Process Raman spectroscopy / Data integration
From instrument evidence to connected records.
Connect Raman Rxn batch spectra and model predictions to bioreactor and process records. Seal connects the exported evidence to the experiments, samples and review workflows your team already uses.
- ExportThe vendor software saves the measurement or evaluated result.
- CollectSeal IoT on the site computer uploads the completed file.
- MapA Seal Script validates the data and writes your chosen fields.
Instrument acquisition, analysis and control stay in the vendor software. This connection uses the existing Seal file and entity APIs. Configure the collection and mapping for your installed software and export layout; selecting the instrument in the directory does not install a parser or start acquisition.
What evidence can Endress+Hauser Raman RunTime provide?
- Source
- Endress+Hauser / Endress+Hauser Raman RunTime
- Evidence to collect
- Raman RunTime batch spectra in SPC format and model results in CSV.
Export options depend on the installed software and licences. The Endress+Hauser Raman RunTime setup guide covers source preparation and collection in Seal.
File collection follows exported batches. It is not a live control loop or automatic historian backfill. Agree interval granularity and measure volume before enabling frequent exports.
Keep one representative export and its matching instrument report. Record the software version, licensed export options, chosen columns or sheets, and which results the file actually contains. Check the layout after a software or report-template change.
Vendor reference: Endress+Hauser Raman RunTime 6.5 operating instructions. Product pages describe the source software; verify detailed export options against the manual for your installed version.
Choose the records you need.
One batch or defined sampling interval with selected model predictions, linked to the original spectral files.
These are suggested fields, not a required schema. Reuse your existing sample, experiment, equipment and method references. Preserve complete exports as File content; store selected values as typed fields when people need to filter, compare, calculate or review them.
| Information | Seal fields | Mapping rule |
|---|---|---|
| Batch and probe | Reference and text fields | Resolve vessel, probe channel and collection timestamp with timezone. |
| Model prediction | Number and unit fields | Name the model output and retain prediction validity. |
| Model and spectra | File and reference fields | Link model revision and source spectra for the same interval. |
| Source and revision | File and text fields | Retain the original export, instrument identity and analysis revision. Recognize an identical source revision before creating another result. |
Context to retain: Batch and probe; Model prediction; Model and spectra; Source and revision. A missing, failed or qualified result should stay distinguishable from a valid zero. Require explicit sample identities and reject unexpected layouts before the Script writes results.
Ready to configure the connection?
The platform docs contain the setup steps: prepare the source export, choose your System and template, configure collection, and test a mapping Script against your actual fields.
Follow the Endress+Hauser Raman RunTime setup guide ↗
Start with one representative export and its matching vendor report. The on-site equipment docs cover IoT installation, credentials and the first upload. The instrument mapping docs explain File content, typed records and ordinary Seal Scripts.
The site computer connects to Seal over outbound HTTPS. A cloud Script cannot read the instrument’s local drive. Collection frequency follows the export workflow: a folder watcher observes saved files and does not request new measurements.
Keep the source files until delivery and mapping are verified. For frequent or large datasets, choose useful summaries or intervals, retain coverage information and measure the intended volume before unattended operation.
Use the evidence in your existing workflow.
Imported records use normal Seal permissions and references. Compare typed results in views and formulas, link the analysis to a procedure, and use @ references to give neil the accessible samples, experiments and source evidence needed for a question.
Review the imported Endress+Hauser Raman RunTime results for these samples. A predicted analyte value is a model output, not a direct reference assay. Show the source evidence and any missing context.
neil can help write the mapping Script from a representative file and the actual fields in your System. Ask for a proposed mapping and a dry run first. Review its sample matching and scientific assumptions before allowing writes. An instrument name alone does not define a file layout or an analysis method.
Define which fields the import owns. Preserve user-entered notes, review decisions and formula-controlled values. Keep source acquisition time, analysis revision and import time separate so later reanalysis remains understandable.
Prove the workflow with representative data.
A useful first acceptance check
Import a batch spanning a model change and an invalid prediction. Verify the change boundary and that an interval summary reports missing or invalid coverage.
- Compare the result, not only the upload. Check the destination sample, selected values, units, qualifiers and linked File against the vendor report.
- Replay identical input. Verify the Script recognizes the source identity and revision before creating another result. File-upload calls alone do not prevent duplicate scientific records.
- Reprocess the analysis. Export a changed evaluation and check that it retains its prior evidence. Decide explicitly whether this creates a linked analysis or updates allowed fields under your normal review rules.
- Interrupt and recover. With disposable data, test a network interruption and restart. Also test failure after upload but before mapping completes. Confirm retained files, visible failures and correct retry behavior for the installed IoT app and Script.
Reconcile the source run or sample list with the imported analyses. A watcher cannot recover measurements the instrument never exported. Retain the checks and their evidence with the configured connection before relying on it for your intended use.
When the result is missing or looks wrong.
- No file arrives
- Confirm the instrument created a completed export in the watched folder. Check the IoT account’s file access, upload logs, destination System and template permissions.
- The File exists, but result fields are empty
- Check whether the mapping Script ran and recognized the export. Resolve missing sample identities or fields explicitly; uploading bytes does not populate an analysis automatically.
- The values differ from the instrument report
- A predicted analyte value is a model output, not a direct reference assay. Preserve model version, prediction status and probe identity; do not hide invalid predictions in averages. Compare the exact export, report and evaluation revision using the field rules above.
- Results are repeated or an analysis is missing
- Check export selection, source identities and revision matching. Reconcile at the analysis level; a filename or successful network response alone cannot establish complete coverage.