What the Data for This Category Looks Like
Data sources for inspection and process record audit for photovoltaic module encapsulation include battery sheet grading systems, glass adhesive film backplane batch ledgers, factory MES systems, string welding and lamination production equipment, EL/IV testers, and others. One test record is generated per module, and one batch ledger is generated per material batch. Daily documents include layout drawings, material batch records, lamination process records, EL defect images, IV power test reports, insulation voltage resistance records, packing lists, and others. Data fields include battery sheet current gear, lamination temperature and duration, EL crack location and area, IV test power value, and others, with corresponding units: ampere, degree Celsius, second, watt, megohm, and others.
Constraints on Deployment and Upgrade
Multiple heterogeneous data sources require adapting docking protocols for different systems during deployment, and pre-configuring data import preprocessing rules. High-frequency update cycles require configuring database connection pools and cache services to support high-frequency writes and queries, preventing performance bottlenecks. Diverse document types require file storage services to support large-volume storage of different formats, including drawings and test reports. Precise processing of core time-consuming links requires vector retrieval services to support parsing of long-text and image data, with corresponding parsing configurations updated during upgrades.
Configuration Settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
DB_MAX_LINK | 100–200 | Adapts to high-frequency database write requirements per module, avoids service interruptions caused by exhausted connections |
MONGODB_URI | Points to the replica set address mounting photovoltaic module business data | Stores bulk unstructured data such as EL images and IV test reports, replica set deployment ensures data high availability |
REDIS_URL | Configured as local cache instance address, set key expiration time to 1800 seconds | Caches frequently queried battery sheet current gear matching rules, reduces query pressure on backend databases |
STORAGE_S3_ENDPOINT | Points to the access address of the enterprise object storage service | Stores large-volume documents such as layout drawings and packing lists, adapts to storage and retrieval needs of multiple file types |
ROOT_KEY | A random 32-bit string is generated according to enterprise internal permission specifications | Configures administrator access permissions, ensures system access security after deployment |
FE_DOMAIN | Enter the server public IP or enterprise intranet domain name | Configures front-end access address, adapts to internal access scenarios of departments such as string welding, lamination, and testing |
Three Common Misconfigurations
- After starting containers, the front-end page cannot be accessed, returning 502 or timeout errors. The cause is that
FE_DOMAINis not correctly configured as the server public IP or domain name, and inbound access rules for port 3000 are not enabled. - After uploading EL images or IV test reports, corresponding files cannot be retrieved in the knowledge base. The cause is that
STORAGE_S3_ENDPOINTis not correctly configured to point to an available object storage service, resulting in failed normal storage of files. - No return result occurs when calling external AI channels, with the interface reporting 401 errors or request timeouts. The cause is that channel secret parameters are not correctly configured, or expired images are used, and the latest version adapted to FastGPT v4.17.0 has not been re-pulled.
How to Verify Successful Configuration
- Execute the
docker pscommand, confirm that the FastGPT container and dependent MongoDB and Redis containers are running. - Access the configured
FE_DOMAINaddress, verify that the front-end page loads normally and administrator login is completed. - Upload a small photovoltaic module EL image test file, verify that the file uploads normally and the corresponding entry appears in the knowledge base.
- Batch import 10 simulated module test data, observe database connection pool monitoring indicators, and confirm that no connection exhaustion alarms are triggered.
The values provided are common starting points and should be evaluated against the reader's own samples.