What the data for this category looks like
Data sources for this category include battery cell grading records, glass adhesive film backplane batch ledgers, factory MES systems, string welding and lamination production equipment, EL defect detection and IV power testers. Data updates follow the rhythm that each module has a dedicated set of records, and batch ledgers are generated synchronously for each batch of raw materials. Daily generated documents include layout diagrams, material batch records, lamination process parameter records, EL defect images, IV test reports, insulation voltage withstand test records, and finished product packing lists. Each document includes dedicated fields such as battery cell current gear, lamination temperature and pressure, and EL crack location.
How These Characteristics Impose Constraints on Workflow Orchestration
- The high-frequency updated data for each module requires the workflow to support variable calls at the single batch or single module level, to prevent mixing data across batches.
- Multi-source heterogeneous data from production equipment, testing instruments and ledger systems requires the workflow to support parallel pulling of files and fields from different sources across multiple nodes.
- Dedicated fields such as battery cell current gear and EL crack location require the workflow to have fixed field mapping rules configured, to ensure drafted documents match the parameters of the corresponding production stage.
- Exception reports need to be linked to EL images and IV test data, requiring the workflow to support passing uploaded test files as parameters to subsequent nodes.
How to Set the Configurations
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
variables | Bind dedicated fields such as battery cell batch number, EL image ID, and set as required | Match the dedicated production data of each module, avoid cross-batch call errors |
| Single node execution timeout | 300–600 seconds | Adapt to the pull and parsing duration of device logs and test reports |
| File upload parameters | Support jpg, png, csv formats, maximum 100 MB per file | Cover common document formats such as EL images and IV test reports, comply with the file upload specifications of FastGPT v4.17.0 |
| Node trigger condition | Triggered after form completion | Ensure input parameters are complete, avoid generating incomplete documents |
nodeId | Bind dedicated node IDs of the three core departments: string welding, lamination, and testing | Match the document drafting requirements of different production links |
| Context window size | Previous 20 conversation logs | Retain the context of the most recent production batch, avoid interference from redundant data |
Three Common Mistakes
- Phenomenon: The workflow allows drafting processes to be triggered when the form is not fully completed, resulting in documents missing core fields such as battery cell gear and lamination parameters. Cause: The node trigger condition is not configured to trigger after form completion, allowing incomplete input to start the process.
- Phenomenon: Uploaded EL images and IV test reports cannot be passed as parameters to the document drafting node, resulting in the generated exception report missing corresponding test data. Cause: The file upload variable binding function is not enabled in the
variablesconfiguration, or the factory's existing file format specifications are not matched. - Phenomenon: In the Docker-deployed FastGPT, the model test works normally and there are response logs in the background, but the workflow conversation shows a failure. Cause: The correct model node ID is not bound in the workflow's
nodeIdconfiguration, or the network policy between containers restricts communication between workflow nodes.
How to Verify the Configuration Is Correct
- Submit a complete form and the EL image of the corresponding batch, check whether the workflow can normally generate process documents or exception reports containing battery cell gear and EL defect information.
- View the workflow running logs, confirm that the uploaded files have been passed as
variablesto subsequent nodes, and the field mapping meets the requirements of the production link. - After triggering the workflow, check whether the conversation logs retain the relevant records of the most recent production batch, with no empty fields.
- Simulate incomplete form input, confirm that the workflow cannot start the document drafting process.
The values provided here are common starting points, and should be evaluated against the reader's own samples and operational requirements.