What this category of data looks like
Data sources for this category include EAM/CMMS systems, equipment sensor data, operator inspection records, maintenance history archives, spare parts inventory lists, and statutory inspection documents. Data update cycles cover shifts, daily, weekly, monthly periods, and real-time updates when faults occur. Daily documents include equipment ledgers, inspection forms, maintenance plans, work orders, fault analysis reports, spare parts lists, and special equipment inspection records. Core fields include equipment number, inspection time, fault code, spare part model, inspection validity period, and others. Units include units, times, hours, days, and others.
How These Characteristics Impact Tool Calling and Plugin Usage
Multi-source heterogeneous data sources require tool calling to adapt to parsing logic for different data formats. This prevents data extraction failures caused by format incompatibility. Diverse update frequencies require tool calling to support both scheduled pull and real-time trigger modes. This meets the needs of different business scenarios. The complexity of document structures requires tools to accurately identify fields for different types of operation and maintenance documents. For example, distinguish inspection items in inspection forms from fault descriptions in work orders. Equipment operation and maintenance data must comply with production scheduling. The trigger timing of tool calling must align with production windows. This avoids disrupting normal production activities.
This page is compatible with FastGPT v4.17.0.
How to Configure Settings
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
| Workflow/API tool selection | auto | Automatically match and call compatible tools for multiple types of equipment operation and maintenance data sources, covering inspection, maintenance, spare parts and other types of data |
| Workflow/API function-call configuration | Enabled | Enable precise function call mode to avoid failure to extract professional equipment fields with general question answering logic |
| Maximum number of data sources per call | First 6 latest data sources | Control resource consumption of tool calls, while covering recent core equipment operation and maintenance data |
| Timeout period | 300-600 seconds | Adapt to long-duration scenarios such as equipment data pulling and ledger parsing |
| Number of historical conversations carried | First 3 conversations | Focus on the context of current equipment operation and maintenance scenarios, avoiding interference from redundant information on tool calls |
| Number of recalled results | First 8 entries | Cover major equipment faults and maintenance records, meeting the basic data requirements for report generation |
Three Common Mistakes
- Phenomenon: Empty results are returned when calling equipment ledger tools. Cause: Access permissions for the corresponding EAM/CMMS data source are not configured, or the data source update frequency setting does not match actual business rhythms.
- Phenomenon: Empty conversation logs appear after workflow runs. Cause: The workflow's log recording switch is not enabled, or the trigger timing does not match the real-time nature of equipment fault reports.
- Phenomenon: Unable to generate replacement solutions matching equipment spare part models when calling code generation tools. Cause: The spare parts list is not used as a context data source for tool calls, or the
functionCallparameter is not enabled for precise function matching.
How to Confirm the Configuration Is Correct
- Manually trigger tool calls. Check whether the returned equipment data fields cover the core items required by the business.
- View tool call logs. Confirm that the value of the
toolChoiceparameter matches the configured setting. - Simulate an equipment fault report scenario. Verify that historical conversations are correctly carried into the tool call context.
- Adjust the update frequency of data sources. Check whether the tool can pull the latest equipment operation and maintenance data at the set rhythm.
The values provided are common starting points and should be measured against the reader's own samples.