Efficient integration of machine learning models in building management systems

Patent metadata

Field	Value
Publication number	US-12045048-B2
Application number	US-202117218661-A
Country	US
Kind code	B2
Filing date	Mar 31, 2021
Priority date	Mar 31, 2021
Publication date	Jul 23, 2024
Grant date	Jul 23, 2024

Abstract

A building management system including one or more memory devices and one or more processors, the one or more memory devices configured to store instructions to be executed on the one or more processors, the processors configured to generate a first predictive model using a machine learning technique and an operating data set based on one or more operating parameters associated with at least one of a plurality of BMS subsystems. The processors are further configured to store the first predictive model at a first time interval, to receive a prediction request from a user input at a second time interval following the first time interval, and to retrieve the first predictive model. The one or more processors are further configured execute the retrieved predictive model to generate a first prediction in response to the prediction request; and initiate an automated control response based on the first prediction.

First claim

What is claimed is:

1. A building management system (BMS) comprising:
one or more memory devices configured to store instructions that, when executed on one or more processors, cause the one or more processors to:
generate, on a remote computing system, a first predictive model using one or more machine learning techniques, wherein the first predictive model is generated using a first operating data set in response to determining a size of the first operating data set exceeds a threshold, the first operating data set collected at a first time interval based on one or more operating parameters associated with at least one of a plurality of BMS subsystems;
store the first predictive model at the first time interval in the remote computing system in an unbuilt state;
receive a prediction request from a user input at a second time interval following the first time interval;
retrieve the first predictive model in the unbuilt state from the remote computing system;
rebuild, locally, the first predictive model;
execute the rebuilt predictive model to generate a first prediction in response to the prediction request;
initiate an automated response based on the first prediction, wherein initiating the automated response comprises performing an automated control action comprising at least one of (i) controlling equipment to enter a safety mode, (ii) operating the equipment at a derated state, (iii) shutting down the equipment, (iv) controlling the equipment using an adjusted setpoint, (v) controlling the equipment to work-around a fault, or (vi) controlling the equipment to prevent the fault from occurring; and
automatically train, at an end of an update period beginning when the first predictive model is generated, the first predictive model using the first operating data set and a second operating data set collected at a fourth time interval, wherein the fourth interval time is after the first interval time, wherein the update period is based on the size of the first operating data set.

2. The BMS of claim 1 , where in the BMS subsystems are virtual subsystems and the first operating data set is a virtual operating data set.

3. The BMS of claim 1 , wherein the instructions cause the one or more processors to:
in response to the prediction request, receive a current data set; and
generate the first prediction using the first predictive model based on the prediction request and the current data set.

4. The BMS of claim 1 , the instructions further comprising:
determining a diagnosis associated with the first prediction to facilitate taking preventative measures; and
performing the automated control action based on the diagnosis to prevent a predicted fault from occurring.

5. The BMS of claim 1 , wherein the first prediction is a fault prediction for at least one of a plurality of the BMS subsystems.

6. The BMS of claim 1 , wherein generating the first predictive model comprises:
transmitting the first operating data set based on the one or more operating parameters associated with at least one of a plurality of BMS subsystems to the remote computing system; and
receiving the first predictive model from the remote computing system.

7. The BMS of claim 1 , wherein storing the first predictive model at a first time interval comprises:
transmitting the first predictive model to the remote computing system.

8. The BMS of claim 1 , wherein retrieving the first predictive model comprises:
receiving the first predictive model from the remote computing system.

9. The BMS of claim 1 , wherein initiating an automated response based on the first prediction comprises:
transmitting a notification to a user indicating the first prediction.

10. The BMS of claim 9 , wherein the notification to the user comprises:
generating and displaying a GUI based on the first prediction.

11. The BMS of claim 1 , wherein the first predictive model is associated with at least one of a plurality of types of building equipment, and wherein the update period is further based on the at least one of the plurality of types of building equipment.

12. A method of controlling building equipment, comprising:
generating, on a remote computing system, a first predictive model using one or more machine learning techniques, wherein the first predictive model is generated using a first operating data set in response to determining a size of the first operating data set exceeds a threshold, the first operating data set collected at a first time interval based on one or more operating parameters associated with the building equipment;
storing the first predictive model at the first time interval in the remote computing system;
receiving a prediction request from a user input at a second time interval following the first time interval;
retrieving the first predictive model from the remote computing system;
rebuilding, in a local computing system, the first predictive model;
executing the rebuilt predictive model to generate a first prediction in response to the prediction request;
initiating an automated response based on the first prediction, wherein initiating the automated response comprises performing an automated control action comprising at least one of (i) controlling equipment to enter a safety mode, (ii) operating the equipment at a derated state, (iii) shutting down the equipment, (iv) controlling the equipment using an adjusted setpoint, (v) controlling the equipment to work-around a fault, or (vi) controlling the equipment to prevent the fault from occurring; and
automatically training, at an end of an update period beginning when the first predictive model is generated, the first predictive model using the first operating data set and a second operating data set collected at a fourth time interval, wherein the fourth interval time is after the first interval time, wherein the update period is based on the size of the first operating data set.

13. The method of claim 12 , further comprising:
in response to the prediction request, receiving a current data set; and
generating the first prediction using the first predictive model based on the prediction request and the current data set.

14. The method of claim 12 further comprising:
determining a diagnosis associated with the first prediction to facilitate taking preventative measures; and
performing the automated control action based on the diagnosis to prevent a predicted fault from occurring.

15. The method of claim 12 , wherein the first predictive model is associated with at least one of a plurality of types of building equipment, and wherein the update period is further based on the at least one of the plurality of types of building equipment.

16. A building management system (BMS) configured to:
generate, on a remote computing system, a first predictive model using one or more machine learning techniques, wherein the first predictive model is generated using a first operating data set in response to determining a size of the first operating data set exceeds a threshold, the first operating data set collected at a first time interval based on one or more operating parameters associated with HVAC equipment;
store the first predictive model at the first time interval in the remote computing system;
receive a prediction request from a user input at a second time interval following the first time interval;
retrieve the first predictive model from the remote computing system;
rebuild, locally, the first predictive model;
execute the rebuilt predictive model to generate a first prediction in response to the prediction request;
initiate an au

Assignees

• Johnson Controls Tech Co
• Tyco Fire & Security Gmbh

Inventors

• Shinde Priti
• Uppal Deepak

Classifications

• G06N20/00

  Machine learning · CPC title

• G05B23/0216

  Human interface functionality, e.g. monitoring system providing help to the user in the selection of tests or in its configuration · CPC title

• G05B23/024 Primary

  Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks · CPC title

• G06N5/01

  Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound · CPC title

• G06N20/10

  using kernel methods, e.g. support vector machines [SVM] · CPC title