What the Data for This Category Looks Like
Data for this category is generated for individual transformers or high-voltage electrical products. Daily documents include winding and core assembly records, vacuum drying records, insulating oil reports, partial discharge and withstand voltage test reports, and nameplate parameter tables. Data sources include design BOM, silicon steel sheet and insulating material procurement data, vacuum and drying equipment operation data, oil test results, and test station test records. Data updates occur once per product or per raw material batch. Fields include insulation parameters, test values, material batch numbers, and similar items. There are no universal batch-standardized fields. Single record size fluctuates based on the number of test items.
What Constraints These Characteristics Impose on Workflow Orchestration
Exclusive data for individual products requires workflows to bind unique identifiers for single batches or units, to avoid cross-product data confusion. The per-unit or per-batch update rhythm requires workflows to support dynamic triggering, to match the non-fixed cycle of data generation. Mixed calls of multiple document types require workflow nodes to support reading parameters across document types, and adapt to format verification for different fields such as insulation parameters and test values. Fluctuations in single record size require workflows to reserve sufficient context processing space, to avoid node execution exceptions caused by overly long data. Product traceability requirements require workflows to retain complete node execution logs, to facilitate subsequent quality problem tracing.
Configuration Settings
This configuration is based on FastGPT v4.17.0.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
variables | Assigned using the combination of single product ID + material batch number | Matches the per-unit or per-batch data update rhythm, avoids cross-product data confusion |
nodeId | Assigned a unique ID based on document type + test project number | Adapts to mixed calls of multiple document types, facilitates traceability of node execution logs |
| Context length limit | 8000–12000 characters | Matches the fluctuation characteristics of single record size, reserves sufficient processing space |
| Node trigger condition | Triggered based on data upload completion time | Matches the non-fixed cycle data update rhythm, enables dynamic triggering |
| Maximum documents per single call | First 5 | Adapts to mixed calls of multiple document types, avoids node execution overload |
| Log retention period | 365 days | Meets traceability requirements for the full product lifecycle |
Three Common Misconfigurations
- Phenomenon: The workflow returns an error with empty fields, or the verification result does not match actual test data. Cause: The
variablesparameter is not bound to the single product ID, resulting in calls to incorrect cross-batch data. - Phenomenon: Workflow nodes execute with a timeout, or returned context content is incomplete. Cause: The context length limit is not set reasonably, resulting in single record size exceeding the node processing upper limit.
- Phenomenon: The workflow only supports referencing a single document, and cannot call both the insulating oil report and the withstand voltage test report at the same time. Cause: Multi-variable binding rules are not configured correctly, or the maximum documents per single call setting is too low.
How to Confirm Correct Configuration
- Upload the complete set of test documents for a single product, trigger workflow execution, check whether returned parameters match the content of the uploaded documents. Adjust the binding rules of the
variablesparameter until the match is correct. - Upload test data of different sizes, adjust the value range of the context length limit, confirm that workflow nodes have no execution exceptions, and determine a reasonable length range based on actual data size.
- Configure upload triggers for multiple documents of different types, check whether the workflow can simultaneously read and verify the core fields of each document. Adjust the maximum documents per single call based on the number of documents.
- View workflow execution logs, confirm that each node execution has a unique
nodeIdidentifier, to facilitate subsequent quality problem tracing.
The values given are common starting points and should be measured against the reader's own samples.