What is OPC HDA?

A raw read returns exactly what the historian stored: the actual values at the exact timestamps they were recorded. If your historian recorded values at irregular intervals because it uses exception-based compression, a raw read reflects that irregularity. If the historian stored a value at 10:03:47 and the next at 10:04:12, those are the timestamps you get back.

An interpolated read returns values at the timestamps you specify, regardless of when the historian actually recorded them. If you ask for values every 60 seconds, you get one value per minute, estimated by interpolating between the nearest raw data points on either side. This is useful for generating consistent time-series data for trend displays or for normalizing tags recorded at different rates to a common time base.

Use raw reads when the actual recorded data matters, for example in compliance reporting or root cause investigation. Use interpolated reads when you need a regular time axis and estimated values are acceptable.

OPC HDA specifies that bad-quality values are excluded from aggregate calculations by default. If an interval contains a mix of good and bad values, the aggregate function is computed only from the good values. The result carries a quality code indicating that some values in the interval were excluded, so the client knows the aggregate was computed from a subset of the available data.

The treatment of uncertain-quality values is server-dependent. Each historian's OPC HDA server documentation should state how it handles uncertain values in aggregates. Some servers include them in calculations; others treat them as bad. For compliance-sensitive applications, it is worth verifying this behavior against the specific historian platform in use.

OPC HDA includes optional write-back interfaces that allow clients to insert, modify, or delete historical data values in the historian archive. These are optional interfaces: not all OPC HDA servers implement them, and write access typically requires specific authorization.

Write-back capability is used in scenarios such as inserting manually entered lab results into the process historian archive, correcting erroneous values caused by a known instrument fault, or loading historical data from one historian into another during a migration. It should be implemented carefully, as modifying historian records has compliance implications in regulated industries.

OPC HDA remains in active use across a large installed base of historian integrations. The vast majority of industrial plants running process historians have applications integrated via OPC HDA that are working reliably and have no operational reason to migrate.

OPC UA Part 11 (Historical Access) provides equivalent functionality over TCP/IP without DCOM, and is the correct choice for new integrations. But the existence of OPC UA historical access does not obsolete working OPC HDA integrations any more than OPC UA obsoleted working OPC DA connections. The two coexist in the same environment, and most major historian vendors expose both interfaces simultaneously to serve both legacy and modern clients.

TOP Server supports OPC HDA on the same installation as OPC DA, OPC UA, and OPC A&E. This means TOP Server can expose historical data to OPC HDA clients for any device it has a driver for. TOP Server's Local Historian plug-in collects and stores tag values locally, and the OPC HDA interface provides client access to that stored data.

This is particularly useful in scenarios where devices do not have their own historian, or where you want a standardized HDA interface to device-level data without deploying a full enterprise historian. It is also useful as a data collection point at the edge, buffering and storing data locally with HDA access, before forwarding aggregated or selected values to an enterprise historian upstream.

A simple average (the "Average" aggregate in OPC HDA) adds up all the raw values in the interval and divides by the number of values. Each value has equal weight regardless of how long it was held before the next change.

A time-weighted average (the "Time Average" aggregate) weights each value by how long it was held. If a flow rate was 100 units for 55 seconds and 200 units for 5 seconds, the simple average is 150. The time-weighted average is closer to 108, because the 100-unit value was held for most of the interval.

Time-weighted averages are more accurate for physical quantities that do not change instantaneously, which describes most process variables. They are required for utility billing calculations and certain environmental reporting standards. Simple averages are appropriate for counting-based variables or when all samples have equal weight regardless of duration.