Diagnostics in building automation
AssigneeSiemens Industry, Inc.
InventorOsman Ahmed
Life of this patent
Estimating the term…
Abstract
Using data from various sources, clustering or other unsupervised learning determines a relationship of the data to performance. Meta data or business data different than building automation data is used to diagnose building automation. Relationships of building automation to the meta or business data are determined with clustering or other case-based reasoning. For multiple building situations, clustering with or without the meta data identifies poor performing buildings, equipment, automation control, or enterprise function.
Claims20
A method of building automation diagnosis in a building management system, the method comprising:
- accessing first data related to a plurality of buildings by a building analytics system of the building management system, the first data including building management system data and enterprise data different than building management system data, the enterprise data for an enterprise associated with the buildings of the plurality and the building management system data being for the buildings of the plurality;
- applying, by the building analytics system, the first data to unsupervised machine learning with a measure of performance for case based reasoning;
- and
- outputting, as a result of the applying, at least one component of the first data indicative of poorer measures of the performance of the building management system identified in the first data, the at least one component including the enterprise data,
- wherein the accessing, applying, and outputting are performed without prior training of the unsupervised machine learning.
A building management system for building automation diagnosis, the system comprising:
- building automation systems for heating, ventilation, and air conditioning, the building automation systems being for multiple buildings and configured to output operational data;
- a building processor configured to apply unsupervised machine learning to the operational data of the building automation systems and to identify a sub-set of the building automation systems based on the operational data onto which said unsupervised machine learning is applied;
- and
- a display configured to output the identification of the sub-set,
- wherein said unsupervised machine learning is unsupervised machine learning without prior training.
A method for building automation diagnosis in a building management system, the method comprising:
- accessing, by a building analytics system of the building management system, enterprise data in an enterprise database, the enterprise data related to one or more buildings, the enterprise data being other than building management system data;
- clustering, by the building analytics system, for performance of a building management system with the enterprise data;
- and
- presenting on a display of the building analytics system, results of the clustering, the results including a relationship of the enterprise data to operation of the one or more buildings identified by the clustering,
- wherein the accessing, clustering, and presenting are performed without prior training of the unsupervised machine learning.
Dependencies are inferred from the claim text.