System and Method for predicting short cycling of HVAC systems and generating alerts for remediation

Abstract

A multisystem controller communicatively coupled to a first and second HVAC system includes a memory configured to store sensor data, a set of cycling times, and operating conditions data associated with the first or second HVAC system. The multisystem controller includes a processor communicatively coupled to the memory and configured to access the sensor data, the set of cycling times, and the operating conditions data, preprocess the set of cycling times, and train a predictive algorithm based on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data. The processor is configured to execute a predictive model to generate an expected cycling time, compute a difference between the expected cycling time and one or more cycling times of the set of cycling times, and generate an alert comprising an indication of a short cycling of the first or second HVAC system.

Description

TECHNICAL FIELD

[0001] The present disclosure relates generally to heating, ventilation, and air conditioning systems, and, more specifically, to a system and method for predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation.BACKGROUND

[0002] A heating, ventilation, and air conditioning (HVAC) system may be utilized generally to regulate temperature within an enclosed space. Specifically, air is cooled via heat transfer with refrigerant flowing through the HVAC system and returned to the enclosed space as conditioned air. While some enclosed spaces may include multiple HVAC systems, temperatures may nevertheless vary across enclosed spaces due to irregular operations of the HVAC systems.SUMMARY

[0003] The system and methods implemented by the system as disclosed in the present disclosure provide technical solutions to the technical problems discussed above by predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation. The disclosed system and methods provide several practical applications and technical advantages. Specifically, the present embodiments improve the power consumption and reliability of HVAC systems by predicting the short cycling of HVAC systems and flagging the HVAC systems for remediation before the short cycling is allowed to repeatedly occur and overrun the HVAC systems or cause damage to the HVAC systems. The present embodiments may further improve the power consumption and reliability of HVAC systems by correlating short cycling with indoor humidity, in that a high frequency of short cycling corresponds generally to high indoor humidity and vice-versa, thus providing an additional symptom (e.g., high indoor humidity) to be monitored for detection of short cycling of HVAC systems.

[0004] For example, short cycling occurs when an HVAC system turns “ON” (e.g., “starts up”) and turns “OFF” (e.g., “shuts down”) much more frequently than the HVAC system would otherwise under normal operating conditions. In such instances, the HVAC system may not complete a normal heating cycle or cooling cycle, leading to indoor temperature and humidity discomfort for users and greater power consumption and operations for the HVAC system (e.g., “start-up” corresponds to a time in which HVAC systems generally consumes the most power and performs the most mechanical operations).

[0005] There are many potential reasons as to why the HVAC system may be short cycling. For example, potential reasons for short cycling may include an oversized HVAC system, one or more clogged air filters of the HVAC system (e.g., an air filter of the HVAC system may experience poor air flow due to being full of dust or debris, etc.), a malfunctioning thermostat associated with the HVAC system, low refrigerant levels for the HVAC system, or a blocked evaporator coil of the HVAC system. Thus, without the presently disclosed embodiments for predicting short cycling of HVAC systems and generating alerts for remediation of the HVAC systems predicted to be short cycling, the HVAC system may frequently and repeatedly short cycle, which may lead to excessive power consumption of the HVAC system and potential damage to the compressor and motor of the HVAC system. Accordingly, it may be useful to provide techniques to predict short cycling of HVAC systems and generate alerts for remediation, and further to generate clusters of short cycles of HVAC systems and correlating short cycles with indoor humidity.

[0006] In accordance with the presently disclosed embodiments, a system includes a first heating, ventilation, and air conditioning (HVAC) system configured to regulate temperature of one of a first space or a second space of a building and a second HVAC system configured to regulate temperature of one of the first space or the second space of the building. In one embodiment, each of the first HVAC system and the second HVAC may be configured to regulate the temperature of a different one of the first space or the second space of the building. In particular embodiments, the system further includes a multisystem controller communicatively coupled to the first HVAC system and the second HVAC system.

[0007] In particular embodiments, the multisystem controller may include a memory configured to store sensor data, a set of cycling times, and operating conditions data associated with one or more of the first HVAC system or the second HVAC system. In particular embodiments, the multisystem controller may further include a processor communicatively coupled to the memory and configured to access the sensor data, the set of cycling times, and the operating conditions data associated with the one or more of the first HVAC system or the second HVAC system. In particular embodiments, the processor may be further configured to preprocess the set of cycling times. In one embodiment, the processor may be configured to preprocess the set of cycling times by normalizing the set of cycling times to compensate for temperature variations and physical dimensions of the one or more of the first HVAC system or the second HVAC system.

[0008] In particular embodiments, the processor may be further configured to train a predictive algorithm based at least in part on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system or the second HVAC system. In particular embodiments, the processor may be further configured to execute, based at least in part on the sensor data, the preprocessed set of cycling times, and the operating conditions data, and, in response to the training of the predictive algorithm, a predictive model to generate an expected cycling time for the one or more of the first HVAC system or the second HVAC system.

[0009] In particular embodiments, the processor may be further configured to output, by the predictive model, the expected cycling time for the one or more of the first HVAC system or the second HVAC system. In particular embodiments, the processor may be further configured to compute a difference between the expected cycling time and one or more cycling times of the set of cycling times. In particular embodiments, in response to determining that the difference between the expected cycling time and the one or more cycling times is greater than or equal to a predetermined cycling time threshold, the processor may be further configured to generate an alert comprising an indication of a short cycling of the one or more of the first HVAC system or the second HVAC system and instructions for remediating the one or more of the first HVAC system or the second HVAC system.

[0010] For example, in one embodiment, the processor may be configured to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold by comparing the expected cycling time to the standard deviation of the normal cycling time. In one embodiment, the processor may be configured to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold when the expected cycling time is an N number of standard deviations less than the mean cycling time.

[0011] In one embodiment, the alert may include a first alert. In particular embodiments, in response to determining that the difference between the expected cycling time and the one or more cycling times is less than the predetermined cycling time threshold, the processor may be further configured to generate a second alert comprising an indication of a normal cycling of the one or more of the first HVAC system or the second HVAC system. In particular embodiments, the processor may be configured to generate, based at least in part on one or more of the historical sensor data, the historical cycling time data, or the historical operating conditions data, normal cycling times for the one or more of the first HVAC system or the second HVAC system, and further to compute, based at least in part on the normal cycling times, a mean cycling time of the normal cycling times and a standard deviation of the normal cycling times.

[0012] In particular embodiments, the processor may be configured to store a log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system, and further to update the predictive model based at least in part on the log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system. For example, in one embodiment, the processor may be configured to execute the updated the predictive model to identify whether the short cycling of the one or more of the first HVAC system or the second HVAC system is one of a random short cycling occurrence or one of a systemic short cycling occurrence.BRIEF DESCRIPTION OF THE DRAWINGS

[0013] For a more complete understanding of this disclosure, reference is now made to the following brief description, taken in connection with the accompanying drawings and detailed description, wherein like reference numerals represent like parts.

[0014] FIG. 1 is a block diagram of first and second heating, ventilation, and air conditioning (HVAC) systems and a multisystem controller, in accordance with one or more embodiments of the present disclosure;

[0015] FIG. 2 illustrates a workflow diagram of an algorithm for predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation, in accordance with one or more embodiments of the present disclosure;

[0016] FIG. 3 illustrates a plot diagram of a model for generating clusters of short cycles of heating, ventilation, and air conditioning (HVAC) systems and correlating short cycles with indoor humidity, in accordance with one or more embodiments of the present disclosure;

[0017] FIG. 4 illustrates a flowchart of an example method for predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation, in accordance with one or more embodiments of the present disclosure; and

[0018] FIG. 5 illustrates a flowchart of an example method for generating clusters of short cycles of heating, ventilation, and air conditioning (HVAC) systems and correlating short cycles with indoor humidity, in accordance with one or more embodiments of the present disclosure.DETAILED DESCRIPTIONExample System

[0019] FIG. 1 is a block diagram of multiunit heating, ventilation, and air conditioning (HVAC) system 100. In particular embodiments, the system 100 may include a multisystem controller 102, a first HVAC system 103A communicatively coupled to the multisystem controller 102, and a second HVAC system 103B communicatively coupled to the multisystem controller 102. In particular embodiments, the multisystem controller 102 may also be communicatively coupled to a utility-provider and / or third-party service 128, which may provide device identification (ID) data 131 and geolocation data 130 to the multisystem controller 102 intermittingly or whenever requested as discussed in greater detail below.

[0020] In particular embodiments, the first HVAC system 103A may be disposed inside, about, or nearby one of a downstairs space 100A or an upstairs space 100B of a multistory building, such as multistory house, a multistory complex, or other similar multistory building in which a user may dwell. Similarly, the second HVAC system 103B may be disposed inside, about, or nearby one of the downstairs space 100A or the upstairs space 100B of the multistory building. For example, in some embodiments, the first HVAC system 103A may be utilized to regulate temperature or humidity of one of the downstairs space 100A or the upstairs space 100B of the multistory building and the second HVAC system 103B may be utilized to regulate temperature or humidity of the other one of the downstairs space 100A or the upstairs space 100B of the multistory building.

[0021] In general, in one embodiment, the multisystem controller 102 may be utilized to access sensor data 126, a set of cycling times 127, and operating conditions data 129 associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and preprocess the set of cycling times 127. In one embodiment, preprocessing the set of cycling times 127 may include normalizing the set of cycling times 127 to scale for temperature variations and physical dimensions of the one or more of the first HVAC system 103A or the second HVAC system 103B. In particular embodiments, the multisystem controller 102 may then train a predictive algorithm based at least in part on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and a plurality of other HVAC systems 103A, 103B.

[0022] In particular embodiments, the multisystem controller 102 may then execute, based at least in part on the sensor data 126, the preprocessed set of cycling times 127, and the operating conditions data 129, and, in response to the training of the predictive algorithm, a predictive model 122 to generate an expected cycling time 124 for the one or more of the first HVAC system 103A or the second HVAC system 103B. In particular embodiments, the multisystem controller 102 may then output, by the predictive model 122, the expected cycling time 124 for the one or more of the first HVAC system 103A or the second HVAC system 103B and compute a difference between the expected cycling time 124 and one or more cycling times 127 of the set of cycling times 127. In response to determining that the difference between the expected cycling time 124 and the one or more cycling times 127 is greater than or equal to a predetermined cycling time threshold, the multisystem controller 102 may then generate an alert comprising an indication of a short cycling of the one or more of the first HVAC system 103A or the second HVAC system 103B and instructions for remediating the one or more of the first HVAC system 103A or the second HVAC system 103B.

[0023] In another embodiment, the multisystem controller 102 may be utilized to access a set of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and train a clustering algorithm based at least in part on a training data set of an aggregate of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and a plurality of other HVAC systems 103A, 103B. In particular embodiments, the multisystem controller 102 may then execute, based at least in part on the set of short cycling times, and, in response to the training of the clustering algorithm, a clustering model 133 to generate a plurality of clusters 302, 304, and 306 of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B. In one embodiment, each cluster of the plurality of clusters 302, 304, and 306 of short cycling times includes a frequency of short cycles.

[0024] In particular embodiments, the multisystem controller 102 may then output, by the clustering model 133, the plurality of clusters 302, 304, and 306 of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and estimate, based at least in part on the plurality of clusters 302, 304, and 306 of short cycling times, a level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B). In one embodiment, the level of indoor humidity is approximately proportional to the frequency of short cycles. In particular embodiments, in response to estimating the level of indoor humidity associated with one of the first space (e.g., downstairs space 100A) or the second space (e.g., upstairs space 100B), the multisystem controller 102 may then instruct the one or more of the first HVAC system 103A or the second HVAC system 103B to operate in accordance with the level of indoor humidity associated with one of the first space (e.g., downstairs space 100A) or the second space (e.g., upstairs space 100B).Multisystem Controller

[0025] In particular embodiments, the multisystem controller 102 may access or receive from the utility provider and / or third-party service 128 geolocation data 130 and device ID data 131 and utilize the geolocation data 130 and device ID data 131 to ascertain that the first HVAC system 103A and the second HVAC system 103B are within the same multistory building.

[0026] For example, in one embodiment, the geolocation data 130 may include a set of geographic coordinates (e.g., longitude and latitude measurements expressed in units of degrees) indicative of the physical location of the multistory building or multistory house of a user. In another embodiment, the geolocation data 130 may include global positioning system (GPS) data that may be received by the multisystem controller 102 from the first HVAC system 103A and the second HVAC system 103B. For example, in one embodiment, the multisystem controller 102 may utilize the device ID data 131 to identify and ping the first HVAC system 103A and the second HVAC system 103B and then request a thermostat associated with the first HVAC system 103A and the second HVAC system 103B to provide the GPS data.

[0027] In particular embodiments, the multisystem controller 102 may then compare the set of geographic coordinates (e.g., longitude and latitude values) indicative of the physical location of the multistory building with the GPS data received from the first HVAC system 103A and the second HVAC system 103B. In particular embodiments, in response to the multisystem controller 102 determining that the set of geographic coordinates (e.g., longitude and latitude values) matches with the GPS data received from the first HVAC system 103A and the second HVAC system 103B, the multisystem controller 102 may then determine that the first HVAC system 103A and the second HVAC system 103B are present within the same multistory building.

[0028] In some embodiments, the multisystem controller 102 may determine a power consumption 114 for operating the HVAC system 110 at a temperature or humidity setpoint 112 (e.g., corresponding to current or scheduled setpoint 134A, 134B of the corresponding HVAC system 103A, 103B). For example, the power consumption 114 may be determined based on a predefined relationship between the temperature or humidity setpoint 112 and outdoor temperature or humidity. In some cases, this relationship may be specific to the HVAC system, such that the first HVAC system 103A may have a different power consumption 114 than the second HVAC system 103B for the same temperature setpoint 112.

[0029] In some embodiments, the multisystem controller 102 may control the first HVAC system 103A and the second HVAC system 103B to operate in accordance with automatically generated and defined temperature or humidity setpoints 112, power consumption 114, occupancy 116, temperature / humidity data 120, sensor data 126, the set of cycling times 127, and / or the operating conditions data 129. In other embodiments, the multisystem controller 102 may control the first HVAC system 103A and the second HVAC system 103B to operate in accordance with one or more user inputs 118 (e.g., user inputs 136A, 136B), which may be received, for example, via a thermostat associated with the respective first and second HVAC systems 103A, 103B.

[0030] In particular embodiments, the sensor data 126 may include, for example, indoor temperature of the downstairs space 100A and the upstairs space 100B, ambient temperature of the downstairs space 100A and the upstairs space 100B, indoor humidity data of the downstairs space 100A and the upstairs space 100B, ambient humidity of the downstairs space 100A and the upstairs space 100B, pressure data associated with the first HVAC system 103A or the second HVAC system103B, load data associated with the first HVAC system 103A or the second HVAC system 103B, or other similar sensor data 126 that may be associated with the first HVAC system 103A or the second HVAC system 103B.

[0031] In particular embodiments, the operating conditions data 129 may include, for example, minute-level status of cooling demand associated with the first HVAC system 103A and the second HVAC system 103B, minute-level heating demand associated with the first HVAC system 103A or the second HVAC system 103B, temperature setpoints for the first HVAC system 103A or the second HVAC system 103B, humidity setpoints for the first HVAC system 103A or the second HVAC system 103B, temperature setpoint deviations for the first HVAC system 103A or the second HVAC system 103B, humidity setpoint deviations for the first HVAC system 103A or the second HVAC system 103B, demand response times, device ID data 131 of the first HVAC system 103A or the second HVAC system 103B, and the external weather associated with the building in which the first HVAC system 103A and the second HVAC system 103B are located.

[0032] In particular embodiments, the set of cycling times 127 may include an aggregate of cycling times of the first HVAC system 103A or the second HVAC system 103B that may be computed (at functional block 204 of the algorithm 200) by the multisystem controller 102. For example, in one embodiment, the multisystem controller 102 may compute the cycling times of the first HVAC system 103A or the second HVAC system 103B utilizing one or more lag functions and functions to identify the points in time in which the operational demand status or load status of the first HVAC system 103A or the second HVAC system 103B switches from “OFF” to “ON” (e.g., “starts up”) and from “ON” to “OFF” (e.g., “shuts down”).

[0033] The multisystem controller 102 includes a processor 104, memory 106, and input / output (I / O) interface 108. The processor 104 may include one or more processors operably coupled to the memory 106. The processor 104 is any electronic circuitry including, but not limited to, state machines, one or more central processing unit (CPU) chips, logic units, cores (e.g., a multi-core processor), field-programmable gate array (FPGAs), application specific integrated circuits (ASICs), or digital signal processors (DSPs) that communicatively couples to memory 106 and controls the operation of HVAC system 103A, 103B. The processor 104 may be a programmable logic device, a microcontroller, a microprocessor, or any suitable combination of the preceding. The processor 104 is communicatively coupled to and in signal communication with the memory 106.

[0034] The processor 104 may be utilized to process data and may be implemented in hardware or software. For example, the processor 104 may be 8-bit, 16-bit, 32-bit, 64-bit or of any other suitable architecture. The processor 104 may include an arithmetic logic unit (ALU) for performing arithmetic and logic operations, processor registers that supply operands to the ALU and store the results of ALU operations, and a control unit that fetches instructions from memory 106 and executes them by directing the coordinated operations of the ALU, registers, and other components. The processor may include other hardware and software that operates to process information, control the HVAC systems 103A, 103B, and perform any of the functions described herein (e.g., with respect to FIG. 4). The processor 104 is not limited to a single processing device and may encompass multiple processing devices.

[0035] The memory 106 may include one or more disks, tape drives, or solid-state drives, and may be used as an over-flow data storage device, to store programs when such programs are selected for execution, and to store instructions and data that are read during program execution. The memory 106 may be volatile or non-volatile and may include ROM, RAM, ternary content-addressable memory (TCAM), dynamic random-access memory (DRAM), and static random-access memory (SRAM). The memory 106 is operable to store any suitable set of instructions, logic, rules, and / or code for executing the functions described in this disclosure with respect to FIGS. 1-5.

[0036] The memory 106 may store for each the respective first and second HVAC systems 103A, 103B, a temperature or humidity setpoint 112, power consumption 114, the occupancy 116 (e.g., occupancies 140A, 140B of HVAC systems 103A, 103B), a user input 118, temperature / humidity data 120 (e.g., temperature / humidity data 138A, 138B of HVAC systems 103A, 103B), sensor data 126, the set of cycling times 127, and / or the operating conditions data 129. In particular embodiments, the memory 106 may further store the predictive model 122 utilized to generate an expected cycling time 124 for the first HVAC system 103A or the second HVAC system 103B in response to the training of a predictive algorithm. In particular embodiments, the memory 106 may further store the clustering model 133 utilized to cluster (e.g., clusters of short cycles 135) or correlate the short cycling of the first HVAC system 103A or the second HVAC system 103B with indoor humidity.

[0037] The I / O interface 108 is configured to communicate data and signals with other devices. For example, the I / O interface 108 may be configured to communicate electrical signals with the HVAC systems 103A, 103B and / or the components of the HVAC systems 103A, 103B. The I / O interface 108 may send signals that cause the a staging schedule to be implemented by the HVAC systems 103A, 103B. The I / O interface 108 may use any suitable type communication protocol. The I / O interface 108 may include ports and / or terminals for establishing signal communications between a of each HVAC system 103A, 103B and other devices. The I / O interface 108 may be configured to enable wired and / or wireless communications.Utility Provider / Third-Party Service

[0038] The utility provider and / or third-party service 128 is generally an entity tasked with overseeing and / or regulating energy consumption by first and second HVAC systems 103A, 103B and may further store and housed device ID data 131 associated with first and second HVAC systems 103A, 103B. For example, the utility provider and / or third-party service 128 may be a company or organization that distributes energy to homes and businesses. In another embodiment, the utility provider and / or third-party service 128 may include a service suitable for providing geolocation data 130 indicative of the physical location of the multistory building or multistory house of a user.First and Second HVAC Systems

[0039] The system 100 includes at least two HVAC systems 103A, 103B. For clarity and conciseness only two HVAC systems, first HVAC system 103A and second HVAC system 103B, are illustrated in FIG. 1. However, the system 100 could include three or more HVAC systems 103A, 103B. Each HVAC system 103A, 103B provides conditioned air (e.g., “services”) a corresponding portion of a space. For example, the first HVAC system 103A may provide conditioned air to a portion of a room or rooms in a home or other building, and the second HVAC system 103B may provide conditioned air to other room or rooms in the same home or building. Each HVAC system 103A, 103B is associated with a power schedule 132A, 132B which indicates times when the HVAC systems 103A, 103B will be turned “ON” (e.g., allowed to provide cooling or heating) and turned “OFF” (e.g., not allowed to provide cooling or heating). Turning “OFF” an HVAC system 103A, 103B generally corresponds to turning “OFF” or not powering a compressor or heating element of the HVAC systems 103A, 103B.

[0040] Each HVAC system 103A, 103B may be associated with a temperature or humidity setpoint 134A, 134B indicating target temperatures or humidities that the first HVAC system 103A and the second HVAC system 103B may attempt to reach in the future. Each the first HVAC system 103A and the second HVAC system 103B may be operable to receive one or more user inputs 118 via a thermostat associated therewith. Such user inputs may be provided to the multisystem controller 102 and stored as user inputs 118. In other cases, user inputs 118 may be received via a user interface of the multisystem controller 102 itself or any other appropriate device.Predicting Short Cycling of Heating, Ventilation, and air Conditioning (HVAC) Systems and Generating Alerts for Remediation

[0041] Embodiments of the present disclosure discuss techniques for predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation.

[0042] FIG. 2 illustrates a workflow diagram of an algorithm 200 for predicting short cycling of HVAC systems and generating alerts for remediation, in accordance with one or more embodiments of the present disclosure. In particular embodiments, the workflow of the algorithm 200 may be performed utilizing the multisystem controller 102 as described above with respect to FIG. 1. In particular embodiments, the workflow of the algorithm 200 may begin (at functional block 202) with the multisystem controller 102 collecting one or more of sensor data 126, a set of cycling times 127, and operating conditions data 129 associated with one or more of the first HVAC system 103A or the second HVAC system 103B.

[0043] For example, in one embodiment, the multisystem controller 102 may access the sensor data 126, which may include, for example, indoor temperature of the downstairs space 100A and the upstairs space 100B, ambient temperature of the downstairs space 100A and the upstairs space 100B, indoor humidity data of the downstairs space 100A and the upstairs space 100B, ambient humidity of the downstairs space 100A and the upstairs space 100B, pressure data associated with the first HVAC system 103A or the second HVAC system 103B, load data associated with the first HVAC system 103A or the second HVAC system 103B, or other similar sensor data 126 that may be associated with the first HVAC system 103A or the second HVAC system 103B.

[0044] Similarly, the operating conditions data 129 may include, for example, minute-level status of cooling demand associated with the first HVAC system 103A and the second HVAC system 103B, minute-level heating demand associated with the first HVAC system 103A or the second HVAC system 103B, temperature setpoints for the first HVAC system 103A or the second HVAC system 103B, humidity setpoints for the first HVAC system 103A or the second HVAC system 103B, temperature setpoint deviations for the first HVAC system 103A or the second HVAC system 103B, humidity setpoint deviations for the first HVAC system 103A or the second HVAC system 103B, demand response times, device ID data 131 of the first HVAC system 103A or the second HVAC system 103B, and the external weather associated with the building in which the first HVAC system 103A and the second HVAC system 103B are located.

[0045] In particular embodiments, the set of cycling times 127 may include an aggregate of cycling times of the first HVAC system 103A or the second HVAC system 103B that may be computed (e.g., calculated at functional block 204 of the algorithm 200) by the multisystem controller 102. For example, in one embodiment, the multisystem controller 102 may compute the cycling times of the first HVAC system 103A or the second HVAC system 103B utilizing one or more lag functions and functions to identify the points in time in which the operational demand status or load status of the first HVAC system 103A or the second HVAC system 103B switches from “OFF” to “ON” (e.g., “starts up”) and from “ON” to “OFF” (e.g., “shuts down”).

[0046] In particular embodiments, the multisystem controller 102 may then compute a duration between the points in time corresponding to the first HVAC system 103A or the second HVAC system 103B switching from “OFF” to “ON” (e.g., “starts up”) and from “ON” to “OFF” (e.g., “shuts down”) to determine individual cycling times 127 for the first HVAC system 103A or the second HVAC system 103B. In particular embodiments, the multisystem controller 102 may then aggregate (e.g., sum) the durations of all of the cycling times 127 within a given period to aggregate the total cycling times for both the heating cycle and cooling cycle of the first HVAC system 103A or the second HVAC system 103B.

[0047] In particular embodiments, the workflow of the algorithm 200 may then continue (at functional block 206) with the multisystem controller 102 calculating a set of derived variables based on the durations of the individual cycling times 127 for the first HVAC system 103A or the second HVAC system 103B. For example, in one embodiment, the set of derived variables may include a normalization (e.g., scaling) of the individual cycling times 127 to scale for temperature variations and the respective specifications (e.g., physical dimensions, size, type, manufacturer, make and model, and so forth) of the first HVAC system 103A or the second HVAC system 103B. In other embodiments, the set of derived variables may include one or more statistical baselines establishing what is just the normal cycling times for the first HVAC system 103A or the second HVAC system 103B. For example, in one embodiment, the multisystem controller 102 may compute of a mean (e.g., “μ”) of the normal cycling times and a standard deviation (e.g., “σ”) of the normal cycling times.

[0048] In particular embodiments, the workflow of the algorithm 200 may then continue (at functional block 208) with the multisystem controller 102 training a predictive algorithm (e.g., building a model using an artificial-intelligence (AI) algorithm, a machine-learning algorithm, a statistical algorithm, and so forth) based on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and a number of other HVAC systems 103A, 103B that may be installed within buildings or houses of different users across a wide geographical region (e.g., a city, a municipality, a province, a country). For example, in one embodiment, the training data set of historical sensor data, historical cycling time data, and historical operating conditions data may include historical data associated with or sourced from the residential cold climate heat pump (CCHP) challenge available from the U.S. Department of Energy.

[0049] In particular embodiments, the workflow of the algorithm 200 may then continue (at functional block 210) with the multisystem controller 102 executing a predictive model 122 to generate an expected or predicted cycling time 124 for the first HVAC system 103A or the second HVAC system 103B in response to the training of the predictive algorithm. For example, in one embodiment, the predictive model 122 may include an artificial-intelligence (AI) model (e.g., one or more generative AI models); a machine-learning model, such as a regression model (e.g., linear regression model, logistic regression model) a gradient boosting model (e.g., adaptive boosting (AdaBoost) model, extreme gradient boosting (XGBoost) model, light gradient boosted machine (LightGBM) model, and so forth), or a neural network (e.g., convolutional neural network (CNN), deep neural network (DNN)); a statistical algorithm; or other similar predictive model 122 that may be suitable for generating an expected cycling time 124 for the first HVAC system 103A or the second HVAC system 103B based on the sensor data 126, the set of cycling times 127, and the operating conditions data 129.

[0050] In particular embodiments, upon the predictive model 122 outputting the expected cycling time 124 for the first HVAC system 103A or the second HVAC system 103B, the workflow of the algorithm 200 may then continue (at functional block 212) with the multisystem controller 102 computing a difference (e.g., calculating residuals) between the expected cycling time 124 for the first HVAC system 103A or the second HVAC system 103B and one or more individual cycling times 127 (e.g., actual or “real-world” cycling times) for the first HVAC system 103A or the second HVAC system 103B. Specifically, in one embodiment, the multisystem controller 102 may subtract the one or more individual cycling times 127 from the expected cycling time 124 as outputted by the predictive model 122.

[0051] In particular embodiments, the workflow of the algorithm 200 may then continue (at functional block 214) with the multisystem controller 102 determining whether the difference between the expected cycling time 124 and the one or more individual cycling times 127 is greater than or equal to a predetermined cycling time threshold as an indication of whether there is anomaly with respect to the one or more individual cycling times 127 (e.g., actual or “real-world” cycling times) for the first HVAC system 103A or the second HVAC system 103B. For example, in particular embodiments, the multisystem controller 102 may determine whether the difference between the expected cycling time 124 and the one or more individual cycling times 127 is greater than or equal to the predetermined cycling time threshold by comparing the expected cycling time 124 to the standard deviation (e.g., “σ”) of the normal cycling times. Specifically, in one embodiment, the multisystem controller 102 may determine that the difference between the expected cycling time 124 and the one or more individual cycling times 127 is greater than or equal to the predetermined cycling time threshold when the expected cycling time 124 is an N number (e.g., 2 or 3) of standard deviations (e.g., “σ”) less than the mean (e.g., “μ”) cycling time.

[0052] In one embodiment, in response to the multisystem controller 102 determining that the difference between the expected cycling time 124 and the one or more individual cycling times 127 is less than the predetermined cycling time threshold (e.g., at functional block 214), the workflow of the algorithm 200 may then conclude (at functional block 216) with the multisystem controller 102 generating an alert including an indication of a normal cycling (e.g., anomaly is not detected) of the first HVAC system 103A or the second HVAC system 103B. On the other hand, in response to the multisystem controller 102 determining that the difference between the expected cycling time 124 and the one or more individual cycling times 127 is greater than or equal to the predetermined cycling time threshold (e.g., at functional block 214), the workflow of the algorithm 200 may then conclude (at functional blocks 218 and 220) with the multisystem controller 102 generating an alert including an indication of a short cycling (e.g., anomaly is detected) of the first HVAC system 103A (e.g., at functional block 218) or the second HVAC system 103B and instructions for remediating (e.g., at functional block 220) the first HVAC system 103A or the second HVAC system 103B.

[0053] For example, in one embodiment, the alert including the indication of a short cycling of the first HVAC system 103A or the second HVAC system 103B and the instructions for remediating the first HVAC system 103A or the second HVAC system 103B may include a flagging or a service ticketing of the first HVAC system 103A or the second HVAC system 103B for a repair service to be performed by a technician. In another embodiment, the alert may include some indication of the reason for the short cycling, such as an indication that the first HVAC system 103A or the second HVAC system 103B is oversized, one or more air filters of the first HVAC system 103A or the second HVAC system 103B is potentially clogged, a thermostat associated with the first HVAC system 103A or the second HVAC system 103B has potentially malfunctioned or misoperated, refrigerant levels of the first HVAC system 103A or the second HVAC system 103B may be potentially too low, or an evaporator coil of the first HVAC system 103A or the second HVAC system 103B may be blocked (e.g., an evaporator coil may experience poor air flow due to a build up of dust, debris, ice, etc.).

[0054] In particular embodiments, the multisystem controller 102 may further accumulate and store a log of the frequency of short cycling of the first HVAC system 103A or the second HVAC system 103B and utilize the log of the frequency of short cycling of the first HVAC system 103A or the second HVAC system 103B to update (e.g., fine-tune) the predictive model 122 to identify and differentiate between random or one-off instances of short cycling of the first HVAC system 103A or the second HVAC system 103B and more systemic instances of short cycling of the first HVAC system 103A or the second HVAC system 103B.

[0055] FIG. 3 illustrates a plot diagram 300 of a model for generating clusters of short cycles of heating, ventilation, and air conditioning (HVAC) systems and predicting impacts of short cycles on indoor humility, in accordance with one or more embodiments of the present disclosure. Specifically, upon the multisystem controller 102 predicting the short cycling of the first HVAC system 103A or the second HVAC system 103B, the multisystem controller 102 may then utilize a clustering model 133 to cluster or correlate the short cycling of the first HVAC system 103A or the second HVAC system 103B with indoor humidity, in that a high frequency of short cycling corresponds generally to high indoor humidity and vice-versa.

[0056] For example, in particular embodiments, the multisystem controller 102 may access a log of short cycling times for the first HVAC system 103A or the second HVAC system 103B and train a clustering algorithm based on a training data set of an aggregate of short cycling times for the first HVAC system 103A or the second HVAC system 103B and a number of other HVAC systems. For example, in one embodiment, the training data set of the aggregate of short cycling times may include a publicly available unlabeled training data set or a proprietary unlabeled training data set sourced from a large number of HVAC systems 103A, 103B that may be installed within buildings or houses of different users across a wide geographical region (e.g., a city, a municipality, a province, a country).

[0057] As generally depicted by the plot diagram 300 of FIG. 3, in particular embodiments, the multisystem controller 102 may then execute the clustering model 133 to generate a number of clusters 302, 304, and 306 of short cycling times for the first HVAC system 103A or the second HVAC system 103B in response to the training of the clustering algorithm. Specifically, the multisystem controller 102 may utilize the clustering model 133 to output the number of clusters 302, 304, and 306 of short cycling times for the first HVAC system 103A or the second HVAC system 103B. In particular embodiments, each of the number of clusters 302, 304, and 306 may include a frequency of short cycles.

[0058] In particular embodiments, the clustering model 133 may include a k-means clustering model trained to generate a k number of clusters (e.g., 3 clusters) of short cycling times for the first HVAC system 103A or the second HVAC system 103B. Specifically, in accordance with the presently disclosed embodiments, the clustering model 133 may include any clustering algorithm or model suitable for dividing data points (e.g., HVAC systems 103A, 103B identified as short cycling and represented by “X”) into different predefined k clusters 302, 304, and 306, in which a datapoint in each cluster 302, 304, and 306 belongs only to that respective cluster 302, 304, and 306. Thus, in accordance with the presently disclosed embodiments, each cluster 302, 304, and 306 includes data points which share similarities with one another, and it follows that the data points in different clusters 302, 304, and 306 are dissimilar to those in other clusters 302, 304, and 306.

[0059] In particular embodiments, the multisystem controller 102 may execute the clustering model 133 by selecting a value of k (e.g., 3) for the k number of clusters 302, 304, and 306 of short cycling times. In one embodiment, a user may select the value of k (e.g., 3) for the k number of clusters 302, 304, and 306. In particular embodiments, the multisystem controller 102 may further execute the clustering model 133 by randomly selecting k short cycle times from the unlabeled aggregate of short cycling times, in which each of the k short cycling times includes a cluster centroid.

[0060] For example, in one embodiment, the cluster centroid of each of the k number of clusters 302, 304, and 306 may include a mean of all data points (e.g., HVAC systems 103A, 103B identified as short cycling and represented by “X”) assigned to that respective cluster 302, 304, and 306. In particular embodiments, the multisystem controller 102 may further execute the clustering model 133 by assigning each other short cycling time from the unlabeled aggregate of short cycling times to the cluster centroid nearest thereto and updating the position of each cluster centroid until each cluster centroid includes an average of the short cycling times assigned thereto.

[0061] For example, in one embodiment, the cluster 302 may correspond to HVAC systems 103A, 103B that experience a high frequency of short cycles (e.g., between 10,000 and 20,000 short cycles) and an indoor humidity range from a moderate level to a high level. The plot diagram 300 thus illustrates that such HVAC systems 103A, 103B within the cluster 302 experience frequent short cycles may generally not allow sufficient time for humidity removal, leading to higher indoor humidity levels.

[0062] Similarly, the cluster 304 may correspond to HVAC systems 103A, 103B that experience a low frequency to moderate frequency of short cycles (e.g., a few dozen short cycles to a few hundred short cycles) and an indoor humidity range from a very low level to a moderately low level. The plot diagram 300 thus illustrates that such HVAC systems 103A, 103B within the cluster 304 may generally operate more efficiently in terms of humidity control, as the lower frequency of short cycles allows for more effective moisture removal during each cycle. The cluster 304 thus illustrates that fewer short cycles contribute to better humidity management and improved efficiency and reliability of HVAC systems 103A, 103B.

[0063] In particular embodiments, the cluster 306 may correspond to HVAC systems 103A, 103B that experience a mixed frequency of short cycles (e.g., from a few dozen short cycles up to 10,000 short cycles) and a corresponding varying indoor humidity range. The cluster 306 may include potentially anomalous HVAC systems 103A, 103B or outlier HVAC systems 103A, 103B that may be influenced by factors not captured solely by short cycle count and humidity measurements, such as differences in insulation, installation environments, and so forth.

[0064] In particular embodiments, based on the clustering model 133 ability to cluster or correlate the short cycling of the first HVAC system 103A or the second HVAC system 103B with indoor humidity, in which the level of indoor humidity is approximately proportional to the frequency of short cycles, the multisystem controller 102 may estimate a level of indoor humidity for the downstairs space 100A or the upstairs space 100B. In particular embodiments, in response to estimating the level of indoor humidity for the downstairs space 100A or the upstairs space 100B, the multisystem controller 102 may instruct the one or more of the first HVAC system 103A or the second HVAC system 103B to operate in accordance with the estimated level of indoor humidity for the downstairs space 100A or the upstairs space 100B.

[0065] For example, in one embodiment, in response to estimating the level of indoor humidity for the downstairs space 100A or the upstairs space 100B, the multisystem controller 102 may generate an alert including instructions for remediating the first HVAC system 103A or the second HVAC system 103B to reduce the frequency of short cycles. In one embodiment, the alert may include an indication that the first HVAC system 103A or the second HVAC system 103B is oversized or that one or more air filters of the first HVAC system 103A or the second HVAC system 103B is potentially clogged.

[0066] FIG. 4 illustrates a flowchart of an example method 400 for predicting short cycling of heating, ventilation, and air conditioning (HVAC) systems and generating alerts for remediation, in accordance with one or more embodiments of the present disclosure. The method 400 may be performed utilizing the processor 104 of the multisystem controller 102 as described above with respect to FIG. 1. The method 400 may begin at block 402 with the processor 104 accessing sensor data 126, a set of cycling times 127, and operating conditions data 129 associated with one or more of the first HVAC system 103A or the second HVAC system 103B.

[0067] The method 400 may continue at block 404 with the processor 104 preprocessing the set of cycling times 127 by normalizing the set of cycling times 127 to scale for temperature variations and physical dimensions of the one or more of the first HVAC system 103A or the second HVAC system 103B. The method 400 may then continue at decision 406 with the processor 104 confirming whether the set of cycling times 127 has been normalized. In one embodiment, in response to confirming that the confirming that the set of cycling times 127 has not been normalized, the method 400 may return to block 404. On the other hand, in response to confirming that the set of cycling times 127 has been normalized, the method 400 may then continue at block 408 with the processor 104 training a predictive algorithm based on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and a plurality of other HVACs 103A, 103B.

[0068] The method 400 may then continue at block 410 with the processor 104 executing, based on the sensor data 126, the preprocessed set of cycling times 127, and the operating conditions data 129, and, in response to the training of the predictive algorithm, a predictive model 122 to generate an expected cycling time 124 for the one or more of the first HVAC system 103A or the second HVAC system 103B. The method 400 may then continue at block 412 with the processor 104 outputting, by the predictive model 122, the expected cycling time 124 for the one or more of the first HVAC system 103A or the second HVAC system 103B. The method 400 may then continue at block 414 with the processor 104 computing a difference between the expected cycling time 124 and one or more cycling times 127 of the set of cycling times 127.

[0069] The method 400 may then continue at decision 416 with the processor 104 confirming whether the difference between the expected cycling time 124 and the one or more cycling times 127 is greater than or equal to a predetermined cycling time threshold. In one embodiment, in response to confirming that difference between the expected cycling time 124 and the one or more cycling times 127 is greater than or equal to the predetermined cycling time threshold (e.g., at decision 416), the method 400 may then continue at block 418 with the processor 104 generating a first alert including an indication of a short cycling of the one or more of the first HVAC system 103A or the second HVAC system 103B and instructions for remediating the one or more of the first HVAC system 103A or the second HVAC system 103B.

[0070] On the other hand, in response to confirming that difference between the expected cycling time 124 and the one or more cycling times 127 is less than the predetermined cycling time threshold (e.g., at decision 416), the method 400 may then conclude at block 420 with the processor 104 generating a second alert including an indication of a normal cycling of the one or more of the first HVAC system 103A or the second HVAC system 103B.

[0071] FIG. 5 illustrates a flowchart of an example method 500 for generating clusters of short cycles of heating, ventilation, and air conditioning (HVAC) systems and correlating short cycles with indoor humidity, in accordance with one or more embodiments of the present disclosure. The method 500 may be performed utilizing the processor 104 of the multisystem controller 102 as described above with respect to FIG. 1. The method 500 may begin at block 502 with the processor 104 accessing a set of short cycling times associated with a one or more of the first HVAC system 103A or the second HVAC system 103B. The method 500 may then continue at block 504 with the processor 104 training a clustering algorithm based on a training data set of an aggregate of short cycling times127 associated with the one or more of the first HVAC system 103A or the second HVAC system 103B and a plurality of other HVAC systems 103A, 103B.

[0072] The method 500 may then continue at block 506 with the processor 104 executing, based on the set of short cycling times, and, in response to the training of the clustering algorithm, a clustering model 133 to generate a plurality of clusters 302, 304, and 306 of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B, in which each cluster of the plurality of clusters 302, 304, and 306 of short cycling times includes a frequency of short cycles. The method 500 may then continue at block 508 with the processor 104 outputting, by the clustering model 133, the plurality of clusters 302, 304, and 306 of short cycling times associated with the one or more of the first HVAC system 103A or the second HVAC system 103B.

[0073] The method 500 may then continue at block 510 with the processor 104 estimating, based on the plurality of clusters 302, 304, and 306 of short cycling times, a level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B), in which the level of indoor humidity is approximately proportional to the frequency of short cycles. The method 500 may then continue at decision 512 with the processor 104 confirming whether the level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B) has been estimated.

[0074] In one embodiment, in response to confirming that the level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B) has not been estimated (e.g., at decision 512), the method 500 may return to block 510. On the other hand, in response to confirming that the level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B) has been estimated (e.g., at decision 512), the method 500 may then conclude at block 514 with the processor 104 instructing the one or more of the first HVAC system 103A or the second HVAC system 103B to operate in accordance with the level of indoor humidity associated with one of a first space (e.g., downstairs space 100A) or a second space (e.g., upstairs space 100B).

[0075] While several embodiments have been provided in the present disclosure, it should be understood that the disclosed systems and methods might be embodied in many other specific forms without departing from the spirit or scope of the present disclosure. The present examples are to be considered as illustrative and not restrictive, and the intention is not to be limited to the details given herein. For example, the various elements or components may be combined or integrated with another system or certain features may be omitted, or not implemented.

[0076] In addition, techniques, systems, subsystems, and methods described and illustrated in the various embodiments as discrete or separate may be combined or integrated with other systems, modules, techniques, or methods without departing from the scope of the present disclosure. Other items shown or discussed as coupled or directly coupled or communicating with each other may be indirectly coupled or communicating through some interface, device, or intermediate component whether electrically, mechanically, or otherwise. Other examples of changes, substitutions, and alterations are ascertainable by one skilled in the art and could be made without departing from the spirit and scope disclosed herein.

[0077] To aid the Patent Office, and any readers of any patent issued on this application in interpreting the claims appended hereto, applicants note that they do not intend any of the appended claims to invoke 35 U.S.C. § 112(f) as it exists on the date of filing hereof unless the words “means for” or “step for” are explicitly used in the particular claim.

Claims

1. A system, comprising:a first heating, ventilation, and air conditioning (HVAC) system configured to regulate temperature of one of a first space or a second space of a building;a second HVAC system configured to regulate temperature of one of the first space or the second space of the building, wherein each of the first HVAC system and the second HVAC is configured to regulate the temperature of a different one of the first space or the second space of the building; anda multisystem controller communicatively coupled to the first HVAC system and the second HVAC system, the multisystem controller comprising:a memory configured to store sensor data, a set of cycling times, and operating conditions data associated with one or more of the first HVAC system or the second HVAC system; anda processor communicatively coupled to the memory and configured to:access the sensor data, the set of cycling times, and the operating conditions data associated with the one or more of the first HVAC system or the second HVAC system;preprocess the set of cycling times, wherein preprocessing the set of cycling times comprises normalizing the set of cycling times to scale for temperature variations and physical dimensions of the one or more of the first HVAC system or the second HVAC system;train a predictive algorithm based at least in part on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system or the second HVAC system and a plurality of other HVAC systems;execute, based at least in part on the sensor data, the preprocessed set of cycling times, and the operating conditions data, and, in response to the training of the predictive algorithm, a predictive model to generate an expected cycling time for the one or more of the first HVAC system or the second HVAC system;output, by the predictive model, the expected cycling time for the one or more of the first HVAC system or the second HVAC system;compute a difference between the expected cycling time and one or more cycling times of the set of cycling times; andin response to determining that the difference between the expected cycling time and the one or more cycling times is greater than or equal to a predetermined cycling time threshold, generate an alert comprising an indication of a short cycling of the one or more of the first HVAC system or the second HVAC system and instructions for remediating the one or more of the first HVAC system or the second HVAC system.

2. The system of claim 1, wherein the alert comprises a first alert, and wherein the processor is further configured to:in response to determining that the difference between the expected cycling time and the one or more cycling times is less than the predetermined cycling time threshold, generate a second alert comprising an indication of a normal cycling of the one or more of the first HVAC system or the second HVAC system.

3. The system of claim 1, wherein the processor is further configured to:generate, based at least in part on one or more of the historical sensor data, the historical cycling time data, or the historical operating conditions data, normal cycling times for the one or more of the first HVAC system or the second HVAC system; andcompute, based at least in part on the normal cycling times, a mean cycling time of the normal cycling times and a standard deviation of the normal cycling times.

4. The system of claim 1, wherein the processor is further configured to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold by comparing the expected cycling time to a standard deviation of the normal cycling times.

5. The system of claim 1, wherein the processor is further configured to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold when the expected cycling time is an N number of standard deviations less than a mean cycling time.

6. The system of claim 1, wherein the processor is further configured to:store a log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system; andupdate the predictive model based at least in part on the log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system.

7. The system of claim 6, wherein the processor is further configured to execute the updated the predictive model to identify whether the short cycling of the one or more of the first HVAC system or the second HVAC system is one of a random short cycling occurrence or one of a systemic short cycling occurrence.

8. A method, comprising:accessing sensor data, a set of cycling times, and operating conditions data associated with one or more of a first heating, ventilation, and air conditioning (HVAC) system or a second HVAC system, wherein the first HVAC system is configured to regulate temperature of one of a first space or a second space of a building and the second HVAC system is configured to regulate temperature of a different one of the first space or the second space of the building;preprocessing the set of cycling times, wherein preprocessing the set of cycling times comprises normalizing the set of cycling times to scale for temperature variations and physical dimensions of the one or more of the first HVAC system or the second HVAC system;training a predictive algorithm based at least in part on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system or the second HVAC system and a plurality of other HVAC systems;executing, based at least in part on the sensor data, the preprocessed set of cycling times, and the operating conditions data, and, in response to the training of the predictive algorithm, a predictive model to generate an expected cycling time for the one or more of the first HVAC system or the second HVAC system;outputting, by the predictive model, the expected cycling time for the one or more of the first HVAC system or the second HVAC system;computing a difference between the expected cycling time and one or more cycling times of the set of cycling times; andin response to determining that the difference between the expected cycling time and the one or more cycling times is greater than or equal to a predetermined cycling time threshold, generating an alert comprising an indication of a short cycling of the one or more of the first HVAC system or the second HVAC system and instructions for remediating the one or more of the first HVAC system or the second HVAC system.

9. The method of claim 8, wherein the alert comprises a first alert, the method further comprising:in response to determining that the difference between the expected cycling time and the one or more cycling times is greater is less than the predetermined cycling time threshold, generating a second alert comprising an indication of a normal cycling of the one or more of the first HVAC system or the second HVAC system.

10. The method of claim 8, further comprising:generating, based at least in part on one or more of the historical sensor data, the historical cycling time data, or the historical operating conditions data, normal cycling times for the one or more of the first HVAC system or the second HVAC system; andcomputing, based at least in part on the normal cycling time, a mean cycling time of the normal cycling times and a standard deviation of the normal cycling times.

11. The method of claim 8, further comprising determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold by comparing the expected cycling time to a standard deviation of the normal cycling times.

12. The method of claim 8, further comprising determining that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold when the expected cycling time is an N number of standard deviations less than a mean cycling time.

13. The method of claim 8, further comprising:storing a log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system; andupdating the predictive model based at least in part on the log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system.

14. The method of claim 13, further comprising executing the updated the predictive model to identify whether the short cycling of the one or more of the first HVAC system or the second HVAC system is one of a random short cycling occurrence or one of a systemic short cycling occurrence.

15. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:access sensor data, a set of cycling times, and operating conditions data associated with one or more of a first heating, ventilation, and air conditioning (HVAC) system or a second HVAC system, wherein the first HVAC system is configured to regulate temperature of one of a first space or a second space of a building and the second HVAC system is configured to regulate temperature of a different one of the first space or the second space of the building;preprocess the set of cycling times, wherein preprocessing the set of cycling times comprises normalizing the set of cycling times to scale for temperature variations and physical dimensions of the one or more of the first HVAC system or the second HVAC system;train a predictive algorithm based at least in part on a training data set of historical sensor data, historical cycling time data, and historical operating conditions data associated with the one or more of the first HVAC system or the second HVAC system and a plurality of other HVAC systems;execute, based at least in part on the sensor data, the preprocessed set of cycling times, and the operating conditions data, and, in response to the training of the predictive algorithm, a predictive model to generate an expected cycling time for the one or more of the first HVAC system or the second HVAC system;output, by the predictive model, the expected cycling time for the one or more of the first HVAC system or the second HVAC system;compute a difference between the expected cycling time and one or more cycling times of the set of cycling times; andin response to determining that the difference between the expected cycling time and the one or more cycling times is greater than or equal to a predetermined cycling time threshold, generate an alert comprising an indication of a short cycling of the one or more of the first HVAC system or the second HVAC system and instructions for remediating the one or more of the first HVAC system or the second HVAC system.

16. The non-transitory computer-readable medium of claim 15, wherein the alert comprises a first alert, and wherein the instructions further cause the one or more processors to:in response to determining that the difference between the expected cycling time and the one or more cycling times is less than the predetermined cycling time threshold, generate a second alert comprising an indication of a normal cycling of the one or more of the first HVAC system or the second HVAC system.

17. The non-transitory computer-readable medium of claim 15, wherein the instructions further cause the one or more processors to:generate, based at least in part on one or more of the historical sensor data, the historical cycling time data, or the historical operating conditions data, normal cycling times for the one or more of the first HVAC system or the second HVAC system; andcompute, based at least in part on the normal cycling times, a mean cycling time of the normal cycling times and a standard deviation of the normal cycling times.

18. The non-transitory computer-readable medium of claim 15, wherein the instructions further cause the one or more processors to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold by comparing the expected cycling time to a standard deviation of the normal cycling times.

19. The non-transitory computer-readable medium of claim 15, wherein the instructions further cause the one or more processors to determine that the difference between the expected cycling time and the one or more cycling times is greater than or equal to the predetermined cycling time threshold when the expected cycling time is an N number of standard deviations less than a mean cycling time.

20. The non-transitory computer-readable medium of claim 15, wherein the instructions further cause the one or more processors to:store a log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system; andupdate the predictive model based at least in part on the log of the indication of the short cycling of the one or more of the first HVAC system or the second HVAC system.

Images