Industrial Data Broker Contextualization for AI Analytics Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial automation systems face challenges in deriving value from large amounts of unstructured and uncorrelated industrial data, leading to inefficiencies in data storage, processing, and the identification of spurious correlations, which hinders the extraction of actionable insights.
Innovation Solution
A smart gateway platform that leverages domain expertise to model and contextualize industrial data, reducing the data space for AI analytics by pre-defining relevant data items and correlations, and using an industrial publish-subscribe approach to stream only relevant data to higher-level systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all industrial data is collected and stored for AI analytics, then the completeness of data is improved, but the data storage requirements and processing complexity increase significantly
Solution Approach 1:
The patent extracts only the relevant data items needed for specific AI analytics tasks by using domain expertise to identify and select pertinent data from the industrial data stream, rather than collecting and storing all available data. This extraction approach reduces storage requirements while maintaining the completeness of relevant information.
Solution Approach 2:
The patent applies local quality by tailoring the data collection and processing approach to specific analytics tasks and business objectives. Different data subsets are collected and processed based on their relevance to particular AI models and analytical needs, rather than uniformly processing all industrial data.
2Measurement precision
If domain expertise is used to pre-define relevant data items and correlations, then the accuracy of analytics insights is improved, but the system complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining relevant data items, correlations, and data models using domain expertise before the actual AI analytics execution. This preliminary structuring of data relationships and relevance rules simplifies the subsequent analytics process while improving accuracy, as the system already has predefined guidance on what data to collect and how to interpret correlations.
3Productivity
If data is streamed selectively based on publish-subscribe topics, then the processing efficiency is improved, but the data filtering complexity increases
Solution Approach 1:
The patent segments the industrial data stream into distinct topics and categories based on domain expertise and business relevance. Each data item is tagged with relevant topics, enabling selective streaming to different analytics consumers. This segmentation reduces processing overhead by avoiding the transmission and processing of irrelevant data, while the topic-based organization simplifies the filtering mechanism.
Data Source
AI summary
An industrial data broker system receives contextualized industrial data from one or more industrial devices that support data modeling at the device level. The received industrial data is augmented with contextualization metadata that defines correlations between the data relevant to an analytical objective, and labels specifying analytic topics to which each data item is relevant. The broker system allows external systems, such as analytic systems, to subscribe to topics of interest, and streams a subset of contextualized device data relevant to the topic of interest to the external system for analysis.