Process historian

A process historian is a time-series database that collects, compresses and stores data from plant control systems, such as sensor readings, setpoints and alarms, so it can be trended, analyzed and retrieved later.

A historian connects to the DCS, PLCs or SCADA system and records values by tag. A tag is one measured or calculated point, such as a pump discharge pressure. Each value is stored with a timestamp and often a quality flag. Collection runs continuously, often at one-second resolution or faster.

Plants generate large volumes of such data, so historians compress it. A common method stores a new value only when it deviates from the previous trend by more than a set band. Retrieval tools interpolate between stored points and compute aggregates like hourly averages. Access is typically through OPC HDA, SQL or REST interfaces. Some sites run a local historian per unit and an enterprise historian that aggregates selected tags for the business.

Two limits matter for analytics. First, compression settings chosen years ago may have discarded the small variations a model needs. Second, historians store measurements, not context: which product was running, which batch, which shift. That context lives in MES, LIMS or paper logs and must be joined in. Historians are the usual training data source for soft sensors, anomaly detection and predictive maintenance models.

Related terms