Building HVAC system with fault-adaptive model predictive control

1. A method for automatically adapting a predictive model used to control a heating, ventilation, or air conditioning (HVAC) system in a building to make adjustments based on a detected fault in the HVAC system, the method comprising:

determining that the detected fault in the HVAC system is based on a stuck actuator fault from a stuck actuator;

determining an impact of the detected fault on an operational performance of at least one of the HVAC system or a zone of the building;

adjusting one or more parameters of the predictive model based on the impact of the detected fault to generate a fault-adapted predictive model, the adjusting the one or more parameters comprising adjusting one or more constraints affecting inputs to the predictive model based on the stuck actuator; and

operating the HVAC system to control an environmental condition of the building using the fault-adapted predictive model, the operating the HVAC system comprising operating a portion of the HVAC system to provide a constant amount of heating or cooling upon determining that the detected fault is based on the stuck actuator fault from the stuck actuator.

2. The method of claim 1 , wherein the detected fault is at least one of: a fault that results in a constant cooling or heating, a fault that renders the HVAC system unavailable, a fault that reduces an available capacity, or a fault that reduces an efficiency of the HVAC system.

3. The method of claim 1 , wherein the adjusting the one or more parameters of the predictive model comprises triggering an identification process of at least one of:

one or more dynamics of the HVAC system; or

one or more dynamics of the building,

wherein the identification process is based on an accuracy error exceeding a threshold.

4. The method of claim 1 , wherein the determining the impact of the detected fault comprises calculating a thermal capacitance within a building zone, the thermal capacitance related to the operational performance of the HVAC system.

5. The method of claim 1 , wherein the one or more parameters define a set of dynamics of the HVAC system or a set of dynamics of one or more zones in the building; and

the method further comprises adjusting one or more threshold values for control variables in the HVAC system.

6. The method of claim 1 , further comprising selecting an optimal control scheme by:

determining the detected fault is a decreased efficiency type of fault and an effect of the detected fault on at least one of the HVAC system or the zone of the building, the effect comprising a tangible amount of energy added to or removed from the HVAC system; and

adjusting the one or more parameters of the predictive model to incorporate the effect of the detected fault.

7. The method of claim 1 , comprising obtaining an indication of the detected fault in the HVAC system by:

receiving data from one or more HVAC sensors, the data comprising operational information relating to at least one of the HVAC system or the zone of the building;

comparing the data to a predetermined range of normal operation; and

determining the detected fault based on the data being outside of the predetermined range of normal operation.

8. The method of claim 1 , wherein the adjusting the one or more parameters of the predictive model comprises re-identifying the predictive model by performing an event-based system identification process.

9. The method of claim 1 , further comprising performing a fault detection process to generate an indication of the detected fault, the fault detection process comprising using a neural network to detect abnormal operation of the HVAC system based on a set of training data that indicates normal operation of the HVAC system.

10. The method of claim 1 , wherein the operating the HVAC system to control the environmental condition of the building comprises using the fault-adapted predictive model to generate constraints for an optimization problem.

11. The method of claim 1 , comprising obtaining an indication of the detected fault in the HVAC system by obtaining an indication of abnormal operation in one or more building management systems (BMSs) or degradation in compressor operation or air flow restriction in an air duct or refrigerant leakage.

12. A controller for automatically adapting a predictive model used to control a heating, ventilation, or air conditioning (HVAC) system in a building to make adjustments based on a detected fault in the HVAC system, the controller comprising a processing circuit configured to:

determine an amount of constant cooling energy or heating energy provided by a portion of the HVAC system as a result of the detected fault;

adjust one or more parameters of the predictive model based on the result of the detected fault to generate a fault-adapted predictive model; and

operate the HVAC system to control an environmental condition of the building using the fault-adapted predictive model.

13. The controller of claim 12 , wherein the processing circuit is further configured to select an optimal control scheme by:

determining the detected fault is a decreased efficiency type of fault and an effect of the detected fault on the HVAC system, the effect comprising a tangible amount of energy added to or removed from the HVAC system; and

adjusting one or more parameters of the predictive model to incorporate the effect of the detected fault.

14. One or more non-transitory computer-readable storage media having computer-executable instructions stored thereon that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

determining an impact of a detected fault on an operational performance of a HVAC system, the detected fault being failure of a HVAC device of the HVAC system;

adjusting one or more values used in an optimization problem based on the impact of the detected fault to generate a fault-adapted predictive model, the adjusting comprising reducing the one or more values by an amount proportional to a capacity of the HVAC device relative to a total capacity of the HVAC system; and

operating the HVAC system to control a variable state or condition of the HVAC system using the fault-adapted predictive model.

15. A method for automatically adapting a predictive model used to control a heating, ventilation, or air conditioning (HVAC) system in a building to make adjustments based on a detected fault in the HVAC system, the method comprising:

determining that the detected fault is based on a sensor in the HVAC system;

determining an impact of the detected fault on an operational performance of at least one of the HVAC system or a zone of the building;

adjusting one or more parameters of the predictive model based on the impact of the detected fault to generate a fault-adapted predictive model by reducing the one or more parameters by an amount proportional to a capacity of a HVAC device relative to a total capacity of the HVAC system; and

operating the HVAC system to control an environmental condition of the building using the fault-adapted predictive model.