About Prompt
- Prompt Type – Dynamic
- Prompt Platform – ChatGPT, Grok, Deepseek, Gemini, Copilot, Midjourney, Meta AI and more
- Niche – Real-Time Issue Documentation
- Language – English
- Category – Anomaly Reporting
- Prompt Title – AI Prompt for Automated Reporting of Production Line Anomalies
Prompt Details
**Prompt Type:** Dynamic
**Purpose:** Real-time Anomaly Reporting for Production Lines
**Target Platform:** All AI Platforms
**Instructions:**
This prompt aims to generate a detailed, real-time report of anomalies detected within a production line. You will receive input data representing various sensor readings, machine statuses, and product quality metrics. Your task is to analyze this data, identify anomalies, and generate a comprehensive report. The report should be structured, informative, and actionable, enabling swift intervention and minimizing production disruptions.
**Input Data Format:**
The input data will be provided as a JSON object containing the following key-value pairs:
* `timestamp`: ISO 8601 timestamp (e.g., “2024-10-27T10:30:00Z”)
* `line_id`: Unique identifier of the production line (e.g., “Line_A”)
* `station_id`: Unique identifier of the station within the production line where the data originates (e.g., “Station_3”)
* `sensor_data`: A JSON object containing sensor readings with sensor names as keys and their corresponding values. (e.g., `{“temperature”: 85, “pressure”: 120, “vibration”: 0.5}`)
* `machine_status`: A string indicating the machine’s operational status (e.g., “Running”, “Stopped”, “Error”)
* `product_quality_metrics`: A JSON object containing product quality metrics. (e.g., `{“weight”: 10.2, “dimensions”: [10, 5, 2]}`)
**Anomaly Detection Logic:**
1. **Historical Data Comparison:** Consider historical data for the same production line and station to identify deviations from normal operating ranges. The historical data timeframe should be configurable and provided as part of the input (e.g., `historical_data_window`: “1 week”). If historical data is unavailable, focus on identifying sudden changes or extreme values.
2. **Threshold Breaches:** Detect anomalies based on pre-defined thresholds for each sensor reading and product quality metric. These thresholds can be provided as part of the input or learned from the historical data if available. (e.g., `thresholds`: `{“temperature”: {“min”: 70, “max”: 90}, “pressure”: {“min”: 100, “max”: 140}}`)
3. **Correlation Analysis:** Investigate correlations between different sensor readings and metrics to identify potential anomalies that might not be apparent when considering individual data points in isolation.
4. **Machine Status:** Analyze machine status changes to detect unexpected stops or errors.
**Report Format:**
The report should be formatted as a JSON object with the following key-value pairs:
* `timestamp`: ISO 8601 timestamp of the report generation time.
* `line_id`: Unique identifier of the production line.
* `station_id`: Unique identifier of the affected station.
* `anomalies`: An array of JSON objects, each representing a detected anomaly. Each anomaly object should contain:
* `anomaly_type`: A descriptive name of the anomaly type (e.g., “Temperature Spike”, “Pressure Drop”, “Weight Deviation”).
* `affected_sensor/metric`: Name of the sensor or metric exhibiting the anomaly.
* `current_value`: Current value of the affected sensor/metric.
* `expected_value/range`: Expected value or range based on historical data or thresholds.
* `severity`: Severity level of the anomaly (e.g., “Low”, “Medium”, “High”). Base severity on the deviation from the expected value/range and potential impact on production.
* `description`: A concise human-readable description of the anomaly.
* `recommendations`: Suggested actions to address the anomaly (e.g., “Check cooling system”, “Inspect pressure regulator”).
**Example Input:**
“`json
{
“timestamp”: “2024-10-27T10:30:00Z”,
“line_id”: “Line_A”,
“station_id”: “Station_3”,
“sensor_data”: {“temperature”: 105, “pressure”: 120},
“machine_status”: “Running”,
“product_quality_metrics”: {“weight”: 10.2},
“historical_data_window”: “1 day”,
“thresholds”: {“temperature”: {“min”: 70, “max”: 90}}
}
“`
**Example Output:**
“`json
{
“timestamp”: “2024-10-27T10:30:05Z”,
“line_id”: “Line_A”,
“station_id”: “Station_3”,
“anomalies”: [
{
“anomaly_type”: “Temperature Spike”,
“affected_sensor/metric”: “temperature”,
“current_value”: 105,
“expected_value/range”: “[70, 90]”,
“severity”: “High”,
“description”: “Temperature exceeded the defined threshold of 90 degrees.”,
“recommendations”: “Check cooling system and inspect heating elements.”
}
]
}
“`
**Additional Notes:**
* Adapt the input and output formats to your specific needs.
* Clearly define the expected level of detail and the desired tone of the report.
* Experiment with different prompting techniques to optimize the performance of the AI model for your specific use case. Provide feedback to the AI based on the quality of the generated reports to further refine its capabilities.