Industrial Alarm Explanation Using ML Feature Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current industrial automation systems face challenges in efficiently managing alarms generated by machine-learning-driven systems, leading to delayed or inappropriate responses that can result in catastrophic conditions or resource wastage due to the complexity of interpreting model-driven predictions.
Innovation Solution
An explainer system that utilizes techniques like LIME and SHAP to identify influential features, extracts contextual information from databases and ontologies, and provides historical context through annotated P&IDs and knowledge graphs to enhance operator understanding of alarms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine-learning-driven systems are used to predict anomalous behavior, then the capability to detect safety or performance issues is enhanced, but the complexity of interpreting model-driven predictions increases
Solution Approach 1:
The patent introduces an explainer system as an intermediary component that sits between the machine learning model and the human operator. This explainer system translates complex model predictions into interpretable explanations, allowing operators to understand why alarms are raised without needing to interpret the raw model output directly. The explainer acts as a mediator that bridges the gap between sophisticated ML capabilities and human comprehension.
Solution Approach 2:
The patent replaces the traditional mechanical approach of direct model-output interpretation with an information-processing system that generates explanatory narratives. Instead of operators directly analyzing complex model predictions, the system automatically processes the predictions through explanation generation algorithms that produce human-understandable reasoning, substituting manual interpretation mechanics with automated explanatory processing.
2Ease of operation
If operators analyze alarms based on their experience, then decisions can be made, but the analysis can take a significant amount of time
Solution Approach 1:
The explainer system performs preliminary analysis and explanation generation before the operator needs to make a decision. By pre-processing the alarm information and preparing interpretive explanations in advance, the system reduces the cognitive load and time required for operators to analyze alarms. The explanations are ready when the operator needs them, eliminating the need for time-consuming on-the-spot analysis.
Solution Approach 2:
The system enables operators to self-serve by providing them with pre-prepared explanations and contextual information that they can immediately use for decision-making. Instead of requiring operators to invest significant time in analysis, the system serves them with ready-to-use interpretive content that supports rapid decision-making while maintaining operational autonomy.
3Reliability
If operators respond to all alarms, then safety is maintained, but resources may be wasted on non-critical issues
Solution Approach 1:
The explainer system applies local quality by providing differentiated explanations tailored to the specific characteristics of each alarm. Rather than treating all alarms uniformly, the system generates context-specific explanations that highlight the criticality and nature of each individual alarm, enabling operators to prioritize responses based on actual need rather than applying a blanket response strategy to all alarms.
Solution Approach 2:
The system changes the parameter of information quality by transforming raw alarm data into contextualized explanations with varying levels of detail and urgency. By adjusting the information parameters based on the specific alarm characteristics and operational context, the system enables operators to distinguish between critical and non-critical issues, optimizing resource allocation while maintaining safety.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
There is provided an explainer system (100) for explaining an alarm raised by a machine learned model (152) of an industrial automation system. The explainer system is configured to: receive model output from the machine learned model trained to predict anomalous behaviour in the industrial automation system and to raise the alarm; process the model output using at least one prediction explanation technique to identify at least one influential feature which contributed to the model output; use the identified at least one influential feature to extract contextual information from at least one machine-readable information source (106, 108, 110, 114, 118) pertaining to the industrial automation system; and prepare the extracted contextual information for display to an operator of the industrial automation system, to enable the operator to select an appropriate action to take in response to the alarm for ensuring proper functioning of the industrial automation system.