From Flood to Cause

My name is Grigorios Agathangelidis, and my professional background is in Electrical Engineering and Software Engineering. I am not a nuclear physicist, nuclear engineer, control-room operator, or specialist in industrial diagnostics. The ideas presented in From Flood to Cause are not derived from direct operational experience in nuclear facilities or other safety-critical plants. Rather, they are the result of extensive independent study and a long-standing interest in how complex systems fail, recover, and explain themselves.

This book emerged as a natural extension of the NEXUS-1 project. While From Grid to Core explored the architecture of a deterministic, explainable, and auditable digital-twin platform, From Flood to Cause focuses on a more specific question: how can we identify the true origin of a failure when hundreds of alarms, symptoms, and signals compete for attention?

My interest in this problem extends beyond the nuclear domain. Alarm floods, cascading failures, misleading symptoms, and hidden root causes appear in every sufficiently complex system: industrial plants, power grids, telecommunications networks, data centres, transportation systems, and modern software platforms. Although the examples in this book are drawn from a nuclear-inspired environment, the underlying principles are intentionally broader.

Throughout this work I have tried to combine disciplines that are rarely discussed together: root-cause analysis, fault propagation, systems engineering, digital twins, observability, retrieval-augmented generation (RAG), explainable AI, verification and validation (V&V), and software architecture. My goal has been to explore how deterministic reasoning and modern language models can coexist without compromising traceability, auditability, or engineering rigor.

A central theme of this book is that language models should explain decisions, not make them. The architecture presented here places deterministic inference, telemetry corroboration, and grounded evidence ahead of AI-generated explanations. Whether or not the specific implementation proves useful, I believe the broader principle—that trustworthy systems require clear boundaries between reasoning and narration—will remain relevant far beyond the domains discussed here.

The cases, architectures, and mechanisms presented throughout this book are educational and conceptual in nature. They are intended to illustrate engineering principles, design patterns, and analytical approaches rather than provide certified operational solutions. Any resemblance to real facilities, events, or systems serves the purpose of explanation and learning, not operational guidance.

What interests me most is not any single technology but the intersection of technologies: how telemetry becomes evidence, how evidence becomes explanation, how explanations remain accountable, and how humans retain authority in systems increasingly assisted by automation and artificial intelligence. These questions sit at the crossroads of engineering, software, safety, and decision-making, and they continue to shape my work.

From Flood to Cause is therefore not only a book about alarm management, root-cause analysis, or AI-assisted diagnostics. It is also an exploration of a broader engineering philosophy: that understanding should precede automation, that evidence should precede conclusions, and that every claim made by a machine should be traceable to the facts that support it.

The ideas collected here represent a snapshot of an ongoing journey of learning and experimentation. Some concepts will undoubtedly evolve; some assumptions may prove incomplete. That is the nature of engineering. Progress comes not from claiming certainty, but from continually refining our understanding in the face of new evidence.

If there is a single message I hope readers take away from this work, it is that complex systems do not become understandable through more data alone. They become understandable when data is organized into evidence, evidence into causality, and causality into explanations that humans can verify, challenge, and trust.

Above all, this book reflects a belief that curiosity, disciplined reasoning, and intellectual humility remain among the most powerful tools available to engineers. Technology changes rapidly; the principles of careful observation, evidence-based analysis, and honest acknowledgment of uncertainty do not.