What data for this category looks like
This category’s data comes from daily operational documents of winding, insulation, final assembly, and testing departments at electric motor and generator factories. These documents include winding process cards, insulation treatment records, dynamic balance reports, voltage withstand insulation test records, and type and factory release test reports. Data sources include electromagnetic design data, copper wire and insulation material batch records, winding equipment, balancing machines, electrical test benches, and ERP/MES systems. Updates trigger when a product or production batch finishes, and after each product type change. Documents include professional fields with units: winding turns, paint dipping and drying parameters, dynamic balance deviation values, and withstand voltage test voltage. They also associate with material batches and equipment operation data.
What constraints these characteristics impose on deployment and upgrade
This category’s data is scattered across device logs, business systems, and paper documents. Deployments require adaptation to multi-source data access formats to avoid field parsing errors. Data update frequency triggers with production batches and product type changes. Upgrades require ensuring incremental synchronization logic is compatible with temporary document submission rhythms to avoid synchronization delays. Professional documents include specific fields with units such as winding turns and withstand voltage test voltage. Deployment phases require configuring field and unit retention rules to prevent semantic information loss during embedding. Test data involves security scope, so deployments require configuring layered data access permission rules to avoid sensitive data leakage.
How to set configurations
This configuration guide applies to FastGPT v4.17.0. Specific parameter configurations are listed below:
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–150 connections | Data for this category comes from multiple types of equipment and business systems, so a stable database connection pool is required to handle concurrent document query requests |
STORAGE_S3_ENDPOINT | Fill in according to the enterprise's own object storage address | Documents such as test reports and process cards for this category must be stored on the enterprise's own storage media to ensure data controllability |
REDIS_URL | Local Redis instance on port 6379 | High-frequency queried process parameters and test data must be cached to reduce database load |
| Single document upload limit | 1000 MB | Type test reports for this category may include multiple sets of test data and waveforms, so support for large-volume file storage and reading is required |
PG_URL | Fill in according to actual measurement and calibration | Business data for this category must be stored in a PostgreSQL database, so it is necessary to match the connection address and permissions of the enterprise's existing database |
FE_DOMAIN | Enterprise internal access domain name | When deployed on the enterprise intranet, configure the internal domain name to ensure normal internal access |
Three common errors
- Phenomenon: After deployment, calling the AI chat component returns an uncaught exception. Cause: In a private deployment scenario, configuration file parameters do not match the actual addresses of the enterprise database and storage, resulting in connection failure.
- Phenomenon: After upgrading the version, the internal browser cannot normally load the Q&A page. Cause: Compatible front-end resource loading rules are not configured for old-version browsers used on the enterprise intranet, resulting in failure to load styles or scripts.
- Phenomenon: Unable to access production data from the enterprise's own equipment. Cause: Access protocol adaptation rules for the corresponding data source are not configured, and the equipment data format is mistakenly converted to a general template, resulting in loss of professional field information such as winding turns and withstand voltage values.
How to confirm correct configuration
- Perform a database connection test, confirm that the configuration value of
DB_MAX_LINKmatches the current concurrent request volume, and adjust it so that the connection pool has no overflow or excessive idle connections. - Upload a typical winding process card and test report, confirm that the document can be stored and retrieved normally, and matches the configuration rules of
STORAGE_S3_ENDPOINT. - Initiate a process parameter query request, confirm that the returned result includes professional fields such as winding turns and withstand voltage test voltage, and that units are not lost.
- Check the configuration of the front-end access domain name
FE_DOMAIN, confirm that the Q&A page can be normally accessed in the internal network environment.
The values provided here are common starting points and should be evaluated against the reader's own samples.