BMS Point Auto-Tagging Using Name Segmentation and Confidence Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual tagging of points in building management systems is labor-intensive and costly, and existing machine learning approaches require large training datasets and are inefficient due to inconsistent naming conventions and irrelevant information.
Innovation Solution
A system that uses a combination of machine learning and expert systems to auto-tag points by extracting metadata tags from building management system data, leveraging a conditional random field model and neural network to predict equipment, location, and function, and assigning confidence scores, with expert systems resolving conflicts and improving tag quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual tagging of points is performed by technicians or operators, then tagging accuracy can be maintained, but labor intensity and cost increase significantly
Solution Approach 1:
The system enables automatic self-tagging of BMS points by utilizing the existing point name structure and descriptor information. The neural network model processes point names and descriptors to automatically generate appropriate tags without requiring manual intervention, thereby maintaining accuracy while dramatically improving efficiency and reducing labor costs.
2Extent of automation
If basic machine learning methods are applied with n-gram breakdown, then automation is achieved, but massive training data is required which is rarely available
Solution Approach 1:
The system extracts and utilizes the inherent structure within point names and descriptor information as features for the neural network model. By taking out and leveraging the existing semantic information already present in the point naming conventions and descriptors, the system achieves effective tagging automation without requiring external massive training datasets.
3Loss of information
If point names include building name and other irrelevant information, then complete point identification is achieved, but machine learning performance deteriorates due to noise
Solution Approach 1:
The system segments the point name into distinct components (building name, equipment identifier, point descriptor) and processes each segment appropriately. The neural network model learns to focus on the relevant segments (equipment identifier and descriptor) while treating the building name as contextual information, thereby maintaining complete point identification while improving tagging accuracy by reducing noise from irrelevant information.
Data Source
AI summary
There is described a building management system and a method for auto-tagging points. Data associated with multiple points of a site are received, and each point is associated with a point name and a point descriptor. A building name is identified based on the point name for each point by extracting a first part detected frequently among the data associated with the points. A point equipment is determined from a second part of each point name and a point function is determined from a third part of each point name. A set of point tags is generated based on the point equipment, the point function, and the point descriptor. Confidence scores are created for the set of point tags based on matching characteristics to a common tag set.