What the data for this category looks like
Data for petroleum refining scenarios comes from DCS/APC control systems, LIMS laboratory information management systems, tank farm metering systems, crude oil evaluation reports, equipment management systems, and energy management systems. Daily documents include operating procedures, production plans, material balance sheets, distillate outlet quality reports, interlock ledgers, shift handover records, and change management documents. Data update frequencies cover real-time, per shift, and daily. Fields include unit load, distillation range, flash point, density, and more. Corresponding units include tons, cubic meters, degrees Celsius, and others.
Constraints during deployment and upgrade
Multiple data sources and staggered update schedules in petroleum refining require adaptive parameters for multi-system integration to be configured during deployment. This ensures real-time data synchronization delay meets process fluctuation response requirements. Structural differences across multiple document types require adaptive rules for document parsing to be adjusted during upgrades. This covers business files in different formats such as operating procedures and interlock ledgers. Coupled business logic across multiple units requires retrieval configurations to support associated recall of cross-unit data. This avoids siloed quality audit results. Large volumes of real-time process data writes require database parameters to match the load requirements of high-concurrency writes.
How to set configurations
| Config Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 connections | Matches concurrent connection requirements for multiple DCS and LIMS systems, and adapts to real-time data writes in refining scenarios |
REDIS_URL | Local deployment instance, memory configuration no less than 8 GB | Caches real-time process data and retrieval sessions, enabling fast access to shift handover records updated per shift |
STORAGE_S3_ENDPOINT | Internal enterprise object storage node address | Stores large-volume business documents such as crude oil evaluation reports and interlock ledgers, and supports batch file uploads updated daily |
MONGODB_URI | Cluster address with read-write separation configuration | Stores structured business data such as material balance sheets and distillate outlet quality reports, supporting high-concurrency retrieval requirements |
FE_DOMAIN | Internal enterprise access domain name | Adapts to access restrictions of internal office networks, ensuring secure access by process departments |
FILE_TOKEN_KEY | Key bound to the enterprise internal document permission system | Prevents unauthorized access to proprietary business documents in refining scenarios |
Three common errors
- Docker-pulled reranker calls succeed but return false: The interface returns a 200 status code but the result field is false. The cause is that permission verification rules for
FILE_TOKEN_KEYand business documents are not configured. This prevents the reranker from reading vector data of proprietary documents in the refining scenario. - After upgrading to FastGPT v4.17.0, the plugin edit page prompts unsaved changes: The top-right corner of the plugin edit interface continuously displays an unsaved changes prompt, regardless of whether the save button is clicked. The cause is an exception in synchronization between locally cached configurations and the database. Permission verification configurations related to
ROOT_KEYare not correctly synchronized. - Custom retrieval count cannot be adjusted above 6: When setting the retrieval count to 6 via API calls, the setting does not take effect. Returned results always use a fixed number of entries. The cause is that
DB_MAX_LINKand retrieval concurrency parameters are not correctly configured. This causes the system to limit the maximum recall count for a single retrieval.
How to confirm the configuration is correct
- View the database connection monitoring panel, confirm that the current connection count falls within the range specified by the
DB_MAX_LINKconfiguration. - Upload a distillate outlet quality report, check that the file storage path matches the address configured in
STORAGE_S3_ENDPOINT. - Initiate a retrieval request based on crude oil properties, check that the number of returned results meets the requirements of the business scenario, and confirm that the retrieval configuration is effective.
- Check the network request logs on the frontend page, confirm that the domain name configured in
FE_DOMAINcan normally receive requests and return business data.
The values provided are common starting points and should be measured against the reader's own samples.