Predictive diagnostics system with fault detector for preventative maintenance of connected equipment
A building management system includes connected equipment configured to measure a plurality of monitored variables and a predictive diagnostics system configured to receive the monitored variables from the connected equipment; generate a probability distribution of the plurality of monitored variables; determine a boundary for the probability distribution using a supervised machine learning technique to separate normal conditions from faulty conditions indicated by the plurality of monitored variables; separate the faulty conditions into sub-patterns using an unsupervised machine learning technique to generate a fault prediction model, each sub-pattern corresponding with a fault, and each fault associated with a fault diagnosis; receive a current set of the monitored variables from the connected equipment; determine whether the current set of monitored variables correspond with one of the sub-patterns of the fault prediction model to facilitate predicting whether a corresponding fault will occur; and determining the fault diagnosis associated with the predicted fault.
1. A building management system comprising:
equipment configured to measure a plurality of variables; and
a predictive diagnostics system configured to:
receive the variables from the equipment;
determine a boundary for a probability distribution of the variables using a first machine learning technique to separate normal conditions from faulty conditions indicated by the variables;
separate the faulty conditions into sub-patterns using a second machine learning technique to generate a fault prediction model, wherein each sub-pattern corresponds with a fault, and wherein each fault is associated with a fault diagnosis;
receive a current set of the variables from the equipment;
determine whether the current set of the variables corresponds with one of the sub-patterns of the fault prediction model to facilitate predicting whether a corresponding fault will occur; and
determine the fault diagnosis associated with a predicted fault to facilitate taking preventative measures.
2. The building management system of claim 1 , wherein the predictive diagnostics system is configured to provide an indication of the fault diagnosis with a recommended action to facilitate manual repair of the equipment based on the fault diagnosis to prevent the predicted fault from occurring.
3. The building management system of claim 1 , wherein the predictive diagnostics system is configured to perform an automated control action based on the fault diagnosis to at least one of (i) prevent the predicted fault from occurring or (ii) compensate for the predicted fault.
4. The building management system of claim 3 , wherein the automated control action includes at least one of (i) entering the equipment into a safety mode, (ii) operating the equipment at a derated state, or (iii) shutting down the equipment.
5. The building management system of claim 3 , wherein the equipment includes an array of equipment, and wherein the automated control action includes (i) at least one of (a) shutting down or (b) derating operation of at least one of connected equipment within the array of equipment that is associated with the predicted fault and (ii) increasing operation of remaining connected equipment within the array of equipment.
6. The building management system of claim 3 , wherein the automated control action includes increasing an output of connected equipment of the equipment to preemptively compensate for an expected decrease in performance in the future based on the predicted fault.
7. The building management system of claim 1 , wherein the predictive diagnostics system is configured to receive an indication of the fault diagnosis for an unprecedented fault in response to the unprecedented fault not corresponding with one of the sub-patterns of the fault prediction model.
8. The building management system of claim 7 , wherein the predictive diagnostics system is configured to define a new sub-pattern of the fault prediction model based on the unprecedented fault.
9. A method for performing predictive diagnostics on equipment of a building system, the method comprising:
connecting a processing circuit of the building system to the equipment, the equipment including one or more sensors positioned to acquire data regarding a plurality of variables indicative of operation of the equipment;
generating, by the processing circuit, a probability distribution of the variables;
determining, by the processing circuit, a boundary for the probability distribution to separate normal conditions from faulty conditions indicated by the variables;
separating, by the processing circuit, the faulty conditions into sub-patterns to generate a fault prediction model, wherein each sub-pattern corresponds with a fault, and wherein each fault is associated with a fault diagnosis;
receiving, by the processing circuit, a current set of the variables from the connected equipment;
determining, by the processing circuit, whether the current set of the variables corresponds with one of the sub-patterns of the fault prediction model to facilitate predicting whether a corresponding fault will occur;
determining, by the processing circuit, the fault diagnosis associated with a predicted fault based on which of the sub-patterns of the fault prediction model the current set of the variables corresponds with; and
at least one of:
providing, by the processing circuit to an output device, an indication of the fault diagnosis with a recommended action to facilitate manual repair of the connected equipment by an operator based on the fault diagnosis to prevent the predicted fault; or
performing, by the processing circuit, an automated control action based on the fault diagnosis to at least one of prevent or compensate for the predicted fault.
10. The method of claim 9 , wherein the automated control action includes at least one of (i) entering the connected equipment into a safety mode, (ii) operating the connected equipment at a derated state, or (iii) shutting down the connected equipment.
11. The method of claim 9 , wherein the connected equipment includes an array of equipment, and wherein the automated control action includes (i) at least one of (a) shutting down or (b) derating operation of at least one of the connected equipment within the array of equipment associated with the predicted fault and (ii) increasing operation of remaining connected equipment within the array of equipment.
12. The method of claim 9 , wherein the automated control action includes increasing an output of the equipment to preemptively compensate for an expected decrease in performance in the future based on the predicted fault.
13. The method of claim 9 , further comprising receiving, by the processing circuit from an input device, an indication of the fault diagnosis for an unprecedented fault in response to the unprecedented fault not corresponding with one of the sub-patterns of the fault prediction model, wherein the input device is or is not the same device as the output device.
14. The method of claim 13 , further comprising defining, by the processing circuit, a new sub-pattern of the fault prediction model based on the unprecedented fault.
15. A building management system for a building, the building management system comprising:
a building subsystem comprising equipment comprising at least one of heating, ventilation, and air conditioning (HVAC) equipment, lighting equipment, or security equipment; and
a predictive diagnostics system configured to:
receive variables from the equipment;
determine a boundary for a distribution of the variables using a first machine learning technique to separate normal conditions from faulty conditions indicated by the variables;
separate the faulty conditions into sub-patterns using a second machine learning technique to generate a fault prediction model, wherein each sub-pattern corresponds with a fault, and wherein each fault is associated with a fault diagnosis;
receive a current set of the variables from the equipment;
determine whether the current set of the variables corresponds with one of the sub-patterns of the fault prediction model to facilitate predicting whether a corresponding fault will occur; and
determine the fault diagnosis associated with a predicted fault to facilitate taking preventative measures.
16. The building management system of claim 15 , wherein the predictive diagnostics system is configured to provide an indication of the fault diagnosis with a recommended action to facilitate manual repair of the equipment based on the fault diagnosis to prevent the predicted fault from occurring.
17. The building management system of claim 15 , wherein the first machine learning technique is a supervised technique and the second machine learning technique is an unsupervised technique.
18. The building management system of claim 15 , wherein the predictive diagnostics system is configured to perform an automated control action based on the fault diagnosis to at least one of (i) prevent the predicted fault from occurring or (ii) compensate for the predicted fault.
19. The building management system of claim 18 , wherein the automated control action includes at least one of (i) entering the equipment into a safety mode, (ii) operating the equipment at a derated state, or (iii) shutting down the equipment.
20. The building management system of claim 18 , wherein the equipment includes an array of equipment, and wherein the automated control action includes (i) at least one of (a) shutting down or (b) derating operation of at least one of connected equipment within the array of equipment that is associated with the predicted fault and (ii) increasing operation of remaining connected equipment within the array of equipment.