⚡ electrical engineering

Algorithmic Verification of Energy Conservation in Discrete Cyber Physical Fluid Networks Using Mathematical Induction

This paper proposes a formal, non-experimental algorithmic framework that uses mathematical induction to verify energy conservation in discrete, acyclic cyber-physical fluid networks, achieving linear-time anomaly detection and significantly reducing computational latency compared to traditional numerical simulation methods.

Original authors: Syed Eirfan Atthar

Published 2026-10-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Syed Eirfan Atthar

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Modern cities and factories rely on invisible networks of pipes to move water, cool data centers, and distribute heat. These systems are physical, but they are increasingly managed by computers that watch sensors and make decisions. This marriage of physical machinery and digital logic is known as a cyber-physical system. For these networks to work safely, they must obey the fundamental laws of physics, specifically the rule that energy cannot simply vanish or appear out of nowhere. If a computer thinks a pipe is carrying a certain amount of heat, but the actual sensors show a different amount, something is wrong. It could be a leak, a broken sensor, or a hacker trying to trick the system. The challenge for engineers has always been how to check these massive, complex networks quickly enough to catch problems before they cause damage.

For decades, the standard way to check these systems has been to use powerful computers to simulate the flow of fluids. These simulations try to solve complex equations that describe how water moves and changes temperature. While this works well for designing a new system, it is too slow for watching an existing one in real time. As a network grows larger, the time required to run these calculations grows explosively, making it impossible to check a huge city's water grid instantly. A new study by Syed Eirfan Atthar at Dibrugarh University in India proposes a different approach. Instead of trying to solve the physics equations to predict what the system should be doing, the researcher suggests simply checking if the measurements we already have make sense. By treating the network as a logical structure rather than a fluid puzzle, the study shows that we can verify the system's integrity almost instantly, even for very large networks.

The core idea relies on a simple logical principle: if you can prove that a single part of a system is working correctly, and you can prove that adding one more part to a working system keeps the whole thing working, then the entire system must be correct. The researcher applied this logic, known as mathematical induction, to fluid networks. First, the continuous flow of water in pipes is broken down into a series of distinct points, or nodes, connected by lines. The study focuses on networks that flow in one direction without loops, like a tree with branches spreading out from a trunk, rather than a closed circle. For every single point in this network, the computer checks if the energy coming in matches the energy going out, allowing for a tiny margin of error caused by imperfect sensors.

If the energy balance holds true for the first point, the logic moves to the next. Because the network has no loops, the computer can move from one node to the next in a specific order, checking each one individually. If every single point passes the check, the entire network is verified as safe. This method bypasses the need for the heavy, slow calculations used in traditional fluid simulations. Instead of solving a giant puzzle where every piece depends on every other piece, the new method checks the pieces one by one. The result is a verification process that is not just fast, but mathematically guaranteed to be correct, provided the sensors are accurate and the network flows in one direction.

To test this idea, the researcher created a computer simulation of a network with one hundred nodes, representing a municipal cooling system. The system was fed with realistic data, including small amounts of random noise that real sensors always produce. The researcher then introduced deliberate errors, such as simulating a leak where five percent of the water flow disappeared, or injecting false data to trick the system. The new algorithm, which the study calls AVEC, successfully identified these problems. It caught the leaks and the fake data with high accuracy while ignoring the normal background noise of the sensors. The study found that the algorithm could distinguish between a real problem and a sensor glitch by using a dynamic threshold that adjusts based on how much the sensors usually vary.

The most striking result of the study is the speed difference. When the researcher compared the new method against traditional simulation tools, the difference was dramatic. For a small network of ten nodes, the traditional method took about two milliseconds to check, while the new method took only a fraction of that. But as the network grew, the traditional method slowed down drastically. Checking a network of one thousand nodes took the traditional method over thirty-eight seconds, and a network of five thousand nodes took more than five minutes. In contrast, the new method checked the five-thousand-node network in just twenty-four milliseconds. This shift from minutes to milliseconds means that a computer could theoretically watch a massive industrial network in real time, spotting a leak or a cyberattack the moment it happens, rather than waiting for a slow calculation to finish.

The study is clear about its boundaries. It does not claim to solve problems for systems with loops, where water circulates back on itself, nor does it work for systems where the flow changes rapidly over time. It is designed specifically for steady, one-way flows where every part of the system is fully observed by sensors. If a sensor is missing or the data is incomplete, the method cannot work because it relies on having the full picture. However, for the specific type of network it targets, the research demonstrates that a rigorous, mathematical proof of safety can be turned into a fast, efficient computer program. This approach offers a way to secure critical infrastructure without the heavy computational cost that has long limited real-time monitoring. By proving that the laws of physics hold true at every single step of the way, the system ensures that the whole network is safe, turning a complex physical problem into a straightforward logical check.

Technical Summary: Algorithmic Verification of Energy Conservation in Discrete Cyber Physical Fluid Networks

Problem Statement
Modern Cyber-Physical Systems (CPS) managing critical infrastructure, such as municipal water distribution and industrial cooling grids, rely on strict adherence to thermodynamic and fluid mechanical laws. Traditional anomaly detection methods utilize continuous numerical simulations (e.g., Computational Fluid Dynamics or Newton-Raphson solvers) to verify system states. However, these methods suffer from two primary limitations when applied to real-time verification of fully observed networks:

Computational Inefficiency: Solving non-linear algebraic equations derived from Navier-Stokes formulations requires repeated matrix inversions, resulting in a computational complexity of $O(V^3)$ , where $V$ is the number of nodes. This cubic scaling renders real-time verification of large-scale networks computationally prohibitive.
Methodological Flaw: There is a conflation in existing literature between solving for unknown fluid states and verifying already-measured states. Continuous solvers are designed for prediction, not the deterministic verification of sensor data integrity.

Methodology
The paper proposes a strictly theoretical, non-experimental framework that transforms the verification problem from a continuous physics-solving task into a discrete data-integrity check. The methodology proceeds as follows:

Discretization and Graph Formulation:
- The physical thermo-fluid network is mapped to a Directed Acyclic Graph (DAG), $G=(V, E)$ .
- Constraints: The framework is bounded to single-phase, incompressible, steady-state flows. It strictly requires full observability, meaning high-fidelity sensors must provide all state variables (mass flow, pressure, velocity, elevation, internal energy) for every edge. Cyclic networks (e.g., closed cooling loops) are excluded as they require non-linear iterative methods outside the proposed deterministic structure.
- Data Structure: Each edge $e_{ij}$ is defined by a discrete data tuple $S_{ij} = \langle \dot{m}_{ij}, P_{ij}, v_{ij}, z_{ij}, u_{ij} \rangle$ .
Formal Proof via Mathematical Induction:
- The core logic utilizes the Principle of Mathematical Induction (PMI) to prove that if local energy conservation holds for every node, global conservation holds for the entire network.
- Base Case ( $k=1$ ): Validates energy balance for a single independent node using sensor data.
- Inductive Step: Assumes a verified network of $n$ nodes ( $G_n$ ) is balanced. It proves that adding a new node ( $v_{n+1}$ ) preserves global balance if and only if the local conservation proposition $P(v_{n+1})$ is true. The bridging edges between the existing network and the new node cancel out mathematically, isolating the verification to the new node's local data.
Algorithmic Implementation (AVEC):
- The proof is translated into the Algorithmic Verification of Energy Conservation (AVEC) algorithm.
- The algorithm performs a topological sort of the DAG and traverses nodes linearly.
- Dynamic Thresholding: To account for sensor noise, the algorithm does not require a perfect zero residual. Instead, it calculates a dynamic error threshold $\epsilon_i$ based on the cumulative variance of associated sensors and a strictness factor $z$ . An anomaly is flagged only if the energy residual exceeds this statistically derived threshold.

Key Contributions

Paradigm Shift in Verification: The paper introduces a formal verification approach that bypasses the need for inverse modeling or continuous simulation for anomaly detection, shifting the problem from $O(V^3)$ matrix operations to a linear-time data traversal.
Deterministic Security: By isolating nodes via induction, the algorithm can instantly pinpoint the exact physical location of a failure (sensor drift, leakage, or data injection) without error propagation across the network.
Complexity Reduction: The proposed AVEC algorithm achieves a time complexity of $O(V + E)$ , a significant reduction compared to traditional solvers.

Results
The framework was validated using a synthetic 100-node directed acyclic graph representing a municipal cooling network.

Anomaly Detection: With a strictness factor of $z=3$ , the system achieved a True Positive Rate (TPR) exceeding 0.94 for detecting leaks and data spoofing, while maintaining a False Positive Rate (FPR) below 8% against Gaussian sensor noise.
Computational Latency: Empirical benchmarks on an Intel Core i7 processor demonstrated a dramatic reduction in execution time:
- 100 nodes: Traditional solver: ~450 ms; AVEC: ~0.48 ms.
- 1,000 nodes: Traditional solver: ~38 seconds; AVEC: ~4.95 ms.
- 5,000 nodes: Traditional solver: >5 minutes (prohibitive); AVEC: ~24.10 ms.
Scalability: The results confirm that AVEC maintains ultra-low, linear latency regardless of network size, whereas traditional solvers scale cubically.

Significance and Claims
The paper claims to offer a "highly efficient integrity check strictly downstream of physical sensors." Its significance lies in three areas:

Mitigating State-Space Explosion: By discretizing physical laws and using induction, the framework avoids the state-space explosion problems inherent in continuous model checking for fluid systems.
Real-Time Security: The linear-time capability allows for continuous, high-frequency verification of data integrity, enabling immediate detection of malicious data injection or physical leaks before they compromise system controllers.
Green ICT: By eliminating the need for heavy CFD or numerical matrix simulations for monitoring, the approach reduces the processing power and energy consumption of central monitoring servers, aligning with Green Information and Communication Technology goals.

The study concludes that for fully observed, acyclic, steady-state fluid networks, global energy conservation can be mathematically guaranteed through localized verification, providing a scalable, computationally inexpensive security layer for modern CPS infrastructure.

Technical Summary: Algorithmic Verification of Energy Conservation in Discrete Cyber Physical Fluid Networks

More like this