US 10,921,760 B2Granted

Predictive control loops using time-based simulation and building-automation systems thereof

AssigneePassiveLogic, Inc.

InventorTroy Aaron Harvey

Life of this patent

Estimating the term…

Loading patent…

Abstract

Apparatuses, systems, methods, and computer program products are presented for a building-automation system controller. A building-automation system controller manages and/or controls energy, thermal, and/or functional systems and subsystems thereof utilizing a sensor, a physical model, a simulation engine, one or more predictive control loops, an optimal cost function, and an error band. A control loop is designed to utilize a simulation engine to predict a simulated predicted sensor value of a controlled system under a simulated control regime. A simulated control regime having an optimal cost function is selected for a controlled system until the controlled system diverges from the simulated predicted sensor value beyond an error band indicating uncertainty in a predicted future behavior so that a control loop is formed utilizing a simulation engine to predict a different future behavior in response to the controlled system diverging from the simulated predicted sensor value.

Claims21

Dependent claims

A building system controller, comprising:

a controlled system;
a sensor;
a physical model, working on the principle of a mathematical model of the controlled system's behavior and characteristics;
a simulation engine, designed to simulate one or more behaviors of the controlled system by evaluating the physical model;
a control loop, designed to utilize the simulation engine to predict a future behavior of the controlled system under a simulated control regime, the future behavior comprising simulated predicted sensor values;
an optimal cost function;
and
an error band indicating an uncertainty in the simulated predicted sensor values such that the error band grows over time as predictions are made farther into the future, the error band determined based on a percentage range of a time series function of the simulated predicted sensor values with bounds ascertained from the optimal cost function;
wherein results of the simulated control regime are compared in terms of the optimal cost function, such that the simulated control regime having the optimal cost function is selected for controlling the controlled system;
and
wherein the controlled system is controlled by the simulated control regime having the optimal cost function until the controlled system diverges from the simulated predicted sensor values beyond the error band, such that the control loop is formed and the control loop utilizes the simulation engine to predict a different future behavior in response to the controlled system diverging from the simulated predicted sensor values beyond the error band.

8 more elements

A building controller comprising:

at least one sensor interface;
at least one control interface;
a weather data source;
a controlled system;
a system physical model that computes a behavior of the controlled system;
a building physical model that computes a behavior of a building under control;
a memory;
and
a processor operatively coupled to the memory, the at least one sensor interface, and the at least one control interface, and configured to execute program code stored in the memory to:
compute the building physical model using the weather data source to generate a load path;
generate a predictive control time series to compute the system physical model such that the system physical model approximates the load path, and outputting a predicted sensor time series;
actuate the controlled system via the at least one control interface using the predictive control time series until sensor data of the at least one sensor interface diverges from the predicted sensor time series beyond an error band indicating uncertainty in the predicted sensor time series and the error band grows as the predicted sensor time series makes predictions farther into the future, such that a control loop is formed, the control loop re-executing the program code to generate a different predictive control time series in response to the sensor data diverging from the predicted sensor time series beyond the error band, the error band determined based on a percentage range of the predictive control time series with bounds ascertained from an optimal cost function.

9 more elements

Dependencies are inferred from the claim text.