Edge Gateway Feature Engineering for KPI Data Context Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial automation systems face challenges in collecting and formatting data from distributed industrial devices into meaningful presentations for users, as much of the data is unstructured and lacks context, requiring developers to define its meaning, which is burdensome and time-consuming.
Innovation Solution
The system employs structured data types, referred to as Basic Information Data Types (BIDTs), which are configured on industrial devices to represent data in a structured format, allowing users to define associations and metadata, and a gateway device that references these BIDTs to generate contextualized presentations and execute scripts for key performance indicators (KPIs).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If developers manually define and format unstructured industrial data, then data meaning and context are established, but the process becomes burdensome and time-consuming
Solution Approach 1:
The system performs preliminary action by automatically generating data models and contextual information before users need to analyze the data. The gateway device pre-processes industrial data, defines data models with relevant properties, and prepares contextualized presentations in advance, eliminating the need for developers to manually define data meaning during analysis.
Solution Approach 2:
The system enables self-service by allowing the gateway device to automatically perform data collection, formatting, and contextualization without requiring developer intervention. The device autonomously defines data models, retrieves data from multiple sources, and generates contextualized presentations, making the system serve itself rather than requiring external developers to perform these tasks.
2Measurement precision
If all industrial data points are collected and analyzed, then comprehensive analysis is achieved, but the complexity and resources required increase significantly
Solution Approach 1:
The system applies extraction by identifying and extracting only the most relevant data points and contextual information needed for specific analytical purposes. Rather than processing all available industrial data, the gateway device selectively retrieves and presents data that directly relates to the queried context or performance indicators, reducing complexity while maintaining analysis accuracy.
Solution Approach 2:
The system implements local quality by providing different levels and types of data contextualization tailored to specific needs. Rather than uniformly processing all data with the same level of detail, the gateway device adapts the data model definition and contextualization depth to match the specific analytical context, device type, or user requirement, optimizing resources while maintaining precision where needed.
3Loss of information
If extensive data analysis is performed on all collected data, then comprehensive insights are obtained, but the time required for assessment increases
Solution Approach 1:
The system performs preliminary action by pre-defining data models and organizing data structures before analysis is needed. The gateway device prepares contextualized data presentations and identifies relevant data relationships in advance, so when analysis is required, the work has already been partially completed, significantly reducing assessment time while maintaining insight completeness.
Solution Approach 2:
The system applies partial action by performing sufficient data analysis and contextualization to achieve the required insight level without over-processing. The gateway device identifies and processes only the necessary portion of data needed to answer the specific query or assess the particular context, avoiding unnecessary analysis of irrelevant data points while still providing comprehensive insights for the given purpose.
Data Source
AI summary
An industrial gateway device supports interactive feature engineering tools that guide a user through an intuitive process for configuring data analytics for key performance indicators (KPIs) of interest. A feature engineering interface renders an interactive model view that displays available data points such that the data points are organized hierarchically according to plant, machine, machine property, or other elements. The user can select, from this model view, data points having an impact on the KPI of interest. The interface also allows the user to define an executable script that defines a mathematical relationship between the selected data points and the KPI. This configuration yields an output model that defines a reduced set of data points to be collected and analyzed, as well as an executable script for assessing a state of the KPI as a function of the reduced data point values.