engineers. The human dimension of this challenge is critical. A process engineer at a large wastewater treatment plant needs to trust the system before acting on its recommendations. When an operator understands not just what the AI recommends, but why it recommends it, skepticism transforms into confidence. This shift from distrust to adoption is the real-world payoff of Physics-Informed AI. When Should Water Utilities Use AI Versus Traditional Models? A critical warning accompanies this enthusiasm around AI tools: not every water-sector challenge requires machine learning. For well-understood processes with stable dynamics and sufficient mechanistic knowledge, a calibrated deterministic model remains more appropriate. These traditional models are more interpretable, more robust to data scarcity, and more defensible to regulators. Machine learning should be deployed where it adds genuine value: handling nonlinearities that physics-based models cannot capture, filling sensor gaps and creating virtual sensors, or accelerating real-time predictions at scale. Choosing the right tool requires intellectual honesty about what the data can and cannot support. XAI features, whether post-hoc explanation methods such as SHAP and sensitivity analysis applied on top of machine learning layers, or structural interpretability built into Physics-Informed AI architectures from the start, should become standard requirements in water engineering. The shift toward Physics-Informed AI represents a fundamental change in how the water sector approaches digital transformation. By combining the predictive power of machine learning with the explanatory capacity of physics-based modelling, utilities can finally build AI systems that operators trust enough to act on. In an industry where a single wrong decision can affect water quality for thousands of people, that transparency is not a luxury; it is essential.
