Building control system with smart edge devices having embedded model predictive control
A smart edge controller for building equipment that operates to affect a variable state or condition within a building. The controller includes processors and non-transitory computer-readable media storing instructions that, when executed by the processors, cause the processors to perform operations including obtaining sensor data indicating environmental conditions of the building and include determining an amount of available processing resources at the smart edge controller or at the building equipment. The operations include automatically scaling a level of complexity of an optimization of a cost function based on the available processing resources and include performing the optimization of the cost function at the automatically scaled level of complexity to generate a first setpoint trajectory. The first setpoint trajectory includes operating setpoints for the building equipment at time steps within an optimization period. The operations include operating the building equipment based on the first setpoint trajectory.
1. A smart edge controller for building equipment that operates to affect a variable state or condition within a building, the controller comprising:
one or more processors; and
one or more non-transitory computer-readable media storing instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
obtaining sensor data indicating one or more environmental conditions of the building;
determining an amount of available processing resources at the smart edge controller or at the building equipment;
automatically scaling a level of complexity of an optimization of a cost function based on the amount of available processing resources;
performing the optimization of the cost function at the automatically scaled level of complexity to generate a first setpoint trajectory for the building equipment, the first setpoint trajectory comprising operating setpoints for the building equipment at a plurality of time steps within an optimization period; and
operating the building equipment based on the first setpoint trajectory to affect the variable state or condition within the building.
2. The controller of claim 1 , the operations further comprising:
selecting a subset of the sensor data for generating a predictive model based on the amount of available processing resources; and
generating the predictive model based on the subset of the sensor data, wherein the optimization of the cost function is performed based on the predictive model.
3. The controller of claim 1 , the operations further comprising generating an active setpoint for the building equipment or for a space of the building based on the first setpoint trajectory, wherein the optimization of the cost function is performed based on a first-order thermal model describing thermal dynamics of the space of the building.
4. The controller of claim 1 , wherein scaling the level of complexity of the optimization of the cost function comprises at least one of:
reducing a number of input variables to the cost function;
reducing a number of time steps within the optimization period; or
reducing a number of decision variables for which values are generated by performing the optimization of the cost function.
5. The controller of claim 1 , wherein the available processing resources comprise at least one of:
available memory;
available clock cycles;
available energy;
available network bandwidth; or
available budget.
6. The controller of claim 1 , wherein functionality of the controller is distributed across a plurality of devices of the building.
7. The controller of claim 1 , the operations further comprising:
determining whether a connection between the controller and a cloud computation system is active; and
in response to a determination that the connection is active, obtaining a second setpoint trajectory from the cloud computation system and using the second setpoint trajectory instead of the first setpoint trajectory to operate the building equipment;
wherein performing the optimization of the cost function to generate the first setpoint trajectory occurs in response to a determination that the connection is not active.
8. An environmental control system for building equipment that operates to affect a variable state or condition within a building, the system comprising:
one or more environmental sensors configured to measure one or more environmental conditions affecting the building;
the building equipment that operates to affect the variable state or condition within the building; and
a controller comprising a processing circuit configured to:
obtain sensor data indicating the one or more environmental conditions of the building from the one or more environmental sensors;
determine an amount of available processing resources at the controller or at the building equipment;
automatically scale a level of complexity of an optimization of a cost function based on the amount of available processing resources;
perform the optimization of the cost function at the automatically scaled level of complexity to generate a first setpoint trajectory for the building equipment, the first setpoint trajectory comprising operating setpoints for the building equipment at a plurality of time steps within an optimization period; and
operate the building equipment based on the first setpoint trajectory to affect the variable state or condition within the building.
9. The system of claim 8 , wherein the processing circuit is configured to:
select a subset of the sensor data for generating a predictive model based on the amount of available processing resources; and
generate the predictive model based on the subset of the sensor data, wherein the optimization of the cost function is performed based on the predictive model.
10. The system of claim 8 , wherein the processing circuit is configured to generate an active setpoint for the building equipment or for a space of the building based on the first setpoint trajectory, wherein the optimization of the cost function is performed based on a first-order thermal model describing thermal dynamics of the space of the building.
11. The system of claim 8 , wherein scaling the level of complexity of the optimization of the cost function comprises at least one of:
reducing a number of input variables to the cost function;
reducing a number of time steps within the optimization period; or
reducing a number of decision variables for which values are generated by performing the optimization of the cost function.
12. The system of claim 8 , wherein the available processing resources comprise at least one of:
available memory;
available clock cycles;
available energy;
available network bandwidth; or
available budget.
13. The system of claim 8 , wherein functionality of the controller is distributed across a plurality of devices of the building.
14. The system of claim 8 , wherein the processing circuit is configured to:
determine whether a connection between the controller and a cloud computation system is active; and
in response to a determination that the connection is active, obtain a second setpoint trajectory from the cloud computation system and using the second setpoint trajectory instead of the first setpoint trajectory to operate the building equipment;
wherein performing the optimization of the cost function to generate the first setpoint trajectory occurs in response to a determination that the connection is not active.
15. A method for operating building equipment to affect a variable state or condition within a building, the method comprising:
obtaining sensor data indicating one or more environmental conditions of the building;
determining a capacity to perform an optimization of a cost function at a smart edge controller or at the building equipment;
automatically scaling a level of complexity of the optimization of the cost function based on the capacity;
performing the optimization of the cost function at the automatically scaled level of complexity to generate a first setpoint trajectory for the building equipment, the first setpoint trajectory comprising operating setpoints for the building equipment at a plurality of time steps within an optimization period; and
operating the building equipment based on the first setpoint trajectory to affect the variable state or condition within the building.
16. The method of claim 15 , further comprising:
selecting a subset of the sensor data for generating a predictive model based on the capacity; and
generating the predictive model based on the subset of the sensor data, wherein the optimization of the cost function is performed based on the predictive model.
17. The method of claim 15 , further comprising generating an active setpoint for the building equipment or for a space of the building based on the first setpoint trajectory.
18. The method of claim 15 , wherein scaling the level of complexity of the optimization of the cost function comprises at least one of:
reducing a number of input variables to the cost function;
reducing a number of time steps within the optimization period; or
reducing a number of decision variables for which values are generated by performing the optimization of the cost function.
19. The method of claim 15 , wherein the optimization of the cost function is performed based on a first-order thermal model describing thermal dynamics of a space of the building.
20. The method of claim 15 , further comprising:
determining whether a connection between the controller and a cloud computation system is active; and
in response to a determination that the connection is active, obtaining a second setpoint trajectory from the cloud computation system and using the second setpoint trajectory instead of the first setpoint trajectory to operate the building equipment;
wherein performing the optimization of the cost function to generate the first setpoint trajectory occurs in response to a determination that the connection is not active.