Heuristic method of automated and learning control, and building automation systems thereof
First Claim
Patent Images
1. A building system controller comprising:
- a controlled system;
a physical model, comprised of a mathematical model of the controlled system;
a simulation engine;
a heuristic;
a data stream, comprised of data from said controlled system; and
a training loop;
wherein the simulation engine can simulate the behavior of the controlled system by means of evaluating the physical model; and
whereby said training loop compares the output of the simulation engine to the data stream using the heuristic, such that the physical model is regressed in a manner that the output of the simulation engine approaches the data stream.
1 Assignment
0 Petitions
Accused Products
Abstract
Disclosed is a method of physical-model based building automation using in-situ regression to optimize control systems.
68 Citations
24 Claims
-
1. A building system controller comprising:
-
a controlled system; a physical model, comprised of a mathematical model of the controlled system; a simulation engine; a heuristic; a data stream, comprised of data from said controlled system; and a training loop; wherein the simulation engine can simulate the behavior of the controlled system by means of evaluating the physical model; and whereby said training loop compares the output of the simulation engine to the data stream using the heuristic, such that the physical model is regressed in a manner that the output of the simulation engine approaches the data stream. - View Dependent Claims (2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
3. The building system controller of claim I, wherein the regressed physical model is utilized to detect faults, by detecting divergent parameters through regression of a previously regressed model.
-
21. A electronic device or network of electronic devices comprising:
-
a coupled building system; a data stream; at least one memory; and one or more processors operatively coupled to the data stream and one or more processors operatively coupled to the memories, where the processors are configured to execute program code stored in the memories to; control the coupled building system; receive sensor measurements obtained with the data stream, and store the data he memory; compute a mathematical model that substantially imitates the behavior of the controlled system, generating a predicted data stream measurement; analyze the variation between the stored data stream and the predicted data stream, using a cost function. iterate the process of computation and analysis, while changing parameters of the mathematical model as to regress the model in a manner that the output of the simulation engine approaches the data stream; control the coupled building system using predicted actuation using the regressed mathematical model; - View Dependent Claims (22, 23, 24)
-
Specification