AI-Driven Anomaly Detection in IoT-Enabled HVAC and Water Heating Systems
Open accessAbstract
As buildings evolve into intelligent, connected ecosystems, Heating, Ventilation, and Air Conditioning (HVAC) systems — along with water heating infrastructure — are becoming increasingly IoT-enabled. These systems generate continuous streams of sensor data that can be harnessed to ensure energy efficiency, occupant comfort, and operational reliability. However, identifying performance degradation or faults within these complex, multi-component systems remains a challenge. This paper explores an AI-driven approach for real-time anomaly detection in IoT-enabled HVAC and water heating systems. By leveraging machine learning (ML) and advanced data analytics, the proposed framework identifies abnormal patterns, predicts potential failures, and facilitates proactive maintenance — enabling energy savings, system longevity, and sustainability.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Alagappan demonstrated the practical applicability of AI-driven anomaly detection in HVAC and water heating ICS deployments domains with highly periodic telemetry patterns and strong process-physics constraints, providing domain-specific empirical grounding for the semantic anomaly detection layer described in Section VIII-C.
“Alagappan demonstrated the practical applicability of AI-driven anomaly detection in HVAC and water heating ICS deployments domains with highly periodic telemetry patterns and strong process-physics constraints providing domain-specific empirical grounding for the semantic anomaly detection layer described in Section VIII-C.[33]”
From Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems (2026) · cites this paper as [33]
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
Experimental — accuracy not yet measuredTransport-layer anomaly detection identifies deviations that may indicate device compromise, unauthorized firmware modification, or replay injection campaigns.
“Transport-layer anomaly detection, operating on statistical baselines of message frequency, payload size distribution, and inter-message timing patterns, identifies deviations that may indicate device compromise, unauthorized firmware modification, or replay injection campaigns.[25], [30], [33]”
From Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems (2026) · cites this paper as [33]
The 2.5 AI/ML Layer section says anomaly-detection models learn normal operational patterns and flag deviations. The Edge Gateway and Communication Layer and Secure Telemetry Ingestion sections discuss communications, but do not describe transport-layer anomaly detection; nor does the paper link detected deviations to device compromise, unauthorized firmware modification, or replay injection. No page numbers were provided. (LLM, conf=0.82)
Experimental — accuracy not yet measuredPhysics-informed distributional baselines enable high-confidence deviation detection with low false-positive rates in ICS environments with highly periodic telemetry patterns and strong process-physics constraints.
“The anomaly detection approach draws on domain-validated techniques for ICS environments with highly periodic telemetry patterns and strong process-physics constraints, where physics-informed distributional baselines enable high-confidence deviation detection with low false-positive rates.[33]”
From Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems (2026) · cites this paper as [33]
The Background and §2.5 AI/ML Layer (pages not provided) describe learning normal HVAC and water-heating behavior from operational data, using domain-specific features, and detecting deviations. They do not establish physics-informed distributional baselines, highly periodic ICS telemetry, high-confidence detection, or low false-positive rates. (LLM, conf=0.85)
Experimental — accuracy not yet measuredContinuous monitoring of model output distributions across time identifies statistical shifts that may indicate data poisoning or model degradation.
“Third, continuous monitoring of model output distributions across time identifies statistical shifts that may indicate data poisoning where the model performs as designed but on corrupted inputs or model degradation.[17], [30], [33]”
From Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems (2026) · cites this paper as [33]
The AI/ML Layer describes continuous monitoring and detecting temporal anomalies, while the Model Drift section notes that system behavior can change over time (page numbers unavailable). The paper does not say that monitoring model output distributions identifies statistical shifts, or link such shifts to data poisoning or corrupted inputs. (LLM, conf=0.83)
Experimental — accuracy not yet measuredProcess-physics-informed distributional baselines substantially outperform purely statistical approaches in distinguishing legitimate process variation from injected anomalies in HVAC and water heating IoT deployments.
“Domain-specific validation of AI anomaly detection for HVAC and water heating IoT deployments confirms that process-physics-informed distributional baselines substantially outperform purely statistical approaches in distinguishing legitimate process variation from injected anomalies.[33]”
From Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems (2026) · cites this paper as [33]
The Methodology section describes domain-specific HVAC and water-heating features and learning normal operational signatures; the Case Study section reports anomaly detection in a smart-building deployment. Neither section reports process-physics-informed distributional baselines, injected anomalies, or a comparison showing substantial improvement over purely statistical approaches. Pages are not available in our copy. (LLM, conf=0.84)
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Still being checked.
Research works printed but not linked: 6
- IEEE IoT Journal, "AI in HVAC fault detection and diagnosis: A systematic review," 2023.
- S. Katipamula and M. Brambley, "Methods for Fault Detection, Diagnostics, and Prognostics for Building Systems --A Review," HVAC&R Research, 2005.
- F.Zhang, N.Saeed, P.Sadeghian, "Deep learning in fault detection and diagnosis of building HVAC systems: A systematic review with meta-analysis", 2023
- M. A. Khan and K. Salah, "IoT Security: Review, Blockchain Solutions, and Open Challenges," Future Generation Computer Systems, 2018.
- F. T. Liu et al., "Isolation Forest," IEEE ICDM, 2008.
- P. Malhotra et al., "LSTM-based Encoder-Decoder for Multi-sensor Anomaly Detection," arXiv:1607.00148, 2016. · arXiv
Cited by
Sorted by citations
Zero Trust Architecture for Smart Factories: Securing Digital Twins and Cyber-Physical OT Systems
Alagappan, V. (2026) · International Journal of Artificial Intelligence, Data Science, and Machine Learning · doi:10.63282/3050-9262.ijaidsml-v7i2p109
- Authors
- -, V. A.
- Venue
- Journal of Advances in Developmental Research
- Published
- 2025-12-13
- Volume
- 16
- Issue
- 2
- ISSN
- 0976-4844 (ISSN-L)
- OpenAlex
- W7125271824
- Cited by
- 0
- References
- 0
- FWCI
- 0.00
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by-sa
- Locations
- Journal of Advances in Developmental Research (publishedVersion, cc-by-sa)
- Topics
- Physical Sciences › Engineering › Control and Systems Engineering › Fault Detection and Control Systems
- Physical Sciences › Engineering › Building and Construction › Building Energy and Comfort Optimization
- Physical Sciences › Computer Science › Artificial Intelligence › Anomaly Detection Techniques and Applications
- Keywords
- anomaly detection, HVAC systems, hot water systems, Internet of Things, building energy efficiency, fault prediction, predictive maintenance, real-time monitoring, sensor data, occupant comfort
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
