Systems and methods for fault detection and handling by assessing building equipment performance

• US 10,281,363 B2
• Filed: 03/24/2016
• Issued: 05/07/2019
• Est. Priority Date: 03/24/2016
• Status: Active Grant

• Current Assignee: Tyco Fire & Security GmbH (Johnson Controls International plc)
• Original Assignee: Johnson Controls Technology Company (Johnson Controls International plc)
• Inventors: Vitullo, Steven R., Boettcher, Andrew J.
• Primary Examiner(s): Le, Toan K
• Application Number:
  US15/080,441
• Publication Number:
  US 20170276571A1
• Time in Patent Office: 1,139 Days
• CPC Class Codes: G01M 99/00 Subject matter not provided for in other groups of this subclass
• CPC Class Codes: G01M 99/005 Testing of complete machines, e.g. washing-machines or mobile phones testing of machine parts G01M13/00; testing of electric apparatus or components G01R31/50
• CPC Class Codes: G05B 23/0248 Causal models, e.g. fault tree; digraphs; qualitative physics
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling

Abstract

A system for detecting faults in building equipment includes an integration fault detector, a kernel density fault detector, and a fault detector selector. The integration fault detector is configured to detect faults in the building equipment by analyzing time series data using an integration fault detection technique. The kernel density fault detector is configured to detect faults in the building equipment by analyzing the time series data using a kernel density estimation fault detection technique. The fault detector selector is configured to select the integration fault detector or the kernel density fault detector for use in detecting faults in the building equipment based on an attribute of the time series data.

Claims

1. A system for detecting faults in building equipment, the system comprising:

an integration fault detector configured to detect faults in the building equipment by analyzing time series data using an integration fault detection technique;

a kernel density fault detector configured to detect faults in the building equipment by analyzing the time series data using a kernel density estimation fault detection technique; and

a fault detector selector configured to select the integration fault detector or the kernel density fault detector for use in detecting faults in the building equipment based on an attribute of the time series data; and

a building equipment controller configured to adjust an operation of the building equipment in response to detecting one or more faults in the building equipment.

2. The system of claim 1, wherein the building equipment comprise at least one of:

one or more individual devices comprising at least one of a heater, a chiller, an air handling unit, a variable air volume unit, a pump, a fan, a cooling tower, a valve, an actuator, a damper, and a thermal energy storage (TES) device;

a building subsystem comprising at least one of a building electrical subsystem, an information communication technology subsystem, a security subsystem, a HVAC subsystem, a lighting subsystem, a lift/escalators subsystem, a fire safety subsystem, a refrigeration subsystem, an advertising or signage subsystem, a cooking subsystem, a vending subsystem, and a printer or copy service subsystem;

one or more subplant systems comprising at least one of a heater subplant, a heat recovery chiller subplant, a chiller subplant, a cooling tower subplant, a hot TES subplant, and a cold TES subplant; and

an energy plant comprising a plurality of the subplant systems.

3. The system of claim 1, wherein:

the attribute of the time series data indicates that the time series data represent an uncontrolled process variable; and

the fault detector selector is configured to select the integration fault detector in response to a determination that the time series data represents an uncontrolled process variable.

4. The system of claim 3, wherein the uncontrolled process variable comprises at least one of an efficiency metric for the building equipment, a coefficient of performance for the building equipment, a power consumption of the building equipment, and an uncontrolled flowrate affected by operating the building equipment.

5. The system of claim 1, wherein:

the attribute of the time series data indicates that the time series data represent a variable controlled to a setpoint; and

the fault detector selector is configured to select the kernel density fault detector in response to a determination that the time series data represents a variable controlled to a setpoint.

6. The system of claim 5, wherein the variable controlled to a setpoint comprises at least one of a controlled flowrate, a controlled temperature, a controlled pressure, a controlled humidity, and a controlled power consumption.

7. The system of claim 1, wherein:

the integration fault detector and the kernel density fault detector are configured to initiate a corrective action in response to detecting a fault in the building equipment; and

the corrective action comprises at least one of generating an alarm, providing a feedback to a controller for the building equipment, shutting off the building equipment, and automatically switching from using the building equipment to using other equipment configured to perform a same function as the building equipment in an automated control process.

8. The system of claim 1, further comprising a modified z-score calculator configured to calculate a modified z-score for each data point of the time series data, wherein the modified z-score is provided as an input to the integration fault detector and the kernel density fault detector.

9. The system of claim 8, wherein the modified z-score calculator is configured to calculate the modified z-score by:

determining a median of the time series data;

determining a median absolute deviation of the time series data; and

using the median and the median absolute deviation to calculate the modified z-score.

10. The system of claim 1, wherein the integration fault detector comprises:

a distance calculator configured to calculate distance values comprising a distance value for each data point of the time series data;

an integrator configured to generate, for each data point of the time series data, a cumulative sum of the distance values calculated for the data point and any previous data points in the time series data; and

a performance metric evaluator configured to calculate calculated performance metrics comprising a performance metric for each data point of the time series data based on the cumulative sum for the data point and to detect faults in the building equipment by comparing the calculated performance metrics to one or more performance thresholds.

11. The system of claim 1, wherein the kernel density fault detector comprises:

a kernel density estimator configured to generate a kernel density function based on the time series data;

a cumulative distribution function generator configured to generate a cumulative distribution function based on the kernel density function;

a threshold calculator configured to determine one or more control chart thresholds using the cumulative distribution function; and

an abnormal data detector configured to detect faults in the building equipment by applying control chart rules and the control chart thresholds to the time series data.

12. A method for detecting faults in building equipment using an integration fault detection technique, the method comprising:

operating the building equipment to generate time series data and receiving the time series data, wherein the time series data comprises a plurality of data points each having a value affected by an operation of the building equipment;

calculating a modified z-score for each of the plurality of data points;

calculating a distance value for each of the plurality of data points, the distance value indicating a difference between the modified z-score for the data point and an outlier threshold;

generating, for each of the plurality of data points, a cumulative sum of the distance values calculated for the data point and any previous data points in the time series data;

calculating, for each of the plurality of data points, a performance metric based on the cumulative sum for the data point;

detecting faults in the building equipment by comparing the calculated performance metrics to one or more performance thresholds; and

adjusting the operation of the building equipment in response to detecting one or more faults in the building equipment.

13. The method of claim 12, wherein calculating the modified z-score comprises:

determining a median of the time series data;

determining a median absolute deviation of the time series data; and

using the median and the median absolute deviation to calculate the modified z-score.

14. The method of claim 12, wherein calculating the distance value comprises, for each of the plurality of data points:

determining whether the modified z-score for the data point is positive or negative;

calculating a first difference between the modified z-score and a first outlier threshold in response to a determination that the modified z-score is positive; and

calculating a second difference between the modified z-score and a second outlier threshold in response to a determination that the modified z-score is negative.

15. The method of claim 12, wherein calculating the distance value comprises:

determining whether the modified z-score for the data point is positive or negative;

subtracting a first outlier threshold from the modified z-score in response to a determination that the modified z-score is positive; and

subtracting the modified z-score from a second outlier threshold in response to a determination that the modified z-score is negative.

16. The method of claim 12, wherein generating the cumulative sum comprises, for each of the plurality of data points:

identifying a previous cumulative sum calculated for a previous data point in the time series data;

adding the distance value for the data point to the previous cumulative sum;

determining whether a result of the adding has a negative value;

in response to a determination that the result of the adding has a negative value, setting the cumulative sum for the data point to zero; and

in response to a determination that the result of the adding does not have a negative value, setting the cumulative sum for the data point to the result of the adding.

17. The method of claim 16, wherein identifying the previous cumulative sum comprises:

determining whether the data point is a first data point of the time series data; and

setting the previous cumulative sum to zero in response to a determination that the data point is the first data point of the time series data.

18. The method of claim 12, wherein the building equipment comprise at least one of:

one or more individual devices comprising at least one of a heater, a chiller, an air handling unit, a variable air volume unit, a pump, a fan, a cooling tower, a valve, an actuator, a damper, and a thermal energy storage (TES) device;

a building subsystem comprising at least one of a building electrical subsystem, an information communication technology subsystem, a security subsystem, a HVAC subsystem, a lighting subsystem, a lift/escalators subsystem, a fire safety subsystem, a refrigeration subsystem, an advertising or signage subsystem, a cooking subsystem, a vending subsystem, and a printer or copy service subsystem;

one or more subplant systems comprising at least one of a heater subplant, a heat recovery chiller subplant, a chiller subplant, a cooling tower subplant, a hot TES subplant, and a cold TES subplant; and

an energy plant comprising a plurality of the subplant systems.

19. The method of claim 12, wherein the time series data represent an uncontrolled process variable comprising at least one of an efficiency metric for the building equipment, a coefficient of performance for the building equipment, a power consumption of the building equipment, and an uncontrolled flowrate affected by operating the building equipment.

20. A method for detecting faults in building equipment using a kernel density fault detection technique, the method comprising:

operating the building equipment to generate time series data and receiving the time series data, wherein the time series data comprise a plurality of data points each having a value affected by an operation of the building equipment;

calculating modified z-scores comprising a modified z-score for each of the plurality of data points;

generating a kernel density distribution based on the modified z-scores;

generating a cumulative distribution function based on the kernel density distribution;

determining one or more control chart thresholds using the cumulative distribution function;

detecting faults in the building equipment by applying control chart rules and the control chart thresholds to the modified z-scores; and

adjusting the operation of the building equipment in response to detecting one or more faults in the building equipment.

21. The method of claim 20, wherein the cumulative distribution function is an empirical cumulative distribution function derived from empirical data.

22. The method of claim 20, wherein generating the kernel density distribution comprises:

generating kernel smoothing function comprising a kernel smoothing function for each of the modified z-scores; and

combining the kernel smoothing functions to form the kernel density distribution.

23. The method of claim 20, wherein the cumulative distribution function expresses a probability that a random modified z-score has a value less than or equal to a threshold value, wherein the random modified z-score is selected from a set of modified z-scores defined by the kernel density distribution.

24. The method of claim 20, wherein determining the one or more control chart thresholds comprises:

identifying one or more probability values;

generating an inverse of the cumulative distribution function; and

using the one or more probability values as inputs to the inverse of the cumulative distribution function to determine a control chart threshold corresponding to each of the probability values.

25. The method of claim 20, wherein detecting faults in the building equipment by applying the control chart rules and the control chart thresholds to the modified z-scores comprises at least one of:

determining that a fault exists in response to a determination that a predetermined number of the modified z-scores are outside of a first range defined by the control chart thresholds; and

determining that a fault exists in response to a determination that a percentage of the predetermined number of the modified z-scores are outside of a second range defined by the control chart thresholds, the second range being wider than the first range.

26. The method of claim 20, wherein the building equipment comprise at least one of:

one or more individual devices comprising at least one of a heater, a chiller, an air handling unit, a variable air volume unit, a pump, a fan, a cooling tower, a valve, an actuator, a damper, and a thermal energy storage (TES) device;

a building subsystem comprising at least one of a building electrical subsystem, an information communication technology subsystem, a security subsystem, a HVAC subsystem, a lighting subsystem, a lift/escalators subsystem, a fire safety subsystem, a refrigeration subsystem, an advertising or signage subsystem, a cooking subsystem, a vending subsystem, and a printer or copy service subsystem;

one or more subplant systems comprising at least one of a heater subplant, a heat recovery chiller subplant, a chiller subplant, a cooling tower subplant, a hot TES subplant, and a cold TES subplant; and

an energy plant comprising a plurality of the subplant systems.

27. The method of claim 20, wherein the time series data represent a variable controlled to a setpoint comprising at least one of a controlled flowrate, a controlled temperature, a controlled pressure, a controlled humidity, and a controlled power consumption.