What the data for this category looks like
Core users of pulp and paper enterprises are typically the stock preparation, pulping, paper making, quality control, and energy departments. Daily technical documents include pulp mixing sheets, beating records, paper machine operation records, online quality curves, laboratory reports, and sheet break and product change records. Data sources include wood pulp and waste paper batch ledgers, DCS/QCS systems, paper machine online measurement equipment, laboratory testing systems, and energy and water treatment systems. Data update frequencies include three rhythms: real-time, per shift, and per roll. Document fields often use industry-specific units, such as °SR for beating degree, g/㎡ for paper grammage, and % for slurry moisture content.
Constraints imposed on citation sources and traceability
The high-frequency updates of real-time data require the traceability function to support quick retrieval of the latest online measurement and operation records, to avoid information lag. Dispersed data from multiple sources requires cross-system correlation verification. The traceability configuration must match the exclusive fields of different documents, to ensure that batch information in pulp mixing sheets and paper machine operation records can be accurately matched. The high-speed continuous web production feature creates a high demand for correlating sheet break and product change records, requiring quick location of production data for the corresponding batch. Differences in units across different documents require that original fields and units be retained during traceability display, to avoid errors in cross-document comparison.
How to set the configuration
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | Adapts to the average length of a single technical document in pulp and paper, to avoid truncation of key traceability fields and batch information |
responseData | Enable full source display | Meets the multi-document correlation needs of cross-departmental reviews, and clearly displays the data source system and collection time |
detail | Retain original fields and units | Retains industry-specific fields such as beating degree °SR and paper grammage g/㎡, to ensure information accuracy during reviews |
| Number of displayed citation sources | Top 3 | Matches the upper limit of document quantity for single-round technical reviews, to avoid information overload that affects judgment |
responseData timeout threshold | 600 seconds | Adapts to the retrieval delay of real-time online measurement data, to ensure complete acquisition of the latest paper machine operation records |
Three common errors
- Phenomenon: When sharing the application without logging in, the citation source button is displayed, but clicking it fails to jump to the source data. Cause: The traceability permission for public sharing is not enabled in the
responseDataconfiguration, so the sharing link cannot call the internal source data interface. - Phenomenon: No citation identifier or field information is displayed at the end of the knowledge base response. Cause: The full field display of the
detailconfiguration is not enabled, or thequoteMaxTokenvalue is too small, truncating the citation marker and source data information. - Phenomenon: After two consecutive rounds of questioning, traceability information for historical documents cannot be correlated. Cause:
quoteMaxTokenis not adjusted to a length suitable for multi-round conversations, so historical traceability data is truncated, making context correlation impossible.
How to confirm the configuration is properly set
- Upload a single pulp mixing sheet document, initiate a technical review request, and confirm that the document name and original field information of the citation source are displayed at the end of the response.
- Generate a non-logged-in sharing link, click the citation source button, and confirm that it jumps to the source data details page.
- Initiate two consecutive rounds of questioning, and confirm that the second round of response can correlate the traceability data from the first round of review.
- Simulate a real-time data update scenario, upload the latest paper machine operation record, and confirm that the traceability information synchronously displays the latest collection time and parameters.
The values provided are common starting points, and should be measured against the reader's own samples.
This document is adapted for FastGPT v4.17.0.