| Technical Capability | Aabcor Platinum | TaKaDu Gold | Gutermann Gold | Innovyze (Info360) Gold |
|---|---|---|---|---|
| Multi‑Source Data Fusion | ✓★★★★★Acoustic + hydraulic sensor fusion (e.g., sound, flow, pressure) with simulated data integration. | ✓★★★★☆SCADA and multiple data inputs unified via central event management. | ✓★★★☆☆Acoustic‑focused (correlating logger networks); integrates external data via partner platforms. | ✓★★★★☆Digital twin combines SCADA with hydraulic models for holistic network view. |
| Modeling Sophistication | ✓★★★★★Deep learning (CNN for acoustics, GNN for network) plus meta‑learner fusion with uncertainty quantification. | ✓★★★★☆Advanced statistical anomaly detection augmented by machine learning algorithms. | ✓★★★☆☆AI acoustic classifier (“AI Predictor”) trained on large leak sound databases for probability of leak. | ✓★★★★☆Library of analytics & modeling tools (forecasting, simulation) for incident detection and response. |
| Deployment Readiness | ✓★★★★★Cloud‑native container packaging (Helm/Kubernetes), API‑first design (OpenAPI), CI/CD with automated rollouts. | ✓★★★★☆Proven SaaS platform (cloud‑based); integration APIs for third‑party data and rapid deployment. | ✓★★★★☆IoT‑enabled hardware with cloud portal (ZONESCAN NET); scalable to tens of thousands of sensors. | ✓★★★★☆Web‑based cloud service (Autodesk Info360); on‑prem/cloud extensibility via Autodesk’s infrastructure. |
| Auditability & Evidence | ✓★★★★★Immutable WORM evidence bundles (signed manifests) per event; full decision trace for regulator audits. | ✓★★★★☆Comprehensive event tracking from detection to resolution (timestamped lifecycle management). | ✓★★★☆☆Leak event records with audio and correlation graphs for operator validation (audit trail). | ✓★★★★☆Incident management logs and compliance reports aligned with utility standards. |
| Explainability | ✓★★★★★Explainable AI outputs: Grad‑CAM spectrograms, SHAP feature importance, network leak heatmaps. | ✓★★★☆☆Contextual dashboards (trends, thresholds, maps) to explain anomalies (no explicit ML explainers). | ✓★★★☆☆Correlation visuals and “AI Predictor” score enabling listen/verify workflows. | ✓★★★★☆Visual overlays on the digital twin; scenario simulations for interpretation. |
| Risk & Threshold Management | ✓★★★★★Conformal risk control with per‑DMA calibrated thresholds; drift detection & auto‑retraining triggers. | ✓★★★☆☆Configurable alert thresholds and KPIs (tuned analytics; no formal conformal guarantees). | ✓★★★★☆Automated daily correlation to minimize false outcomes; confidence scoring for filtering. | ✓★★★★☆Rule‑based and ML alerting with adjustable sensitivity leveraging historical patterns. |
| Synthetic Data Augmentation | ✓★★★★★Extensive synthetic leaks and simulated hydraulics to supplement training (governed; no data leakage). | –Relies on real network data and historical incidents (no known synthetic generation). | –Relies on real acoustic samples; emphasizes field data over synthetic augmentation. | –Primarily actual sensor and model data; scenarios via digital twin rather than generated training data. |
Note: Company names and products are trademarks of their respective owners. This comparative view is technical and indicative; capabilities may vary by deployment.