The Measurement Data Health Check is a paid, fixed-scope assessment by Akan Industrial and Automation that shows a process plant which of its measurements the records support, which show problems, and what to fix first. It covers one site and works on exports that plant staff produce from records the plant already has: trend data, validation and calibration records, maintenance logs and laboratory results. There is no connection to plant systems. The result is a written report with figures for each measurement in the agreed scope, findings ranked by effect and recommended actions. It is the usual starting point for plant measurement data in Akan’s Industrial AI and Software practice.
Why check measurement data?
An analyzer can keep delivering a value while its sampling system is partly blocked or its calibration has drifted. The historian stores the number either way. Decisions on process control, product quality and environmental monitoring may then rest on values whose condition has not been quantified. The Health Check puts figures on the condition of each measurement in the agreed scope.
Who is it for?
The assessment is for plants that run online process analyzers, continuous emission monitoring systems (SEÖS), continuous wastewater monitoring systems (SAİS) or flow metering systems. It also suits plants that compare analyzer readings with laboratory results. The equipment can be of any make, including analyzers and systems that Akan did not supply.
It fits when the plant asks questions such as:
- Which analyzer readings can operators and process engineers rely on?
- Why do analyzer and laboratory results disagree?
- Which measurements lose the most data, and why?
What does the assessment examine?
| Review | What it shows |
|---|---|
| Availability | For each measurement, the share of the review period in which a value was delivered |
| Validity | The share of delivered values that are valid; frozen values, out-of-range values and values recorded during maintenance or calibration are separated out |
| Drift and step changes | Slow shifts and sudden jumps in the trend, set against maintenance and calibration records |
| Control charts | Whether validation results are stable or show a systematic trend over time |
| Analyzer and laboratory comparison | Bias and scatter between analyzer readings and laboratory results for the same sample point |
| Data access | Where the data is kept, in which system and format, and how it can be reached |
What data do we need?
The assessment uses records the plant already keeps.
| Data | Typical source |
|---|---|
| Trend data | Export from the historian or data acquisition system |
| Validation and calibration records | Analyzer logs, maintenance system or spreadsheets |
| Maintenance logs | Maintenance system or logbooks |
| Laboratory results | Laboratory system or spreadsheets |
Gaps in the records are themselves reported as findings.
How does it work?
- Scope. The site, the measurements, the review period and the exports are agreed in writing.
- Data handover. Plant staff produce the exports. Akan does not connect to the control network.
- Analysis. Akan’s engineers carry out the reviews listed above for each measurement in the agreed scope.
- Report and review. The findings are presented to the plant’s engineers and discussed measurement by measurement.
The assessment typically takes three to four weeks.
What do you receive?
- A written report with the figures for each measurement in the agreed scope
- Findings ranked by effect, with recommended actions for maintenance, calibration, the sampling system and data handling
- A data-access map
- A proposal for a pilot with acceptance criteria, if the findings justify one
Every finding states the data and records it is based on. The findings are advisory: the plant’s engineers review them and decide what to act on.
Where does the data stay?
No software or hardware is installed on the control network. Nothing writes to the PLC or DCS, and no setpoint is changed. The exports are analysed on the client’s premises or, if the client agrees in writing, on Akan’s computers. No data leaves the site without that agreement. Exports and working copies are returned or deleted when the assessment ends. Where the records contain personal data, a data-processing agreement is signed. Details are on the Method, Data and Security page.
What is outside the scope?
The Health Check is not accredited laboratory testing and it is not a regulatory audit. It does not filter, alter or delay data that is reported to an authority.
What comes next?
Akan’s Industrial AI and Software work follows four steps: Assess, Pilot, Deploy, Support. For measurement data, the Health Check is the Assess step. If the findings justify continuous checking of the data, the next step is a pilot of Analyzer and Monitoring Data Analytics with acceptance criteria agreed in advance. In a pilot the software runs on the client’s premises over a read-only connection. Findings on the analyzer or its sampling system can be handled by the plant’s own team or with Akan’s analytical instrumentation engineers.
