What the data for this category looks like
The data for this category covers production and testing information across all processes of crystalline silicon solar cells. Daily documents include process cards, diffusion and coating parameters, printing and sintering records, EL inspection images, IV grading data, and yield reports. Data sources include silicon wafer incoming materials, MES systems, production equipment such as diffusion furnaces, coating equipment, and printing presses, EL and IV testing equipment, and laboratory data. Update frequencies cover per-piece testing, per-batch output, and daily summaries. Core fields include batch number, wafer ID, process parameter values, defect coordinates, conversion efficiency, and more. Data types include structured parameters, images, and report files.
What constraints these characteristics impose on the database and operations maintenance link
Multi-source heterogeneous data sources require the database to support mixed storage of structured parameters, images, reports, and other data types. Unified association keys such as batch number and wafer ID are required to enable full-link traceability. The high-frequency update rhythm (per-piece testing, per-batch output) requires the database to have high concurrent write performance to avoid data accumulation or write delays. Diverse data types and update frequencies require the operations maintenance link to divide different storage media and retention strategies. For example, store real-time inspection data in high-performance storage, and store historical yield reports in archive storage. In addition, quality traceability requires full historical data support. Data integrity and traceability must be ensured, and missing fields or broken associations must be avoided.
How to set the configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
MONGODB_URI | Connection string with replica set parameters, for example mongodb://admin:pass@host1:27017,host2:27017,host3:27017/fastgpt?replicaSet=rs0 | Adapt to high-concurrency write operations for single inspection data, and avoid service interruptions caused by single-node failures |
DB_MAX_LINK | 200–300 | Match the high-frequency write demand for thousands of cells per batch, and prevent connection exhaustion |
REDIS_URL | Independent instance connection string with access password | Cache batch association information and high-frequency query results to speed up query speed for full-link traceability |
| Data import batch size | 5000 entries per batch | Match the conventional production scale of single-batch cells, and avoid timeout during single import |
| Archive storage period | Calibrated based on actual measurements | Adapt to compliance requirements for enterprise quality traceability, and avoid invalid data occupying storage resources |
Three common mistakes
- Symptom: A MongoDB replica set configuration failure error occurs when starting FastGPT. Cause: The replica set parameters of
MONGODB_URIare not correctly configured, and only a single-node connection string is used for deployment. - Symptom: Field matching exceptions occur when importing production data in Excel format, and some process parameters cannot be associated with the corresponding batch. Cause: The naming rules of association fields such as batch number and wafer ID are not unified, resulting in failure to establish correct full-link associations between data.
- Symptom: EL inspection images cannot be retrieved or previewed normally after bulk import. Cause: The correct file storage path is not configured, resulting in a mismatch between the image file index and the actual storage location.
How to confirm the configuration is complete
- Log in to the FastGPT v4.17.0 backend management panel, check the database connection status module, and confirm that the connection status of
MONGODB_URIandREDIS_URLshows normal. - Import a batch of simulated photovoltaic cell production data, and verify that batch association query, process parameter retrieval, and EL inspection image preview functions can be called normally.
- Trigger a batch data import task, confirm that there are no connection timeout or data loss errors during the import process, and the import time meets expectations.
- Retrieve the IV grading data of the specified batch, and confirm that the returned results include complete process parameters, inspection records, and associated EL inspection images.
The values provided here are common starting points and should be evaluated against the reader's own samples.