HMI Model State Visualization for Industrial ML Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial automation environments face challenges in effectively integrating machine learning models to assist in the control of OEM assets and in efficiently surfacing the health status of these models to plant operators.
Innovation Solution
A system comprising a machine learning component and a Human Machine Interface (HMI) component is introduced to visualize the status of machine learning models in industrial automation environments. The machine learning component processes inputs from industrial devices to generate outputs that influence device operations and reports operational data to the HMI. The HMI displays visualizations of the machine learning model status, allowing operators to assess the health and relationships of the models with industrial devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine learning models are integrated into industrial automation environments, then the ability to assist in control of OEM assets is improved, but the complexity of the system increases
Solution Approach 1:
The patent introduces an HMI intermediary component that bridges the machine learning model and the industrial automation system. The HMI receives operational data from the ML model, processes it into visual representations, and presents it to operators. This intermediary layer manages the complexity by handling data transformation and presentation, allowing the core ML control functionality to remain focused on its primary task while the HMI handles the complexity of human-machine interaction and data visualization.
2Loss of information
If machine learning model health status is surfaced to plant operators, then operational awareness is improved, but the device complexity increases
Solution Approach 1:
The patent employs color-coded visual indicators to represent machine learning model health status. Different colors (e.g., green for healthy, yellow for degraded, red for failed) are used to convey the operational state of the ML model and its components. This visual encoding system allows operators to quickly assess model health without interpreting complex numerical data, reducing the cognitive load and simplifying the HMI while maintaining comprehensive operational awareness.
Solution Approach 2:
The HMI divides the machine learning model health information into segmented, manageable visual components. Instead of presenting a single complex status indicator, the system breaks down the model health into multiple discrete visual elements representing different aspects (e.g., model performance, data quality, computational status). This segmentation allows operators to understand specific areas of concern while keeping the overall interface simple and organized.
3Loss of information
If expanded views of machine learning components are displayed on HMI, then detailed operational data is improved, but the information display complexity increases
Solution Approach 1:
The HMI implements a nested view structure where high-level summaries of machine learning model status are displayed at the parent level, and detailed operational data is available in child or expanded views. Operators can drill down from overview representations into more detailed information about specific model components, data flows, or performance metrics. This nesting approach allows comprehensive information to be organized in a hierarchical manner, providing detail on demand without overwhelming the interface with all information simultaneously.
Data Source
AI summary
Various embodiments of the present technology generally relate to industrial automation environments. More specifically, embodiments include systems and methods to visualize machine learning model status in an industrial automation environment. In some examples, a machine learning component receives process inputs associated with industrial devices in the industrial automated environment. The machine learning component processes the inputs to generate machine learning outputs and transfers the machine learning outputs to influence one or more functions of the industrial devices. The machine learning component reports operational data characterizing the machine learning outputs. A Human Machine Interface (HMI) component displays a visualization of the machine learning component and receives the operational data from the machine learning component. In response to a user selection, the HMI component displays an expanded view of the machine learning component that comprises the operational data and that indicates relationships between the machine learning component and the industrial devices.