SCADA & HMI
Historian Concepts
Time-series storage: compression, retention, and retrieval for process data.
Overview
A historian is the plant's time-series memory: long-horizon storage of process values, tuned by compression and retention so years of data stay queryable.
Engineering purpose
Trend analysis, incident reconstruction, regulatory evidence, and the raw material for predictive analytics.
How it works
Swinging-door (deadband) compression stores only significant changes per signal; tiered retention keeps raw data short-term and aggregates long-term; interpolated reads reconstruct values between stored points.
- Deadband (swinging-door) compression stores only significant changes — set it per signal dynamics, not globally.
- Resolution and retention trade off: raw short-term, aggregated long-term tiers keep storage bounded.
- Interpolated reads differ from raw reads — trending tools must state which they return.
Common faults
Global compression settings flattening fast dynamics; archive queues backing up when ingest outruns storage; clock skew between sources corrupting event sequence; interpolated reads mistaken for raw evidence.
Diagnostic checks
- 1Compare a raw value against the historian value to validate compression settings.
- 2Check archive backlog/queue size — growth means the store can't keep up with ingest.
- 3Compare a live value against its historian readback to confirm compression is not erasing the dynamics you care about.
Safety notes
Incident investigations need raw, time-synchronized data — verify NTP discipline across sources before you ever need to reconstruct an event.
Commissioning notes
Set compression per signal class, size storage from measured ingest, and test a restore from archive before declaring the system production-ready.
Related concepts
Time-series compression, retention tiers, interpolation vs raw reads, NTP synchronization, trend retrieval.
historian · time series · trend · archive · compression · هیستورین · سری زمانی · ترند · آرشیو
Related articles
SCADA Tag Architecture
Tag naming, polling groups, and data quality in supervisory systems.
Alarm Management
Alarm rationalization, priorities, and flood prevention.
HMI Screen Design
Operator display hierarchy, alarm visibility, and bilingual layouts.
When Hermes Brain uses this article
Cited for questions about trends, missing or flattened history, archive growth, and time-series evidence quality.
…