# US 2019/0377306 A1 · Predictive control loops using time-based simulation and building-automation systems thereof > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2019/0377306 A1 - Document ID: US20190377306A1 - Status: Published application - Kind: A1 - Application number: 16/006,713 - Filed: 2018-06-12 - Published: 2019-12-12 - Earliest claimed: 2018-06-12 - Assignees: PassiveLogic, Inc. - Applicants: PassiveLogic, Inc. - Inventors: Troy Aaron Harvey - Web page: https://patentgrep.com/patent/US20190377306A1 ## CPC - G05B 13/027: {using neural networks only} - G05B 13/048: {using a predictor} - G05B 13/0265: {the criterion being a learning criterion} - G05B 17/02: electric - G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G) ## Abstract Disclosed is a building-automation system controller for management and control of energy, thermal, and functional systems and subsystems thereof utilizing predictive control loops utilizing time-based simulation acting upon the atomic control unit triad of source-transport-sink. ## Claims ### Claim 1 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/or 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; a cost function; and an error band; wherein the results of the simulated control regime are compared in terms of the cost function, such that the simulated control regime having an optimal cost function is selected for controlling the controlled system; and wherein the building system is controlled by the simulated control regime having an optimal cost function until it diverges from the simulated predicted sensor value by the error band, such that a control loop is formed. ### Claim 2 (depends on 1) The building system controller of claim 1, wherein the error band may comprise one of the following but not limited to: an error band of at least one sensor, a mathematical function that computes the offset limits from predicted values, a percentage range of expected values, a time-series function of expected sensor values, a cost function boundary of expected sensor values, a human comfort limit of expected sensor values, and/or an equipment limit of allowed operation as determined by sensor values. ### Claim 3 (depends on 1) The building system controller of claim 1, wherein the control regime having an optimal cost function is performed by the controller, until sensor values diverge, at which point the physical model may be evaluated again so as to seek a new regime having a new optimal cost function given the changed system conditions. ### Claim 4 (depends on 1) The building system controller of claim 1, wherein the cost function may comprise a discount function based on discounted future value of the cost function. ### Claim 5 (depends on 1) The building system controller of claim 1, wherein the cost function may comprise a discount function that devalues the future cost of the control regimes, based on the accuracy or probability of past predictions. ### Claim 6 (depends on 1) The building system controller of claim 1, wherein the cost function utilizes a discount function that may devalue future system costs as compared to current cost such that future uncertainty is accounted for, including but not limited to one or more of: comfort, weather, energy, utility retail pricing, utility wholesale pricing, and/or peer-to-peer energy market prices. ### Claim 7 (depends on 1) The building system controller of claim 1, wherein the cost function may comprise but is not limited to models of one or more of: energy use, primary energy use, energy monetary cost, human comfort, the safety of building or building contents, the durability of building or building contents, microorganism growth potential, system equipment durability, system equipment longevity, environmental impact, and/or energy use CO2 potential. ### Claim 8 (depends on 1) The building system controller of claim 1, wherein the mathematical model may comprise but is not limited to one or more of: building models, human comfort models, climate models, equipment models, HVAC system models, energy systems, renewable energy systems, building envelopes, climate models, human comfort models, resource consumption models, thermodynamic models, electrical response models, heat transfer models, component wear models, convection models, conduction models, radiation models, stratification models, aging models, and/or energy or thermal transport models. ### Claim 9 (depends on 1) The building system controller of claim 1, wherein the mathematical model method may comprise but is not limited to one or more of: finite difference, finite element, finite volume, conduction transfer function, the State-Space method, the Direct Root Finding method, Explicit Finite Difference, the Implicit Finite Difference, and/or the Crank-Nicolson Finite Difference. ### Claim 10 (depends on 1) The building system controller of claim 1, wherein the cost function may compare control regimes for optimality by one or more of but is not limited to: differential comparison, multivariate population selection, statistical classification, neural nets, deep learning, reinforcement learning, deep reinforcement learning, ensemble learning, clustering, feature extraction, preference ranking, and/or benchmarking. ### Claim 11 (depends on 1) The building system controller of claim 1, wherein the physical model parameters are adjusted to better fit the actual behavior of the controlled systems by a method of regression between the simulated data and actual data of the sensors and control regimes. ### Claim 12 (depends on 11) The physical model parameters of claim 11, wherein the regression fitting method may he comprised of but is not limited to one or more of: a heuristic, metaheuristic, or hyperheuristic, including but not limited to, one or more of: particle swarm organization, self-organizing migration algorithm, neural networks, group method of data handling, differential evolution, genetic algorithm, memetic algorithm, random forest, hill climbing algorithm, simulated annealing, monte-carlo methods, random search, fuzzy-logic, arithmetic mean, geometric mean, harmonic mean, trimean, median, mode, mid-range, quadratic mean (RMS), cubic mean, generalized mean, weighted mean, linear regression, logistic regression, polynomial regression, k-means clustering, k-nearest neighbors, decision trees, perceptron, multi-layer perceptron (neural network), kernel methods, support vector machines, ensemble methods, boosting, bagging, naive Bayes, expectation maximization, Gaussian mixture models, Gaussian processes, principal component analysis, singular value decomposition, reinforcement learning, Voronoi decomposition, and/or social theory voting techniques and concepts, including but not limited to social welfare functions, social choice functions, single transferrable vote, Bucklin's rule, social decision schemes, collective utility functions, and/or Condorcet method and extensions including but not limited to Copeland's rule, maximin, Dodgson's rule, Young's rule, and/or ranked pairs. ### Claim 13 A building controller comprising: at least one sensor interface; at least one control interface; a weather source; a controlled system; a system physical model that computes the behavior of the controlled system; a building physical model that computes the behavior of the building under control; a memory; and a processor operatively coupled to the memory, sensor interface(s), control interface(s), and configured to execute program code stored in the memory to: compute the building physical model using the weather source to generate a load path; generate a predictive control time series to compute the system physical model such that it approximates the load path, and outputting a predicted sensor time series; actuate the controlled system via the control interface(s) using the predictive control time series until the sensor data of the sensor interface diverges from the computed sensor time series, whereby restarting the process from the beginning. ### Claim 14 (depends on 13) The building controller of claim 13, wherein the controlled system comprises but is not limited to one or more of the following: building automation systems, heating systems, cooling systems, ventilation systems, power management systems, renewable energy systems, irrigation systems, occupancy systems, lighting systems, environmental control systems, humidity control systems, air quality management systems, window operators, and/or shade systems. ### Claim 15 (depends on 13) The building controller of claim 13, wherein the sensor(s) of a sensor interface comprises but is not limited to one or more of: weather data, temperature sensor, flow sensor, humidity sensor, CO2 sensor, radiant temperature sensor, insolation sensor, wind sensor, occupancy sensor, rotational movement sensor, door movement sensor, rain sensor, snow sensor, light level sensor, time sensor, calendar, and/or external data feeds, data streams, and/or sensors thereof. ### Claim 16 (depends on 13) The building controller of claim 13, wherein the simulation engine comprises one or more computational entities in a distributed network of the simulation engines or remote computational entities. ### Claim 17 (depends on 13) The building controller of claim 13, wherein sensor values of sensor data comprises a mathematical fusion of a multiplicity of individual sensor values. ### Claim 18 (depends on 13) The building controller of claim 13, wherein a data feed of future data from an external source is combined with the sensor time series output of the building controller operating on the same data to provide durability of the data, such that gaps in the feed data or a lost connection do not compromise the availability of the data. ### Claim 19 (depends on 13) The building controller of claim 13, wherein sensor data or weather data of a building controller are combined with an external source of the same data, using local sensors to provide a modification of the data feed information such that it more accurately represents local conditions. ### Claim 20 (depends on 13) The building controller of claim 13, wherein the divergence from the computed time series values may comprise but is not limited to one of: an error band, a mathematical function that computes the offset limits from predicted values, a percentage range of expected values, a time-series function of expected sensor values, a cost function boundary of expected sensor values, a human comfort limit of expected sensor values, and/or an equipment limit of allowed operation as determined by sensor values ### Claim 21 (depends on 13) The building controller of claim 13, wherein the load path generated is computed via a physical model of human comfort of occupants contained within the physical model of the building. ### Claim 22 (depends on 21) The load path of claim 21, wherein the occupants of the human comfort model occupancy function are inferred via direct or indirect data feeds of occupancy activity, including but not limited to: mobile device beaconing, mobile device trilateralization, access control, computer network activity, door sensor activity, mobile device addresses on a local network, acoustic activity, visual occupant tracking, and/or thermal activity.