Skip to main content

Report data lineage

An auditor who reads a reported value needs to know where the value came from. Data lineage answers that question inside the product. You click a value in a report. The product then shows each calculation step back to the per-minute records and the pipeline runs that wrote them. This page explains the structure of that chain. Trace a reported value in the user guide explains how to open and use the view.

Why lineage follows the calculation​

The lineage tree starts at the report cell and follows the calculation, not the database tables. A report cell often depends on more than one tag. For example, a NOx mass for a day needs the NOx concentration and the dry exhaust flow. The conversion mode of the report column decides which tags take part. The tree shows only those tags.

At each level, the tree shows the formula with the actual values in it. Thus you can check each step with a calculator.

The levels of the chain​

  1. Report cell. The value as the report shows it. This level shows the conversion formula and how the period is aggregated. A mass or a mass flow is a sum. A concentration is an average. A mass per heat input or per power output is a weighted average.
  2. VSTA, SSTA and STA. The emission tag always follows the order VSTA, SSTA, STA. The tree never skips the SSTA, because the VSTA is calculated from it. The oxygen STA is a child of the SSTA, because the oxygen correction uses it. A supporting tag, such as a flow or a heat input, is a plain STA directly below the report cell.
  3. STA intervals. A daily or monthly value contains many hourly intervals. A month has more than 700. You open the list and select the hour that you want to examine.
  4. First-level data records. Each hour contains the per-minute records of that hour.
  5. Pipeline run. Each per-minute record comes from one pipeline run. This level shows the steps of the run, their status, and the values that each step read and wrote.

The tree also shows the model output, the input data and the calibration that was active for the period.

If the report column reads a long-term average directly, the tree ends at the LTA node. The LTA is already aggregated, so the tree has no chain below it.

Why the tree is shaped like a diamond​

The tree first spreads out from the report cell to several tags. It then comes together again at the per-minute records. One pipeline run reads all instruments of an emission source at the same time. It writes one record that contains every tag for that minute. The STA of NOx, the STA of oxygen and the STA of the flow all read the same records.

This shared record is useful in an investigation. When the NOx value of one minute is substituted, the same record shows if oxygen and flow were good in that minute. A problem with one instrument then looks different from a general loss of connection.

Each emission source has its own records. A report cell for an asset with several sources therefore has a separate set of records for each source.

How quality moves up the chain​

Each level has a quality. A bad or substituted value at the minute level affects the STA of its hour, then the SSTA and the VSTA of that hour. The tree shows a quality summary at each level, for example the number of good and substituted minutes in an hour. You can follow a flag from the report cell down to the minute where it started.

References to EN 17255​

Each aggregation level refers to the section of EN 17255-1 that applies to it. These are section 8.3 for the STA, section 8.7 for the SSTA and section 8.11 for the VSTA. These references help an inspector compare the product with the standard. The product does not show the text of the standard.

Limits of the chain​

Lineage shows what the product stored. Some records hold less detail:

  • Some pipeline runs keep their data in memory, for example some batch runs. These runs do not store the input and output values of each step. The pipeline run level then shows the steps without these values.
  • If the product cannot find the pipeline run for a record, the level shows the run as unavailable.
  • The report cell shows the stored values. It does not calculate them again. See Why a report reads stored values.