What this category's data looks like
The core data for this category comes from production and testing links including cell grading equipment, MES systems, welding torque tools, BMS software, EOL test benches, etc. Daily maintenance and production documents include grouping rules, BOM lists, welding and torque records, BMS flashing logs, insulation and voltage withstand reports, EOL test reports, and system configuration lists. Data update frequency is once per produced battery pack or energy storage cabinet, and synchronized each time the BMS software version changes. Document fields include professional parameters such as cell capacity, internal resistance, material code, torque value, flashing version number, etc. Data must be stored classified by production link, and there is no unified general storage format.
What constraints do these characteristics impose on the knowledge base retrieval and recall link
The production and maintenance data of this category has high update frequency, and each single piece of data is associated with a specific equipment unit. This requires the retrieval and recall link to support incremental synchronization and accurate matching of the exclusive configuration of a single device. Core documents such as grouping rules and BMS configurations are directly related to the safe operation of equipment, so retrieval results must prioritize matching professional technical documents to avoid interference from irrelevant content. Heterogeneous data formats from multiple sources require the preprocessing link to perform structured conversion according to document types, ensuring that fault phenomena and corresponding processing procedures can be accurately matched during retrieval. In addition, failure consequences such as insulation failure and BMS configuration errors are severe, so the accuracy of retrieval and recall must meet the reliability standards for on-site maintenance.
How to configure the settings
This configuration is based on FastGPT v4.17.0
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
similarity | 0.75–0.85 | Professional parameter matching accuracy requirements for battery PACK and energy storage systems are high. A value that is too low will introduce irrelevant documents, while a value that is too high will result in insufficient valid recall results |
limit | Top 8 entries | The number of relevant documents for single-device inspection is limited. Excessive entries will increase the context length and affect response speed |
usingReRank | Enabled | Documents for this category involve multi-dimensional professional content such as electrical, thermal management, and software configuration. Reranking can improve the sorting priority of highly relevant results while balancing response speed |
searchMode | embedding | Equipment maintenance Q&A requires matching fault phenomena with corresponding processing procedures. Keyword matching alone cannot cover professional scenarios |
datasetSearchUsingExtensionQuery | Enabled | On-site inspection personnel's questions may contain colloquial descriptions. Extended queries can cover more professional term variants |
collectionIds | Bind exclusive collections such as grouping rules and BMS documents for the current project | Avoid recalling knowledge base content from irrelevant categories, ensuring retrieval results align with the business scenario of this category |
Three common misconfigurations
- Phenomenon: Authentication failure is returned when calling the API to synchronize the knowledge base collection. Cause: The exclusive API key bound to the target knowledge base collection is not used, or the key is not granted collection read and write permissions.
- Phenomenon: An operation failure prompt appears when binding the Google Search API. Cause: The real-name verification process for the Google Search billing account has not been completed, or the call frequency exceeds the limit of the current package.
- Phenomenon: Retrieval response speed for equipment inspection Q&A is too slow. Cause: The number of recalled documents per single time is excessive, or the
similaritythreshold is set too low, resulting in a large number of irrelevant documents being recalled and increasing processing time.
How to verify the configuration is correct
- Execute a simulated on-site inspection question, such as "How to troubleshoot insulation alarms for a certain model of energy storage cabinet". Check whether the returned results include the exclusive insulation and voltage withstand reports and BMS configuration documents for this category. The
similarityconfiguration can be adjusted to match business requirements. - Call the API to test the knowledge base collection synchronization process. Confirm that the returned status code is 200 and the number of documents in the collection matches the source production data.
- Test the process of binding the Google Search API. Confirm that it can be called normally and return reference content for equipment maintenance of this category.
- Test the response speed of different questions. Confirm that it meets the real-time requirements of the business scenario. The
limitconfiguration can be adjusted for optimization.
The values given are common starting points and should be measured against the reader's own samples.