FIELD GUIDE
Pre-Winter Assessment: District Heating Network Readiness
Structured pre-winter risk assessment framework for municipal district heating networks — prioritizing sections for inspection before the heating season.
← Rivixi HomeResearch Lab
Rivixi Research Lab
Applied AI. Published Proof.
Peer-reviewed methodologies, production case studies, and technical whitepapers from the team turning research into measurable ROI.
CASE STUDY
False Positive in Acoustic Leak Detection of a Flooded Heating Pipeline: A Case Study in Hybrid AI Diagnostics
When ensemble AI models and a third-party correlator all flag a leak on a flooded 336 m district heating line, hybrid PNR (8.64) and phase coherence (0.4%) produce the correct no-excavation verdict.
RESEARCH PAPER
Vibroacoustic Diagnostics of Degradation and Wall Thinning in Pressurized Pipelines: From Physical Wave Propagation Models to a Hybrid DSP-Deep Learning Pipeline
Field-validated RIVIXI Diagnose module maps corrosive wall thinning (10–18.5% loss) on a ø426×7 mm district heating line — converging with UT measurements within 3.5% while ruling out active leakage.
RESEARCH PAPER
Utilizing Zero-Crossing Rate (ZCR) for Acoustic Leak Detection in Pipelines: From Empirical Models to a Physically Grounded DSP Pipeline
How Zero-Crossing Rate (ZCR) verification in RIVIXI AI v1.3 distinguishes hydrodynamic leak hiss from impulsive mechanical false alarms — reducing FPR by 13.25% on 113 field recordings.
CASE STUDY
Overcoming Data Chaos: How Machine Learning Compensates for Inaccurate Utility Records
How Decision Fusion algorithms process messy utility maintenance logs and GIS records to achieve 93.88% Recall on pre-failure zone detection.
RESEARCH PAPER
Topological AI-Analysis vs. Classical Cross-Correlation: Overcoming Legacy Defectoscope Vulnerabilities
How modern neural network algorithms eliminate the pipeline false alarm scalability barrier inherent in the direct reengineering of classical non-destructive testing.
RESEARCH PAPER
Adapting an Ultrasonic Diagnostics AI Platform to Legacy Hardware: Dynamic DSP Pipeline Reengineering
How Rivixi Lab overcame Nyquist limit collisions and legacy sensor constraints to integrate SebaKMT and Kaskad-3 detectors into the RIVIXI AI v1.3 SaaS platform.
RESEARCH PAPER
Improving Pipeline Failure Prediction via Data Sanitization, Hyperparameter Optimization, and Boosting Blending Ensembles
How a multi-model blending ensemble combined with rigorous data sanitization improved district heating and water pipeline risk prediction, achieving a ROC-AUC of 0.8879.
RESEARCH PAPER
Seeing Sound: A Computer Vision Approach to Ultrasonic Leak Detection in Industrial Pipelines
How converting ultrasonic signals into Mel-Spectrograms and applying ensemble neural networks overcame the limitations of traditional heuristics in noisy industrial environments.
RESEARCH PAPER
AI vs. Metallography: Accelerating Outage Inspection in 2.25Cr-1Mo Steel Pipelines via MBN-RVX-AI
MBN-RVX-AI replaces slow replica metallography during outages with cloud AI screening—over 92% accuracy on 2.25Cr-1Mo steel microstructural degradation.
CASE STUDY
Autonomous AI Agents in Tech Support: The ProstorHelp Architecture
How moving from scripted chatbots to autonomous ReAct agents with Advanced RAG enables zero-hallucination tech support for complex enterprise software.
RESEARCH PAPER
Application of Hybrid ML Models for Pipeline Failure Prediction
A hybrid machine learning architecture combining static pipeline characteristics with temporal accident history to forecast failures in urban heat and water supply systems.