Distributed Energy Resource Analytics & Grid Observability Platform
Primary service
AI-ready Data, Governance & Observability
Secondary service
Real-time Monitoring & Anomaly Detection
Proposed Cadvel capability — not a client case study.
Target client and segment
Distribution network operators and utilities with growing DER portfolios
Challenge
More distributed generation and flexible demand make a partial network view harder to interpret. Operators and planners need to distinguish a real operating anomaly from missing, delayed or incorrectly mapped data.
Key capabilities
- A network-context view of DER observations with explicit freshness and coverage.
- Prioritized investigation of unusual operating patterns and data-quality failures.
- Governed historical datasets for forecasting and model evaluation.
Business Challenge
Coverage is uneven
Some assets stream readings while others report at longer intervals or not at all. An aggregate should expose what it actually measures.
Changing topology distorts analysis
DER readings need the network relationship valid at the time. A stale feeder assignment can create a misleading local trend.
Alerts lack supporting evidence
An unusual load pattern may reflect weather, planned switching or a broken feed. Operators need these contexts together to investigate.
Models inherit untested data
Forecasting features built from late corrections or future information can look accurate offline and fail in operation. Provenance and time-aware validation matter.
Architecture and Tech Stack
Approved OT exports, meter and DER feeds → timestamp and quality validation → versioned asset and topology model → coverage-aware network aggregates → investigation dashboards and governed feature datasets. Preserve both event time and data-arrival time.
Technology options
Kafka for streaming intake; ClickHouse for operational investigations; Snowflake and dbt for governed historical features where needed; Datadog for platform health. Select one or both analytical stores only after workload and cost testing.
Delivery boundary
This is decision-support observability, not electrical state estimation or protection. Any inferred network state or forecasting model needs separate engineering validation, with measured and estimated values clearly distinguished.
Implementation and Solution
Proposed implementation stages
Define decisions and coverage
Choose feeder-level questions, available observations and data owners; document blind spots.
Build network context
Validate timestamps and effective-dated topology links; publish coverage and freshness measures.
Test alerts and features
Replay known incidents; compare contextual alerts and validate historical features without future-data leakage.
Pilot with operators
Run alongside existing tools, measure investigation usefulness and establish model and data-change governance.
What this enables
Measurable success criteria — to be agreed and validated with the client; not achieved results.
Observable asset coverage
In-scope DER assets with valid, timely data and topology links ÷ registered in-scope DER assets.
Investigation time
Median minutes from a reviewed alert to a documented cause or disposition.
Alert precision
Confirmed actionable grid or data issues ÷ reviewed alerts, reported separately by issue type.
Governed feature coverage
Production features with owner, lineage, quality checks and freshness objective ÷ production features.
KPI targets and acceptance tolerances are set with the client during discovery; no achieved outcomes or ROI are asserted.
Source inspiration / reference
Source-derived facts
DXC’s Endeavour Energy story mentions analytics and DER initiatives involving solar, batteries, smart meters and EVs; the delivered engagement described is a data-center move. DXC’s renewable energy offering discusses forecasting and grid analytics. Neither source documents the specific observability platform proposed here.
Reference context only. Proposed Cadvel design and KPIs are independent of the source delivery.
