What data for this category looks like
Non-ferrous metal smelting and electrolysis production ledger data sources include concentrate and auxiliary material analysis reports, tank voltage and current data collected by the DCS system, composition test results issued by the laboratory, energy and environmental protection monitoring records. Daily documents include batching lists, smelting records, electrolytic cell condition records, anode and cathode quality records, environmental protection ledgers, and more. Production teams generate structured records for single-batch raw materials or single-furnace production. Data update frequencies cover real-time (cell status), per furnace/per tank (smelting/electrolysis records), and per batch of raw materials (batching and laboratory reports). Document fields include raw material composition parameters, production process parameters, quality inspection indicators, and more. There is no unified standardized format, and teams must split and map fields according to production links.
What constraints these characteristics impose on deployment and upgrade
Multiple heterogeneous data sources and varied update frequencies require teams to adapt data source access for different synchronization frequencies during deployment. Teams must also support newly added laboratory or environmental protection monitoring data formats during upgrades. Professionally segmented document fields and quality indicators require custom field mapping rules to avoid data loss or misalignment from general configurations. Real-time cell status data with high real-time needs requires teams to configure low-latency caching and synchronization mechanisms. Batch ledger summary tasks require teams to adjust concurrent connection counts. Errors may lead to reduced metal recovery rates or environmental protection over-standard levels, so teams must complete data verification and rollback plan configuration before upgrades.
How to set configurations
This chapter’s configuration applies to FastGPT v4.17.0.
| Configuration Item | Recommended Value | Rationale |
|---|---|---|
DB_MAX_LINK | 15–25 个 | Adapt to the concurrent requirements of multi-source real-time data pulling, avoid database connection exhaustion |
REDIS_URL | redis://default:custom password@internal Redis address:6379/0 | Store real-time electrolytic cell status and temporary ledger cache, match real-time update frequency |
STORAGE_S3_ENDPOINT | internal enterprise object storage service address | Store batch-uploaded smelting and laboratory original documents, adapt to multi-type file access |
FILE_TOKEN_KEY | secret string generated according to FastGPT deployment documentation | Verify ledger file upload permissions, prevent unauthorized data access |
ROOT_KEY | random string with complexity no less than 16 characters | Highest permission configuration for the management platform, adapt to enterprise-level deployment security requirements |
Three common errors
- Symptom: After importing JSON format knowledge base configuration, the optional knowledge base fields are empty, and the target library cannot be directly selected. Cause: Failed to pull file metadata due to incorrect configuration of
FILE_TOKEN_KEY, and preset ledger classification tags are not associated. - Symptom: 401 unauthorized or connection timeout error returned when pulling the Reranker model. Cause: Incorrect configuration of model storage related keys, or the ACCESS Token is not bound to the enterprise internal model warehouse permissions.
- Symptom: Data duplication or missing partial fields appear after multi-furnace ledger summary tasks are executed. Cause:
DB_MAX_LINKvalue is too high, causing database connection conflicts, or the synchronization interval does not match the single-furnace production cycle.
How to confirm the configuration is correct
- Upload batches of documents including batching lists and smelting records, check if the file upload status shows success, and if metadata includes correct batch numbers and furnace numbers.
- Call the real-time data pulling interface, verify if DCS-collected tank voltage and current data synchronizes to the platform database at the set interval.
- Log in to the management backend with the highest permission key, confirm all configuration items are saved and have no format errors.
- Trigger a single-furnace ledger summary task, check if the generated report includes all preset field data for the corresponding batch.
The values provided are common starting points and should be evaluated against the reader’s own samples.