METHOD AND SYSTEM FOR HVAC COMPONENT FAILURE PREDICTION

[0001] This application claims benefit of U.S. provisional patent application No. 63/736,572 filed Dec. 19, 2024 titled Method and System for HVAC Component Failure Prediction.

[0002] This disclosure relates generally to electrical equipment and in particular to a method and system for predicting the future or probability of failure of individual electrical components for HVAC systems.

[0003] HVAC, systems commonly use a plurality of electrical and mechanical components to move air and other fluids into and out of occupied spaces for the comfort of occupants or other considerations. In particular such systems commonly utilized motors for driving fans to move air and pumps to move liquids as well as valves to control flow of fluids along with many other electrical components.

[0004] One difficulty in current HVAC systems is that the components that make up such a system may be prone to failure. However, disadvantageously, when one component fails, the entire HVAC system frequently fails to operate at all with little indication to a user as to the source of the failure. In particular, a fan or a valve may fail resulting in the stopped operation of the HVAC system without any indication as to why. The source of such failure must then be identified by a service technician who may or may not be able to attend within a reasonable period of time.

[0005] A common method of avoiding problematic component failure is the use of regular, such as annual, maintenance on the HVAC system. Such annual maintenance, however only provide a one-time inspection for possible signs of future failure. Many of the operating characteristics which may be indicative of a potential part failure are however, outside the scope possible during such maintenance.

[0006] An additional difficulty for home and building owners is that the failure of specific components is often only apparent after such failure has occurred. While part life and other predictive methods have been used to determine a reasonable lifespan of particular systems and components, such methods only rely on statistical analysis of similar components. Such statistical analysis does not take into account the actual operation of that component and therefore fails to predict any pre-mature failure of the components.

[0007] According to a first embodiment of the present disclosure is a system for predicting the potential failure of components in HVAC systems, the system comprising a sensor operable to measure at least one operating condition of an electrical component in an HVAC system and a prediction module including a processor operable. The prediction module is configured to to receive data from the sensor, store historical data from the sensor and analyse the data and historical data to predict a potential loss of effective operation of the electrical component

[0008] The prediction module may be located across a distributed network and/or server. The prediction module may utilize a machine learning process to analyse the historical data and predict a potential loss of effective operation. The network may include a database containing historical data on a plurality of equivalent electrical components corresponding to the electrical component.

[0009] The prediction module may include a trained model representing failure potentials, timing and conditions for the electrical components. The prediction module may be configured to update the trained model in response to an input of an actual failure of the electrical component utilizing machine learning.

[0010] The component may comprise an electrical motor. The electrical motor may comprise a fan motor. The sensor may be operable to measure at least one of current and amperage consumed by the electrical component. The system may further include a performance sensor adapted to measure a performance of the electrical component. The performance sensor may be configured to measure an airflow caused by the electrical component.

[0011] The system may further comprise a data collection module operable to transmit the received data to a remote prediction module. The prediction module may be located in a cloud server. The system may further comprise an alert module operable to transmit an alert signal to a user in response to the processor predicting a potential failure of the component.

[0012] According to a further embodiment of the present disclosure is a method for predicting potential failure of electrical components in HVAC systems comprising receiving, from a sensor at least one operating condition of an electrical component in an HVAC system at a controller, analysing data from the sensor at the controller, utilizing machine learning with a trained model representing performance and failure of comparable electrical components and operating conditions and predicting a potential failure as determined by the controller based on the trained model.

[0013] Other aspects and features of the present disclosure will become apparent to those ordinarily skilled in the art upon review of the following description of specific embodiments in conjunction with the accompanying figures.

[0019] Referring to FIG. 1, an exemplary system for monitoring and predicting failure of an electrical component is illustrated generally at 10. As illustrated in Figure, 1, the electrical component may be an electrical fan 12 driven by an electrical motor 14 and having a downstream duct 16 being supplied with air from the fan. Although not illustrated in FIG. 1, it will be appreciated that such fan 12 may be a part of a furnace air handling unit or air conditioner, by way of non-limiting example. The motor 14 includes a power supply, 13 to it as is commonly known which may comprise an electrical supply at any known voltage and frequency.

[0020] The system includes a sensor 30 around or in line with the power supply 13 adapted to measure an instantaneous current or voltage supply to or consumed by the motor 14. In particular, the sensor may comprise a current sensor or a voltage detector or a combination thereof, selected from any known sensor types as are known. A controller 20 is in communication with the sensor to receive a measurement of the current and/or voltage supplied to or consumed by the motor 14. The controller includes a processor 22 and a database 24 containing information on the historical measurements of the motor 14 as measured by the sensor 30 as well as a taught model of the data measurements of similar or identical motors 14 and the relationship to such measurements as to the predicted failure of those motors.

[0021] The controller 30 includes instructions configured to identify the current operating state of the motor 14 and determine when such conditions indicate that a failure of the motor may be possible or imminent. The conditions in which the controller may determine when a component may be at risk of failure may be developed and refined in the taught model by commonly known techniques. In particular, the taught model may be configured to recognized operational patterns in the component that lead to previous failures or within other equipment. Subsequent failures and maintenance performed using the outputs from the taught model will refine such recognition. Optionally, a technician may input a threshold for an out of specification operation such as too high or too low of a voltage or current to the component. Such thresholds may be inputted or adjusted from time to time by a technician or operator. The taught model contains a model operable to permit the processor to identify when the operating state of the motor 14 or other electrical component has changed to one in which it is recognized has in previous teaching examples indicated the imminent failure of that electrical component. In particular the processor 22 utilizes the measured data to recognize, utilizing machine learning methods when a given state is possible to or likely to experience a failure of that electrical component. Optionally, the operation of the processor 22 and/or controller 20 may be split across a local or specified processor operably coupled to the remaining components in the system and a remote processor, such as a cloud based server and database. In such embodiments, the local processor may be configured to push the collected data on any desired schedule, including without limitation, periodically, on demand or continuously to the could based server which is configured to perform the analysis and prediction functions described in further detail below. Furthermore in some embodiments, the local controller based functions may be collection of data which is then transmitted to a remote processor for processing as further set out below. The remote processor may be located at one or more discrete locations or may be distributed across cloud or networked computing systems as are known.

[0022] Optionally, the system may include a performance sensor 32 operable to measure the output or performance of the electrical component, such as the fan 14. In particular and in the case of a fan, the performance sensor 32 may comprise an air flow meter operable to measure the air flow output of the fan 14. The processor 22 may then include the performance measurement within the taught model to assess if the measured

[0023] Turning now to FIG. 2, a method for monitoring the operation of the electrical component, such as a fan as illustrated in FIG. 1 is illustrated generally. The system comprises, utilizing the processor, identifying the electrical component. The step of identifying the electrical component may comprise identifying which sensor leads from the sensor 30 are inputted into the controller 20. Such inputting may be accomplished during a commissioning process for the HVAC equipment although it will be appreciated that other identification methods, including through a bus or the like may also be utilized. In some embodiments, the identification of the electrical component may be performed by a user inputting the identification of the component into the controller. The method also includes identifying the associated sensor to be monitored. It will be appreciated that the identification of the electrical component and the associated sensor may be performed contemporaneously or in any order. The processor then monitors the sensor for the measured criteria. When a given criteria is within a or above a threshold set in the taught model that a failure is imminent or probably, the processor is then configured to send an alert to the monitoring station or user device as set above.

[0024] More generally, in this specification, the term “processor” is intended to broadly encompass any type of device or combination of devices capable of performing the functions described herein, including (without limitation) other types of microprocessors, microcontrollers, other integrated circuits, other types of circuits or combinations of circuits, logic gates or gate arrays, or programmable devices of any sort, for example, either alone or in combination with other such devices located at the same location or remotely from each other. Additional types of processor(s) will be apparent to those ordinarily skilled in the art upon review of this specification, and substitution of any such other types of processor(s) is considered not to depart from the scope of the present invention as defined herein. In various embodiments, the processor can be implemented as a single-chip, multiple chips and/or other electrical components including one or more integrated circuits and printed circuit boards.

[0025] Computer code comprising instructions for the processor(s) to carry out the various embodiments, aspects, features, etc. of the present disclosure may reside in memory within the processor or be obtained from outside the system. The code may be broken into separate routines, products, etc. to carry forth specific steps disclosed herein. In various embodiments, the processor can be implemented as a single-chip, multiple chips and/or other electrical components including one or more integrated circuits and printed circuit boards. The processor together with a suitable operating system or control logic may operate to execute instructions in the form of computer code and produce and use data. By way of example and not by way of limitation, the operating system may be Windows-based, Mac-based, or Unix or Linux-based, among other suitable operating systems. Coded instructions, operating systems and other control software and instruction sets are generally well known and will not be described in further detail here. The processor 22 may include adaptive algorithms including by way of non-limiting example, machine learning algorithms programmed to communicate with an ability to interpret answers from the users and provide further analysis and interpretation thereof. Such algorithms may include regression analysis, natural language processing, or any other machine learning methods as are available.

[0026] Memory may include various tangible, non-transitory computer-readable media including Read-Only Memory (ROM) and/or Random-Access Memory (RAM). As is well known in the art, ROM acts to transfer data and instructions uni-directionally to the processor, and RAM is used typically to transfer data and instructions in a bi-directional manner. In the various embodiments disclosed herein, RAM includes computer program instructions that when executed by the processor cause the processor to execute the program instructions described in greater detail below. More generally, the term “memory” as used herein encompasses one or more storage mediums and generally provides a place to store computer code (e.g., software and/or firmware) and data. It may comprise, for example, electronic, optical, magnetic, or any other storage or transmission device capable of providing the processor with program instructions. Memory may further include a floppy disk, CD-ROM, DVD, magnetic disk, memory chip, ASIC, FPGA, EEPROM, EPROM, flash memory, optical media, or any other suitable memory from which processor can read instructions in computer programming languages.

[0027] The processor may also interface with input and output or a remote monitoring system 28 and/or user device 29 such as a smart phone, tablet, computer or the like and/or building management systems including HVAC management. The processor 22 may be configured to transmit to the remote monitoring system 28 or user device that a component failure is possible or imminent.

[0028] Turning now to FIG. 3, according to some embodiments, the system may comprise a monitoring module 40 operably connected to a plurality of electrical components 14 either within a common piece of HVAC equipment or a common building or site. In particular the monitoring module 40 may be configured to receive the data from the sensors 30 and performance sensors 32 for each electrical component 14, compile the data and transmit the data through a network 50 to a server or cloud based server which may include one or more processors and databases operable to function as the controller 20 as set out above. As illustrated in FIG. 4, the cloud 50 may be connected to a plurality of monitoring modules 40 at a plurality of locations configured to monitor each electrical component at each location. The cloud or distributed network 50 may include a prediction module 52 and an alert module 54.

[0029] The database contains a plurality of entries for each piece of electrical component and may be stored in any known manner. In particular it has been found useful to store the information, details and operating conditions of each electrical component as individual data points within a relational dataset within the database. As part of each data set for that particular electrical component a technician or other personnel may input or download as an initial set up for that electrical component, details of the component itself, including by way of non-limiting example, model, make, year, serial number as well as surrounding operating environments and conditions. The technician may also input a baseline of the operating characteristics of that component such as the voltage and current and performance outputs as the component is known to operate under a new, or proper working order. Such baseline data may be utilized by the predictive algorithm and machine learning within the prediction module to determine when the operation or performance of the component has degraded. Additional data for each component may then be stored within the database 24 as measured by the sensor 30 and/or performance sensor 32. In particular the sensor and/or performance sensor may provide such instantaneous operation measure at sampling rate selected to provide indication of the operation of the component over time. The sampling rate should be selected to provide a detailed enough illustration of the operation while not missing short term fluctuations or spikes in performance while utilizing manageable bandwidth across the network By way of non-limiting examples it has been found that a sampling frequency of approximately every 2 seconds has been useful for residential HVAC equipment although it will be appreciated that shorter periods may be desired for more critical equipment or as greater bandwidth becomes possible or reduced sampling rates for equipment that is less prone to short term fluctuations.

[0030] The prediction module 52 may contain the machine learning and/or taught model. As set out above the machine learning algorithm within the prediction module is configured to analyse the data contained within the database 24 and identify conditions, operation and/or patterns which may indicate that a failure or degradation of a particular component may be possible. It will be appreciated that as the prediction module analyses the entire dataset within the database, such analysis would not be capable by a human operator. The prediction module 52 may also be configured to prioritize or isolate the data considered to components identical or similar to the one in question as may be determined to be optimal by the learning model over time.

[0031] The alert module 54 is configured to output to a user, either through a remote monitoring station 28 or directly to a user device 29 as set out above one of a plurality of notices. Such notice may take any format as are commonly known, including, without limitation, pop up notice, text message, email, telephone call or any other auditory or visual indication. In particular the alert module may be configured to output a fault notice to a user when the sensor and/or performance sensor indicates that the electrical component has stopped operating or has ceased to turn on when required. Examples of such failure may include, without limitation, no current to a fan, or zero air flow from the fan indicating it is not turning. The alert module may also be configured to output an alert notice to an operator that either the operating data of the component as reported by the sensor and/or the performance of the component as reported by the performance sensor is within a pattern recognized by the prediction model 52 to predict an imminent or potential future failure of the component. The system may furthermore be configured to receive a lower priority signal, including, such as by way of non-limiting example, warnings and the like of conditions outside of normal operating parameters but not reaching the level of a failure or alert. The prediction module 52 may be configured to receive and analyse such warnings or notices and recognize a trend of such warnings towards an eventual failure or alert and provide a notice to the user of such prediction similar to as set out above.

[0032] According to some embodiments, the baseline data for the HVAC system and in particular each component therein is stored. These baselines can be used to detect changes indicating impending or existing failures. By way of non-limiting example, of particular measured current signatures of the electrical components for failures of those components may be pre-programmed, and may be updated based on observed evidence from contractors within the taught model. For example, once a malfunction in an HVAC system is recognized, the system may note the frequency of particular data types leading up to the malfunction and correlate that data type with frequency signatures associated with potential causes of the malfunction. The data signal types and indications utilized by the system may be unique to each model and type of electrical component but may share common characteristics. These common characteristics may be adapted based on the specific type of HVAC system being monitored.

[0033] It will be appreciated that, the taught model and processor may comprise a computer learning system, such as a neural network or a genetic algorithm such as a deep learning algorithm an anomaly detection algorithm, a linear regression algorithm, or a logistical regression algorithm, or a combination of various machine learning algorithms, to name a few examples which may be used to refine frequency signatures. The system may further be configured to provide a dataset containing information on the predictive failure points for individual components and systems including advanced analytics of such failure, including, without limitation mean time to failure and the like. In particular the system may compare operational data of the HVAC system illustrated an in particular the electrical component to other HVAC systems having predictive HVAC equipment monitoring systems installed on the other HVAC systems to generate probabilities of failure across multiple systems and operating conditions.

[0034] Although he present system and method are described above with respect to HVAC systems and the electrical components forming such system, it will be appreciated that such system and methods may also be utilized to monitor and predict the potential failure of other electrical components in other systems, including by way of non-limiting example, drive motors, fluid valves, pumps and motors. It will also be appreciated that the system may be designed into and incorporated into a new system or applied to existing HVAC equipment including, but not limited to furnace HVAC systems.

[0035] While specific embodiments of the disclosure have been described and illustrated, such embodiments should be considered illustrative only and not as limiting the invention as construed in accordance with the accompanying claims.