Field-Level Anomaly Detection for Industrial Process Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current industrial process automation systems require significant time, cost, and engineering effort to detect and diagnose process disturbances and variabilities, relying heavily on process historians and manual data analysis, which is inefficient and costly.
Innovation Solution
Implementing autonomous field devices with integrated anomaly detection algorithms and edge devices to autonomously detect anomalies at the field device level, enabling real-time data capture and analysis, and transmitting data to the cloud for scalable monitoring and optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional process historians and manual data analysis are used to detect and diagnose process disturbances, then comprehensive data collection and analysis capability is achieved, but significant time, cost, and engineering effort are required
Solution Approach 1:
Anomaly detection algorithms are pre-configured and embedded within field devices, enabling them to autonomously perform detection functions without requiring manual data collection, ingestion, or analysis. The field devices continuously monitor their own operational parameters and automatically identify anomalies, eliminating the time-consuming manual processes of data gathering and analysis while maintaining comprehensive detection capability
Solution Approach 2:
Field devices are equipped with self-diagnostic capabilities through integrated anomaly detection algorithms that allow them to autonomously monitor their own operational status, detect anomalies in real-time, and generate diagnostic data without external intervention. This self-service approach eliminates the need for process engineers to manually collect and analyze data, significantly reducing time loss while preserving thorough anomaly detection
2Reliability
If process engineers manually analyze data and create remediation plans, then accurate diagnosis and remediation are achieved, but high cost and engineering effort are expended
Solution Approach 1:
Field devices autonomously perform anomaly detection and generate diagnostic data using embedded algorithms, eliminating the need for expensive manual analysis by process engineers. The devices self-monitor operational parameters, automatically identify anomalies, and provide structured diagnostic information that can be directly utilized for remediation, maintaining diagnosis accuracy while dramatically reducing engineering effort and implementation cost
Solution Approach 2:
Edge devices serve as intermediaries between field devices and cloud infrastructure, performing preliminary data processing and anomaly detection at the network edge. This intermediary layer reduces the complexity and cost of cloud-based analysis while maintaining diagnostic accuracy by filtering and preparing data before cloud transmission, thereby reducing overall system implementation cost
3Loss of information
If data is collected and stored in process historians, then historical data availability is achieved, but significant data processing and ingestion work is required
Solution Approach 1:
Field devices perform preliminary anomaly detection and data processing locally using embedded algorithms before data needs to be stored or analyzed centrally. By pre-processing data at the source and only transmitting relevant diagnostic information to cloud or edge systems, the complexity of data ingestion and processing is significantly reduced while maintaining full data availability for historical analysis when needed
Solution Approach 2:
The patent extracts the anomaly detection function from centralized process historians and manual analysis systems, embedding it directly within field devices. This extraction eliminates the need for complex centralized data processing and ingestion pipelines, as each field device independently performs detection and generates structured diagnostic data, thereby reducing data processing complexity while preserving complete data availability
4Productivity
If autonomous field devices with anomaly detection algorithms are implemented, then real-time anomaly detection and faster response times are achieved, but device complexity increases
Solution Approach 1:
The anomaly detection system is segmented and distributed across individual field devices rather than centralized in complex processing systems. Each field device contains simplified anomaly detection logic tailored to its specific function, enabling real-time autonomous detection without requiring high complexity in any single device. This segmentation allows fast local response while keeping individual device complexity manageable
Solution Approach 2:
Field devices are designed with multi-functionality, integrating measurement, diagnostic, and anomaly detection capabilities within a single device platform. This universal design allows the same field device structure to perform multiple functions including data collection, local processing, autonomous anomaly detection, and communication, thereby achieving fast response times without proportionally increasing device complexity through specialized components
Data Source
AI summary
An industrial process automation system includes a field device level including a plurality of autonomous field devices, and an edge device. Each of the plurality of field devices includes a signal generating module generating a measurement signal, a diagnostic module generating diagnostic data based on the measurement signal, an anomaly detection algorithm that identifies an anomaly based on the diagnostic data, and a communication module transmitting the anomaly to the edge device and then to a cloud/server.