Building Automation Label Verification for Fault Diagnostics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional building automation systems struggle with accurately determining whether assigned labels adequately characterize data points, leading to issues with analysis and diagnostics due to manual labeling challenges and lack of standardized naming conventions.
Innovation Solution
A label plausibility approach using a combination of tree-based and image transformation classifiers to independently verify if data points are correctly understood and labeled by analyzing timeseries data, regardless of naming conventions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual labeling of data points is performed, then data points can be assigned labels, but the process is error-prone and difficult for installation and maintenance personnel
Solution Approach 1:
The system automatically performs label plausibility checking using machine learning classifiers (tree-based and image transformation classifiers) that analyze timeseries data to verify whether assigned labels correctly characterize data points. This self-service approach eliminates the need for manual labeling verification by installation and maintenance personnel, reducing errors while maintaining ease of operation.
2Stability of the object's composition
If standardized naming conventions are implemented, then data point labeling becomes more consistent, but the system loses flexibility in handling non-standardized names
Solution Approach 1:
The system segments the labeling verification process into two independent classifier components: a tree-based classifier that handles structured label validation and an image transformation classifier that processes timeseries data patterns. This segmentation allows the system to maintain consistent labeling standards while adapting to various naming conventions, as each classifier can be trained to recognize different naming patterns without compromising overall consistency.
3Reliability
If conventional fault detection systems are used, then basic fault detection is achieved, but the systems fail to determine whether assigned labels adequately characterize data points
Solution Approach 1:
The system merges two distinct classifier approaches (tree-based classifier and image transformation classifier) into a unified label plausibility checking system. Both classifiers receive the same timeseries data input and operate independently to verify label accuracy. This combination improves fault detection reliability by cross-validating labels through multiple analytical methods while managing complexity through a modular architecture where each classifier is independently trained and evaluated.
Data Source
AI summary
Systems and methods detect and diagnose faults of a building automation system. Timeseries data are received from the building automation system. A label plausibility is determined for each set of timeseries data and the corresponding label associated with the set of timeseries data based on a tree-based classifier and an image transformation classifier. The tree-based classifier and the image transformation classifier receive the same data input and operating distinctly from each other.