WO2025108693A1

WO2025108693A1 - Systems and methods to determine a multi-level health index for detection of equipment anomalies

Info

Publication number: WO2025108693A1
Application number: PCT/EP2024/081161
Authority: WO
Inventors: Fahad Abdullah ALDAKHEEL; Omar Abdulaziz MOAWAD; Abdulaziz Ibrahim BIN SALAMAH; Hassan Ali ALBAQSHI; Adel Ali AL-ALI; Abdulrahman Mohammed ALBESHR; Rayan Khalid KHALIL; Fahad Khalef ALZEYADI; Ahmed Ali ALREBH; Fahad Yousef ALSHOWAIRED; Jalal Majed ALRUKHAIMI
Original assignee: SABIC Global Technologies BV
Current assignee: SABIC Global Technologies BV
Priority date: 2023-11-20
Filing date: 2024-11-05
Publication date: 2025-05-30
Anticipated expiration: 2026-05-20

Abstract

Disclosed here are methods and systems to detect equipment anomalies or equipment failure based on a multi-level health index for equipment positioned at industrial plants. In an embodiment, the method may include applying data points from the industrial plants to first principle models to produce a plurality of outputs. The method may include determining the multi-level health index based on application of one or more of (a) the data points or (b) the plurality of outputs to one or more of the plurality of trained machine learning models. The method may include comparing each level of the multi-level health index to historical conditions to detect one or more of equipment anomalies or failure. the method may include, in response to one level of the multi-level health index being outside of the historical conditions, generating an alert.

Description

SYSTEMS AND METHODS TO DETERMINE A MULTI-LEVEL HEALTH INDEX FOR DETECTION OF EQUIPMENT ANOMALIES

FIELD OF DISCLOSURE

[0001] The present disclosure generally relates to systems and methods to determine a multi-level health index for each of one or more industrial plants and to, based on each level of the multi-level health index, detect and/or predict component and/or equipment level anomalies and/or failures.

BACKGROUND

[0002] An industrial plant may include a variety of equipment, such as compressors, pumps, valves, and/or other equipment utilized for various plant operations. The equipment may run or be operated every day and for all or a substantial portion of each day. While some equipment can be serviced without impacting plant production, other equipment cannot be brought down, taken offline, or shut down without causing a significant impact to plant production and/or operations. If such components are, for example, shut down, such an action may cause a reduction of product output and an increase in cost associated with the loss of such product output. Further, each industrial plant may include different and/or in some examples similar equipment and, further, each industrial plant may produce a variety of different products.

BRIEF SUMMARY

[0003] In view of the foregoing, Applicant has recognized these problems and others in the art, and has recognized a need for enhanced systems and methods to determine a multi-level health index for each of one or more industrial plants and, based on each level of the multi-level health index, detect and/or predict component and/or equipment level anomalies and/or failures. Further, the multi-level health index may enable assessment of a plurality of components at a plurality of plants.

[0004] The present disclosure generally relates to systems and methods that address the relevant issues as described above, among other issues. In particular, such systems and methods may enable a user a to monitor overall health, via the multi-level health index, of a plurality of industrial plants and, by selecting a lower level of the health index, monitor health at the component level for each industrial plant. Such a system or method may reduce time to determine whether equipment is experiencing or may experience an issue by centralizing the health index for each of the plurality of industrial plants on a single platform, which is not typical due to each industrial plant (a) generating a large data set daily, as most components run a substantial part of or even all of the day, and (b) including a plurality of different components and equipment. Further, such systems and methods may automatically monitor the multi-level health index against historical conditions or historical optimal conditions and automatically generate alerts for anomalies at the component and/or equipment level. [0005] Such systems and methods may capture data, in real-time and/or, in another embodiment, may acquire or capture data from previous or historical industrial plant operations (for example, from a database or other type of storage), and preprocess the captured data to form a preprocessed data set. Such preprocessing may include applying the data to a first principle model (also referred to as a physics based model), removing outlying data points, and/or selecting a training set and testing set of data. Such systems and methods may include, if a particular trained machine learning model is not available and/or to re-train a trained machine learning model, applying the preprocessed data to one or more machine learning models to generate one or more classifiers or trained machine learning model.

[0006] After generating one or more trained machine learning models (and/or classifiers), the systems and methods may capture real-time data from each of a plurality of industrial plants. The data may include raw sensor data corresponding to various components and/or equipment. The systems and methods may apply the raw sensor data to a corresponding first principle model from a plurality of first principle models, to generate a first principle output. The systems and methods may then apply the raw sensor data and the first principle outputs to a corresponding trained machine learning model. After such application for each of a plurality of components corresponding to a plurality of equipment, the systems and methods may include generating a multi-level health index.

[0007] The multi-level health index may include component level health index values based on the output of the trained machine learning models, as well as, in some embodiments, the first principle model outputs. Further, each component level health index may be combined to generate an over-all health level index and/or an equipment level health index. In such embodiments, each component level health index may be weighted the same. In other embodiments, each component level health index may be weighted differently based on the type of component and/or the type of equipment corresponding to the component. For example, some equipment may operate continuously or substantially continuously and failure to operate the equipment continuously may result in loss or decrease in production of a product and/or component and, in another example, some components may be used more than others in relation to operation of the equipment, while other components are ancillary and can be serviced and/or replaced without impacting operation or production. As such, the overall health index may be impacted more or less based on the type of equipment and/or the type of component. In an embodiment, a top level of the multi-level health index may correspond to all monitored industrial plants, a next level of the multi-level health index may correspond to a selected industrial plant, a second next level of the multi-level health index may correspond to equipment at the selected plant, and/or a third next level of the multi-level health index may correspond to a component level corresponding to equipment at the selected plant. Thus, a user may “drill down” or traverse through the multi-level health index to root cause and/or predict equipment and/or component anomalies or failures.

[0008] As noted, the multi-level health index may predict future maintenance, service cycles, and/or failures for selected equipment and/or components. For example, after application of the data to a trained machine learning model, the systems and methods may utilize the output to generate a prediction regarding when a particular component and/or corresponding equipment may be maintained, serviced, and/or replaced. In other word, the multi-level health index may include details and/or information regarding potential or predicted component, equipment, and/or plant issues.

[0009] Further still, the systems and methods may automatically alert a user when any health index from any level of the multi-level health index is outside of a selected range or falls below a selected threshold and indicates different levels and/or a timeframe of maintenance to be performed (for example, immediate, within a week, within a month, within a year, and/or within other time frames). [0010] In other embodiments, the systems and methods may include a platform. The platform may monitor for and/or be configured to receive new industrial plants to be monitored via the multi-level health index. The platform may be configured to enable a user to enter credentials and/or other access data to cause the platform to connect to a controller and/or computing device at an industrial plant, thus allowing the platform to monitor the industrial plant and/or generate a health index for the industrial plant. The platform, in an embodiment, may include an interactive display and/or user interface that enables user to select various levels of the multi-level health index.

[0011] Thus, by utilizing the systems and methods described, a user may quickly diagnose and/or be alerted, of a current and/or impending maintenance cycle or equipment failure.

[0012] Accordingly, an embodiment of the disclosure is directed to a method to detect equipment anomalies or equipment failure based on a multi-level health index for equipment positioned at one of a plurality of industrial plants. The method may include training a plurality of machine learning models with historical data associated with a plurality of industrial plants to form a plurality of trained machine learning models. The historical data may include one or more of (a) historical raw sensor data from equipment at the plurality of industrial plants, (b) historical first principle model outputs associated with the plurality of industrial plants, (c) maintenance records of the equipment at the plurality of industrial plants, (d) failure records of the equipment in the plurality of industrial plants, or (e) one or more identifiers to indicate one or more selected equipment, selected chemicals, an equipment type, or a chemical type. The method may include receiving, in real time, a plurality of raw sensor data points from the equipment at the plurality of industrial plants. The method may include applying the data points to one or more first principle models to produce a plurality of outputs from one or more first principle models associated with the plurality of industrial plants. The method may include determining the multi-level health index based on application of one or more of (a) one or more of the plurality of raw sensor data points or (b) one or more of the plurality of outputs from the one or more first principal models to a selected one or more of the plurality of trained machine learning models. The method may include comparing, in real-time, each level of the multi-level health index to historical conditions to detect one or more of equipment anomalies or equipment failure. The method may also include, in response to one level of the multi-level health index being outside of the historical conditions, generating an alert.

[0013] In an embodiment, the method may further include monitoring for detection of a new industrial plant and, in response to detection of the new industrial plant, determining, based on a type of equipment positioned at the new industrial plant, if the plurality of trained machine learning models correspond to equipment positioned at the new industrial plant. The method may include if equipment at the new industrial plant does not correspond to the plurality of trained machine learning models. The method may include obtaining, via the new industrial plant, new plant historical data, the new plant historical data to include one or more of (a) historical raw sensor data from the equipment at the new industrial plant with no corresponding trained machine learning model, (b) historical first principle model outputs associated with the equipment at the new industrial plant with no corresponding trained machine learning model, (c) maintenance records of equipment at the new industrial plant with no corresponding trained machine learning model, (d) failure records of the equipment in the new industrial plant with no corresponding trained machine learning model, or (e) the one or more identifiers to indicate one or more selected equipment, selected chemicals, the equipment type, or the chemical type at the new industrial plant. The method may include training one or more new machine learning models with the new plant historical data.

[0014] In an embodiment, the multi-level health index may comprise a plurality of health indices that indicate a status of each of the plurality of the industrial plants, each equipment at each of the plurality of the one or more of the industrial plants, and each component of each equipment at each of the plurality of the one or more of the industrial plants.

[0015] In another embodiment, the method may include generating a user interface, the user interface to include an interactive display and to initially display a top level of the multi-level health index for each of the plurality of the industrial plants. The method may include in response to selection of one of the plurality of the industrial plants on the interactive display, displaying a next level of the multilevel health index for each equipment at the one of the plurality of the industrial plants. The method may include, in response to selection of one of each equipment at the one of the plurality of the industrial plants, generating a schematic view of the one of each equipment, wherein the schematic view includes each component of the one of each equipment and a corresponding level of the multilevel health index.

[0016] In some embodiments, the alert may include a corrective action selected from a predefined set of corrective actions. For example, the corrective action may include servicing a selected component of selected equipment.

[0017] Another embodiment, of the disclosure is directed to a method to detect equipment anomalies or equipment failure based on a multi-level health index for a plurality of plants. The method may include obtaining real-time component level data from a plurality of equipment positioned at one of the plurality of plants. The method may include determining a plurality of subsets of the real-time component level data based on a type of component and a type of equipment. The method may include determining, based on application of each of the plurality of subsets to one or more machine learning models or first principle models that correspond to each of the plurality of subsets, each level of the multi-level health index for each of the plurality of plants. The method may include generating a top level of the multi-level health index for each of the plurality of plants based on each determined level of the multi-level health index for each of the plurality of plants. The method may include comparing, in real-time, each level of the multi-level health index to corresponding historical conditions to detect one or more of equipment anomalies or equipment failure. The method may include in response to a comparison that is not within the corresponding selected threshold range of each level of the multilevel health index for each of the plurality of plants with a historical optimal condition, generating an alert.

[0018] In another embodiment, the each level of the multi-level health index for each of the plurality of plants may be based on one or more of a corresponding type of equipment, a weight for the corresponding type of equipment, a corresponding type of component, or a weight for the corresponding type of component.

[0019] In another embodiment, the historical condition comprises one or more of previous maintenance, previous pauses in production, previous equipment failures, previous chemical output, or use of redundant equipment. In yet another embodiment, a top level of the multi-level health index of a selected plant may indicate a current operational status of the selected plants. In another embodiment, the multi-level health index may indicate a failure prediction time frame of selected equipment, and wherein the notification includes a next action to prevent failure of the selected equipment.

[0020] Another embodiment of the disclosure is directed to a system to detect equipment anomalies or equipment failure based on a multi-level health index for a plurality of plants. The system may include a communications circuitry configured to obtain, in real-time, component level data corresponding to a plurality of equipment positioned at one of a plurality of plants. The system may include a preprocessing circuitry configured to determine a plurality of subsets of the component level data based on a type of component, a type of equipment, or a type of plant. The system may include a modeling circuitry configured to generate a plurality of probabilities based on application of each of the plurality of subsets to one or more corresponding machine learning models. The system may include a health index circuitry configured to determine a health index for each component based on the plurality of probabilities, generate a multi-level health index for each of the plurality of plants based on the health index for each component, and monitor for one or more of equipment anomalies or equipment failure based on a comparison of each level of the multi-level health index to a corresponding selected range.

[0021] In an embodiment, the multi-level health index may comprise a component level health index, an equipment level health index, and a plant level health index. In another embodiment, the component level health index may comprise each health index for each component, the equipment level health index may comprise a second health index for each equipment based on the health index of each component associated with corresponding equipment, and the plant level health index may comprise a third health index for each plant based on the second health index of each equipment associated with a corresponding plant. The second health index for each equipment may be further based on a weight associated with each component of each equipment and the third health index for each plant may be further based on a weight associated with each equipment positioned at the corresponding plant. [0022] Another embodiment of the disclosure is directed to a computing device to detect equipment anomalies or equipment failure based on a multi-level health index for a plurality of plants, the controller. The computing device may include a plurality of inputs/outputs each in signal communication with a plant. The computing device may also include a processor and a non-transitory machine readable storage medium storing instructions configured to, when executed by the processor, obtain, in real-time, component level data corresponding to a plurality of equipment positioned at one of a plurality of plants. The instructions, when executed, may determine a plurality of subsets of the component level data based on a type of component, a type of equipment, or a type of plant. The instructions, when executed, may generate a plurality of probabilities based on application of each of the plurality of subsets to one or more corresponding machine learning models. The instructions, when executed, may determine a health index for each component based on the plurality of probabilities. The instructions, when executed, may generate a multi-level health index for each of the plurality of plants based on the health index for each component. The instructions, when executed, may monitor for one or more of equipment anomalies or equipment failure based on a comparison of each level of the multi-level health index to a corresponding selected range.

[0023] In an embodiment, the corresponding selected range comprises a value associated with a historical optimal condition, and wherein the controller is configured to generate one or more of a notification, a corrective action, equipment status, or an alert level based on the comparison. In another embodiment, the corrective action may include performing maintenance on selected equipment.

[0024] Additional and/or alternative objects, features and advantages of the present disclosure will become apparent to the skilled artisan from the figures, detailed description, and examples herein. Applicant notes, however, that the figures, detailed description, and examples, while indicating certain embodiments of the instant disclosure, are provided for illustrative purposes only and are not intended to be limiting or to imply a particular limitation. Moreover, certain changes and modifications within the spirit and scope of the disclosed technology will become apparent to those of ordinary in the relevant art from this detailed description.

BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The disclosed aspects, features and advantages of the disclosure will become better understood with regard to the following descriptions, examples, claims, and accompanying drawings. Applicant notes, however, that the drawings illustrate certain embodiments of the disclosure and should not be considered limiting with regards to the breadth and scope of the disclosure:

[0026] FIG. 1A and FIG. IB are schematic diagrams of a system to determine a multi-level health index, in accordance with certain embodiments of the present disclosure;

[0027] FIG. 2 is another schematic diagram of an apparatus to determine a multi-level health index, in accordance with certain embodiments of the present disclosure;

[0028] FIG. 3 is a schematic diagram of a computing device to determine a multi-level health index, in accordance with certain embodiments of the present disclosure;

[0029] FIG. 4 is a flow diagram to determine a multi-level health index, in accordance with certain embodiments of the present disclosure; and

[0030] FIG. 5A, FIG. 5B, FIG. 5C, FIG. 5D, FIG. 5E, FIG. 5F, FIG. 5G, FIG. 5H, FIG. 51, and FIG. 5J are illustrations of user interfaces displaying varying levels of a multi-level health index, in accordance with certain embodiments of the present disclosure.

DETAILED DESCRIPTION

[0031] So that the manner in which the features and advantages of the embodiments of the systems and methods disclosed herein, as well as others that will become apparent, may be understood in more detail, a more particular description of embodiments of systems and methods briefly summarized above may be had by reference to the following detailed description of embodiments thereof, in which one or more are further illustrated in the appended drawings, which form a part of this specification. It is to be noted, however, that the drawings illustrate only various embodiments of the systems and methods disclosed herein and are therefore not to be considered limiting of the scope of the systems and methods disclosed herein as it may include other effective embodiments as well.

[0032] Applicant has recognized these problems and others in the art, and has recognized a need for enhanced systems and methods to determine a multi-level health index for each of one or more industrial plants and, based on each level of the multi-level health index, detect component and/or equipment level anomalies and/or failures. Further, the multi-level health index may enable assessment of a plurality of components at a plurality of plants.

[0033] The present disclosure generally relates to systems and methods that address the relevant issues as described above, among other issues. In particular, such systems and methods may enable a user to monitor overall health of a plurality of industrial plants and, by selecting a lower level of the multi-level health index, monitor health at the component level for each industrial plant. An industrial plant, in an example, may include a chemical plant and/or a manufacturing plant, among other plants configured to produce chemicals and/or products. Such a system or method may reduce time by creating a centralized multi-level health index via one platform, which is not typical due to each industrial plant (a) generating a large data set daily, as most components run a substantial part of or even all of the day, and (b) including a plurality of different components and equipment. Further, such systems and methods may automatically monitor health based on historical conditions or historical optimal conditions and may automatically generate alerts for anomalies from the component level to the plant level.

[0034] Such systems and methods may capture data in real-time and/or, in another embodiment, may acquire or capture data from previous or historical industrial plant operations (for example, from a database or other type of storage), and preprocess the captured data to form a preprocessed data set. Such preprocessing may include applying the data to a first principle model (also referred to as a physics based model) and/or obtaining historical outputs from a first principle model. Such systems and methods may include applying the preprocessed data to one or more machine learning models to generate one or more classifiers or trained machine learning model. Examples of a first principle model include, but are not limited to, a model based on pump hydraulic power (utilizing, for example, Power_hydrauUc and a model based on a compressor pressure ratio (utilizing, for

example, Pressure Ratio_stage = among other first principle models based on other equations

related to one or more other characteristics of components and/or equipment located at a plant.

[0035] For example, a first principle model based on a compressor pressure ratio, as noted above, may utilize a selected formula (listed above). In such examples, the systems and methods described herein may gather various data points to determine a curve or graph representing such an equation or function, such as P_d and P_s. In such embodiments, the P_d may represent the discharge pressure over time and P_smay represent suction pressure over time. Thus, a model utilizing Pressure Ratio_stage of a compressor over a selected time period or frame may represent the discharge pressure of the compressor divided by the suction pressure of the compressor over that selected time period or frame. [0036] In another example, a first principle model based on pump hydraulic power may be utilized. In such an example, the equation or function (listed above) may utilize Q and AP. Q, in such examples may represent the pump flow in meters cubed per hour, while AP may represent discharge pressure minus suction pressure. In such examples, the first principle model based on such equations may represent kilowatt hours of the pump over time, utilizing the factors and equation described above. As noted, such examples are non-limiting and it will be understood that other simple and/or complex first principle models may be utilized.

[0037] After generating one or more trained machine learning models (and/or classifiers), the systems and methods may capture real-time data from each of a plurality of industrial plants. The data may include raw sensor data corresponding to various components and/or equipment. The systems and methods may apply the raw sensor data to a corresponding first principle model, to generate a first principle output. The systems and methods may then apply the raw sensor data and the first principle outputs to a corresponding trained machine learning model. After such application for each of a plurality of components for each of the plurality of corresponding equipment, the systems and methods may generate a multi-level health index, based on the output from the trained machine learning model.

[0038] The multi-level health index may include component level health index values based on the output of the trained machine learning models, as well as, in some embodiments, the first principle model outputs. Further, each component level health index may be combined to generate an over-all health level index. In such embodiments, each component level health index may be weighted the same. In other embodiments, each component level health index may be weighted differently based on the type of component and the type of equipment corresponding to the component. For example, some equipment may operate continuously or substantially continuously and failure to operate continuously may result in loss or decrease in production and, in another example, some components may be used in operation, while other components are ancillary and can be services and/or replaced without impacting operation or production. As such, the overall health index may be impacted more or less based on the type of equipment and/or the type of component. As such, a top level of the multilevel health index may correspond to all the industrial plants, a next level of the multi-level health index may correspond to a selected industrial plant, a second next level of the multi-level health index may correspond to equipment at the selected plant, and/or a third next level of the multi-level health index may correspond to a component level corresponding to equipment at the selected plant. Thus, a user may “drill down” or traverse through the multi-level health index to root cause and/or predict equipment and/or component anomalies or failures.

[0039] Further, the multi-level health index may predict future maintenance or service cycles for selected equipment and/or components. For example, after application of the data to a trained machine learning model, the systems and methods may utilize the output to generate a prediction regarding when a particular component, corresponding equipment, and/or corresponding plant may be maintained, serviced, and/or replaced. In other word, the multi-level health index may include details and/or information regarding or indicating potential component, equipment, and/or plant issues.

[0040] Further still, the systems and methods may automatically alert a user when any health index from any level of the multi-level health index is outside of a selected range or falls below a selected threshold, indicating different levels and/or time frames of maintenance to be performed (for example, immediate, within a week, within a month, within a year, and/or within other time frames). In an embodiment, the threshold may be based on historical conditions or historical optimal conditions.

[0041] In other embodiments, the systems and methods may include a platform. The platform may monitor for and/or be configured to receive new industrial plants to be monitored via the multi-level health index. The platform may be configured to enable a user to enter credentials and/or other access data to cause the platform to connect to a controller and/or computing device at an industrial plant, thus allowing the platform to monitor the industrial plant and/or generate a multi-level health index for the industrial plant. The platform, in an embodiment, may include an interactive display and/or user interface that enables user to select various levels of the multi-level health index.

[0042] Thus, utilizing the systems and methods described, a user may quickly diagnose and/or be alerted, of an impending maintenance cycle or equipment failure. Further, the health index generated by such systems and methods brings visibility as to particular components and/or equipment that should receive focus, in relation to maintenance and/or remediation, as the health index indicates an issue at a particular plant, at particular equipment, and at a particular component..

[0043] The following definitions are provided for clarifying certain terms and phrases of the present disclosure and are in no way intended to unnecessarily or unduly limit any embodiments and aspects related thereto.

[0044] The use of the words “a” or “an” when used in conjunction with the term “comprising,” “including,” “containing,” or “having” in the claims or the specification may mean “one,” but it is also consistent with the meaning of “one or more,” “at least one,” and “one or more than one.” [0045] The words “comprising” (and any form of comprising, such as “comprise” and “comprises”), “having” (and any form of having, such as “have” and “has”), “including” (and any form of including, such as “includes” and “include”) or “containing” (and any form of containing, such as “contains” and “contain”) are inclusive or open-ended and do not exclude additional, unrecited elements or method steps.

[0046] FIG. 1A and FIG. IB are schematic diagrams of a system to determine a multi-level health index, in accordance with certain embodiments of the present disclosure. Turning first to FIG. 1 A, an equipment health care system 102 may include a processor 104 or a plurality of processors and a memory 106. The memory 106 may store instructions. Further, the equipment health care system 102 may connect to or be in signal communication with one or more plants 118A, 118B, and up to 118N, such as an industrial plant, chemical plant, and/or manufacturing plant. Such an connection, in an embodiment, may be encrypted. In other words, data transmitted between the one or more plants 118A, 118B, and up to 118N and the equipment health care system 102 may be encrypted (in other words, the data may be converted into an encrypted format). The equipment health care system 102 may utilize an advanced encryption standard algorithm, a Rivest-Shamir-Adleman (RSA) algorithm and/or other data encryption algorithms, as will be understood by one skilled in the art. Such encryption algorithms may ensure that data transmission remains private.

[0047] The equipment health care system 102 may, in another embodiment, connect to a controller 120A, 120B, and up to 120N, one or more controllers, and/or a supervisory controller positioned at each plantl l8A, 118B, and up to 118N. The equipment health care system 102 may receive and/or obtain data from the plant 118A, 118B, and up to 118N via the controller 120A, 120B, and up to 120N. The equipment health care system 102 may also connect to a computing device 124, for example, the equipment health care system 102 may connect to a user’s computing device (such as computing device 124) via a user interface 122. In such examples, the user may interact with the equipment health care system 102 via the user interface 122.

[0048] As noted, the memory 106 may store instructions and/or various algorithms, models, and/or classifiers. In an embodiment, the instructions may include preprocessing instructions 108. In an embodiment, the equipment health care system 102 may, in addition to using the trained machine learning models to generate a multi-level health index, train one or more of the machine learning models 110. For example, the equipment health care system 102 may receive historical data from one or more of the plants 118A, 118B, and up to 118N, the computing device 124, and/or from a database. In another embodiment, the equipment health care system 102, as noted, may include trained machine learning models (for example, machine learning models 110). As such, to generate the multi-level health index, the equipment health care system 102 may receive data from each of the plants 118A, 118B, and up to 118N in real-time. In other words, as data is generated via a chemical operation, the equipment health care system 102 may receive that data and, further, that data may be received from a plurality of plants (for example, plants 118A, 118B, and up to 118N). Prior to applying the data to a machine learning model, the equipment health care system may preprocess the data via preprocessing instructions 108. Preprocessing the data may include removing noise from the data, removing outlying data points, and/or applying the data to a corresponding one of a plurality of first principle models 112.

[0049] In an embodiment, the equipment health care system 102 may be configured to receive large amounts of data, such as terabytes, petabytes, or, in some examples, even more. For example, one plant may generate that amount of data per day (as noted, terabytes, petabytes, or, in some examples, even more). Further, the equipment health care system 102 may connect to many plants (for example, plants 118A, 118B, and up to 118N), thus increasing the data received based on the amount of plants being monitored.

[0050] Once the data has been preprocessed, the health care instructions 114 may be executed. The health care instructions 114, when executed by the processor 104, may cause the preprocessed data to be applied to one of the first principle models (for example, first principle model 112). The first principle model 112 (also referred to as a physics based model) may include models based on fundamental physical and chemical principles. Such first principle models 112 may be based on, in examples, complex equations, rather than the typical simple first principle models, although simple first principle models may be utilized as well. In such embodiments, the first principle models 112, in particular complex equation based first principle models, may include product compositions and/or characteristics (for example, products including various fluids, such as gasses or liquids) first principle models produced at a particular plant. The product compositions and/or characteristics first principle models may be utilized to determine the health index for one or more components or equipment, in addition to machine learning model outputs and/or other first principle model outputs. Further, such product compositions and/or characteristics first principle models may be dynamic. In other words, the equations for and/or inputs to such a first principle model may change over time based on a type of product, a type of feed or feedstock, and/or desired characteristics for the final product.

[0051] Once the real-time data has been received and preprocessed, then the health care instructions may, when executed, cause the preprocessed data to be applied to a corresponding machine learning mode 110. In an embodiment, the output form application of the real-time data to the first principle models 112 may also be applied, along with the real-time data, the machine learning model. The machine learning model may produce a number, a probability, a curve and/or chart, and/or another value or indicator that may indicate whether a particular component and/or equipment may be maintained or otherwise serviced. Such an output may be utilized, by the health care instructions 114, to generate or form the multi-level health index. For example, the health care instructions 114 may first determine a health index for each component corresponding to each equipment at each corresponding plant. Then using the component health level index, the health care instructions 114 may determine an equipment health level. The health care instructions 114 may then use the equipment health level to determine a plant health level. Finally, the health care instructions 114 may determine the overall health index for all plants based on the health index for each plant. In an embodiment, the multi-level health index may reflect the health (for example, represented by a percentage or other number or indicator) for each corresponding component, equipment, and/or plant. In another embodiment, the health care instructions 114 may utilize various weights for different components and/or equipment to determine the total overall health of each plant 118A, 118B, and up to 118N. Such a weight may be based on the type of component and/or type of equipment the component is position within. Some components and/or equipment may be utilized frequently, as opposed to other components. Other equipment and/or components may be utilized continuously and the loss of such a component or equipment may impact production at a plant 118A, 118B, and up to 118N. As such, those components and/or equipment may be rated or weighted higher than other components. Thus a higher rated or weighted component, with a high health impact, may impact the overall health index more than a lower rated or weighted component.

[0052] In an embodiment, if the health of any component, equipment, and/or plant falls below a selected threshold or outside of a selected range, then the health care instructions 114 may cause an alert to be issued to the computing device 124 via the user interface 122. In an embodiment, the threshold may be based on historical conditions or historical optimal conditions. In another embodiment, the equipment health care system 102 may automatically cause a corrective action, such as maintenance, to occur via the controller 120A, 120B, and up to 120N of the plant 118A, 118B, and up to 118N. Other corrective actions may include replacement of the equipment, replacement of components, maintenance of the equipment, and/or, in some examples, shutting the equipment down. In another embodiment, the equipment health care system 102 may utilize the health index to predict failure of equipment and/or components and, in response to such a prediction, generate an alert indicating the potential failure and a corrective action. The equipment health care system 102 may determine the potential failure based on gradual changes in the health index.

[0053] Once the multi-level health index is available, the equipment health care system 102 may execute the visualization instructions! 16. The visualization instructions 116 may cause the equipment health care system 102 to generate the user interface 122. The user interface may include numerous interfaces depicting different components and/or equipment (for example, as schematic and/or three dimensional drawings). Further, some interfaces may be rendered with a visual or schematic illustration of the component and/or equipment with health overlayed throughout.

[0054] In an embodiment, and as illustrated in FIG. IB, each plant 118A may include a plurality of equipment 126A, 126B, and up to 126N. Each piece of equipment 126A, 126B, and up to 126N may include a plurality of components 128A, 128B and up to 128N and/or a plurality of sensors 130A, 130B, and up to 130N (and/or, in other embodiments, other meters or measuring devices), further each plant 118A may include a computing device or controller 120A to manage, control, and/or capture data associated with each piece of equipment 126A, 126B, and up to 126N. For example, the controller 120 A may cause a component, in one example, to shut down. In another embodiment, the controller 120A may cause a user to be notified regarding a maintenance issue. In yet another embodiment, the controller 120A itself may perform maintenance and/or selected maintenance tasks. [0055] In an embodiment, a subset of the plurality of sensors 130A, 130B, and up to 130N may be grouped together based on the components associated with and/or corresponding to the subset of the plurality of sensors 130A, 130B, and up to 130N and/or based on the processes related to the subset of the plurality of sensors 130A, 130B, and up to 130N. In another embodiment, the subset of the plurality of sensors 130A, 130B, and up to 130N may be grouped together based on feedback from a user. As data is received in real time, the equipment health care system 102 may detect the subset of the plurality of sensors 130A, 130B, and up to 130N based on such groupings.

[0056] In some examples, the equipment health care system 102 may be a computing device. The term “computing device” is used herein to refer to any one or all of programmable logic controllers (PLCs), controllers, supervisory controllers, programmable automation controllers (PACs), industrial computers, servers, virtual computing devices or environments, desktop computers, personal data assistants (PDAs), laptop computers, tablet computers, smart books, palm-top computers, personal computers, smartphones, virtual computing devices, cloud based computing devices, and similar electronic devices equipped with at least a processor and any other physical components necessarily to perform the various operations described herein.

[0057] The term “server” or “server device” is used to refer to any computing device capable of functioning as a server, such as a master exchange server, web server, mail server, document server, or any other type of server. A server may be a dedicated computing device or a server module (e.g., an application) hosted by a computing device that causes the computing device to operate as a server. A server module (e.g., server application) may be a full function server module, or a light or secondary server module (e.g., light or secondary server application) that is configured to provide synchronization services among the dynamic databases on computing devices. A light server or secondary server may be a slimmed-down version of server type functionality that can be implemented on a computing device, such as a smart phone, thereby enabling it to function as an Internet server (e.g., an enterprise e-mail server) only to the extent necessary to provide the functionality described herein.

[0058] As used herein, a “non-transitory machine-readable storage medium” or “memory” may be any electronic, magnetic, optical, or other physical storage apparatus to contain or store information such as executable instructions, data, and the like. For example, any machine-readable storage medium described herein may be any of random access memory (RAM), volatile memory, nonvolatile memory, flash memory, a storage drive (e.g., hard drive), a solid state drive, any type of storage disc, and the like, or a combination thereof. The memory may store or include instructions executable by the processor.

[0059] As used herein, a “processor” or “processing circuitry” may include, for example one processor or multiple processors included in a single device or distributed across multiple computing devices. The processor (such as, processor 104 shown in FIG. 1) may be at least one of a central processing unit (CPU), a semiconductor-based microprocessor, a graphics processing unit (GPU), a field-programmable gate array (FPGA) to retrieve and execute instructions, a real time processor (RTP), other electronic circuitry suitable for the retrieval and execution instructions stored on a machine-readable storage medium, or a combination thereof.

[0060] In an embodiment, the one or more machine learning models 112 may be a supervised or unsupervised learning model. In an embodiment, the one or more machine learning models 112 may be based on one or more of decision trees, random forest models, random forests utilizing bagging or boosting (as in, gradient boosting), K-nearest neighbors, neural network methods, support vector machines (SVM), other supervised learning models, other semi-supervised learning models, other unsupervised learning models, or some combination thereof, as will be readily understood by one having ordinary skill in the art.

[0061] FIG. 2 is another schematic diagram of an apparatus to determine a multi-level health index, in accordance with certain embodiments of the present disclosure. Such a system may be comprised of a processing circuitry 202, a memory 204, a communications circuitry 206, a preprocessing circuitry 208, a machine learning model circuitry 210, a first principle modeling circuitry 212, a health index circuitry 214, and a visualization circuitry 216, each of which will be described in greater detail below. While the various components are only illustrated in FIG. 2 as being connected with processing circuitry 202, it will be understood that the apparatus 200 may further comprise a bus (not expressly shown in FIG. 2) for passing information amongst any combination of the various components of the apparatus 200. The apparatus 200 may be configured to execute various operations described herein, such as those described above in connection with FIG. 1 and below in connection with FIGS. 3 -5 J.

[0062] The processing circuitry 202 (and/or co-processor or any other processor assisting or otherwise associated with the processor) may be in communication with the memory 204 via a bus for passing information amongst components of the apparatus. The processing circuitry 202 may be embodied in a number of different ways and may, for example, include one or more processing devices configured to perform independently. Furthermore, the processor may include one or more processors configured in tandem via a bus to enable independent execution of software instructions, pipelining, and/or multithreading.

[0063] The processing circuitry 202 may be configured to execute software instructions stored in the memory 204 or otherwise accessible to the processing circuitry 202 (e.g., software instructions stored on a separate storage device). In some cases, the processing circuitry 202 may be configured to execute hard-coded functionality. As such, whether configured by hardware or software methods, or by a combination of hardware with software, the processing circuitry 202 represents an entity (e.g., physically embodied in circuitry) capable of performing operations according to various embodiments of the present disclosure while configured accordingly. Alternatively, as another example, when the processing circuitry 202 is embodied as an executor of software instructions, the software instructions may specifically configure the processing circuitry 202 to perform the algorithms and/or operations described herein when the software instructions are executed.

[0064] Memory 204 is non-transitory and may include, for example, one or more volatile and/or nonvolatile memories. In other words, for example, the memory 204 may be an electronic storage device (e.g., a computer readable storage medium). The memory 204 may be configured to store information, data, content, applications, software instructions, or the like, for enabling the apparatus 200 to carry out various functions in accordance with example embodiments contemplated herein.

[0065] The communications circuitry 206 may be any means such as a device or circuitry embodied in either hardware or a combination of hardware and software that is configured to receive and/or transmit data from/to a network and/or any other device, circuitry, or module in communication with the apparatus 200. In this regard, the communications circuitry 206 may include, for example, a network interface for enabling communications with a wired or wireless communication network. For example, the communications circuitry 206 may include one or more network interface cards, antennas, buses, switches, routers, modems, and supporting hardware and/or software, or any other device suitable for enabling communications via a network. Furthermore, the communications circuitry 206 may include the processing circuitry 202 for causing transmission of such signals to a network or for handling receipt of signals received from a network. The communications circuitry 206, in an embodiment, may enable reception of polymerization operation data (including, in an example, real-time plant data and/or historical plant data) and transmission of a multi-level health index and/or alerts.

[0066] The apparatus 200 may include preprocessing circuitry 208 configured to preprocess received data. Preprocessing received data may include removing noise from the real-time data. The preprocessing circuitry 208 may utilize processing circuitry 202, memory 204, or any other hardware component included in the apparatus 200 to perform these operations, as described in connection with FIG. 4 below. The preprocessing circuitry 208 may further utilize communications circuitry 206, as noted above, to gather data (such as real-time and/or historical plant data) from a variety of sources (for example, one or more different components, equipment, computing devices and/or controllers at a plant or plants; a database; and/or other data sources). The output of the preprocessing circuitry 208 may be transmitted to other circuitry of the apparatus 200 (such as the health index circuitry 214, machine learning modeling circuitry 210, and/or first principle modeling circuitry 212).

[0067] In addition, the apparatus 200 further comprises machine learning modeling circuitry 210. The machine learning modeling circuitry may be configured to train one or more machine learning models and/or apply data to the one or more trained machine learning models. The machine relearning modeling circuitry may receive historical data from, for example, the memory or communications circuitry (for example, one or more plants may provide data via the communications circuitry 206) Further, the machine learning modeling circuitry 210 may be configured to utilize the historical data and a known outcome to train one or more machine learning models. Further, the machine learning modeling circuitry 210 may utilize the outputs from the first principle modeling circuitry 212 for data to apply to the trained machine learning models. In another embodiment, the machine learning modeling circuitry 210 may apply current or real time data to trained machine learning models. Such an application of data may produce a number, probability, predictions, simulations, and/or charts or curves. Using such an output, the apparatus 200 may determine the health level index, for example, via the health index circuitry 214. The machine learning modeling circuitry 210 may utilize processing circuitry 202, memory 204, or any other hardware component included in the apparatus 200 to perform these operations, as described in connection with FIG. 4 below. The machine learning modeling circuitry 210 may further utilize communications circuitry 206 to gather data (for example, preprocessed data and/or an output from one or more first principle models) from a variety of sources (such as the preprocessing circuitry 208, first principle modeling circuitry 212, and/or other sources);. The output of the machine learning modeling circuitry 210 may be transmitted to other circuitry of the apparatus 200, such as the health index circuitry 214.

[0068] The apparatus 200 further comprises the first principle modeling circuitry 212 that may apply received data to one or more first principle models based on the type of data received and/or an indicator transmitted along with the data. The first principle models included therein may include complex first principle models based on various formula. Such an application of the data to the one or more trained machine learning models may produce an output including one or more predictions, probabilities, simulations (such as a simulated performance of a potential chemical product), and/or charts or graphs associated with a performance of a resulting chemical product (for example, the chemical product being a result of the chemical operations and the chemical composition). The first principle modeling circuitry 212 may utilize processing circuitry 202, memory 204, or any other hardware component included in the apparatus 200 to perform these operations, as described in connection with FIG. 4 below. The first principle modeling circuitry 212 may further utilize communications circuitry 206 to receive data from the machine learning modeling circuitry 210. The output of the first principle modeling circuitry 212 may be transmitted to other circuitry of the apparatus 200, such as the health index circuitry 214.

[0069] The apparatus 200 further comprises the health index circuitry 214 that may transmit data to the visualization circuitry 216 and/or the machine learning modeling circuitry 210 and/or first principle modeling circuitry 212. The health index circuitry 214 may determine the health index for components, equipment, and/or plants based on the machine learning modeling circuitry 210 outputs. For example, the following equation may be utilized to determine the component level health index:

Health Index (M_n) = (100% — ModelDeviation⁶ + 0.05 * [No. of open long term alerts])

In such embodiments, Health Index (M_n) may represent the health index for one or more components associated with equipment at corresponding plants. As shown, the health index for a component may be based on, in an example, the number of open long term alerts. In an embodiment, a long term alert may include an alert generated by the apparatus 200 that indicates an unresolved chronic issue in the component. In another embodiment, the health index may be based on a number of all open alerts. In yet another embodiment, alerts may not be utilized in the determination of the health index, rather, other data generated by a machine learning model and/or first principle model may produce a number, probability, statistic, or other indicator to indicate health of a particular component. In another embodiment, ModelDeviation⁶ may represent the machine learning output which indicates how far the real time data or output of one or more of the machine learning models is from the historical conditions or historical optimal conditions. In such embodiments, the health index equation for a component may be constructed in such a way that exhibits a behavior similar to a potential failure curve. In other words, as the deviation of the real time data or output increases, then a potential failure is more likely. Further, the use of ModelDeviation⁶ may prevent misrepresentation of the actual condition of a particular component.

[0070] The following equation may be used to determine the health index for equipment:

In such embodiments, the Health Index (An) may represent the health of a particular or selected equipment with a number of selected components (Mr). In an embodiment, each health index for each component may be weighted for corresponding equipment. In other words, the summation may include a corresponding weight for each corresponding component health index.

[0071] The following equation may be utilized for a plant level health indices:

_n :

[0.55 * N_Tierl + 0.35 * N_Tier2 + 0.1 * N_Tier2]

[0072] In such embodiments, the Health Index Pn) may represent the health index for a particular plant. The equation may include weighting based on different ‘tiers’ of equipment. In another embodiment, different values may be utilized for differently weighted components.

[0073] In an embodiment, the health index for a site, which may include one or more plants may be determined based on Health Index (S_n) =

. Such an equation may utilize the health index for

one or more plants that comprise a particular site. Similarly, the health index for all plants and/or sites 11 r c 1 may be determined based on Health Index (C_n) = ^N _N - hi ^an embodiment, for a site and/or all the plants and/or sites, the health index may include different weights for different plants. [0074] In an embodiment, each equation subsequent to the component equation may indicate health based on the actual health of the underlying components. In other words, the health index for one or more components may provide the basis for all other subsequent health indices. The use and/or monitoring of the overall health index, and lower level health indices, may prevent failures and ensure continuous operation of a plant or plurality of plants.

[0075] In embodiments, application of the data to the one or more trained machine learning models may produce an output including one or more predictions, probabilities, simulations, and/or charts or graphs. The health index circuitry 214 may utilize processing circuitry 202, memory 204, or any other hardware component included in the apparatus 200 to perform these operations, as described in connection with FIG. 4 below. The output of the health index circuitry 214 may be transmitted to other circuitry of the apparatus 200, such as the visualization circuitry 216.

[0076] The apparatus 200 further comprises the visualization circuitry 216 that may generate a user interface, allow a user to click through or drill down into the multi-level health index, and/or generate schematic and/or three dimensional visuals of components and/or equipment with health index overlays. In an embodiment, the user click through various screens as they move down the multilevel health index (as illustrated in FIGS. 5A-J). Further, the visualization circuitry 216 may enable the user to initiate or begin the maintenance. The visualization circuitry 216 may utilize processing circuitry 202, memory 204, or any other hardware component included in the apparatus 200 to perform these operations, as described in connection with FIG. 4 below. The visualization circuitry 216 may further utilize communications circuitry 206 to receive data from the health index circuitry 214.

[0077] Although components 202-216 are described in part using functional language, it will be understood that the particular implementations necessarily include the use of particular hardware. It should also be understood that certain of these components 202-216 may include similar or common hardware. For example, the preprocessing circuitry 208, the machine learning model circuitry 210, the first principle modeling circuitry 212, the health index circuitry 214, and the visualization circuitry 216 may each at times utilize of the processing circuitry 202, memory 204, or communications circuitry 206, such that duplicate hardware is not required to facilitate operation of these physical elements of the apparatus 200 (although dedicated hardware elements may be used for any of these components in some embodiments, such as those in which enhanced parallelism may be desired). Use of the terms “circuitry,” and “engine” with respect to elements of the apparatus therefore shall be interpreted as necessarily including the particular hardware configured to perform the functions associated with the particular element being described. Of course, while the terms “circuitry” and “engine” should be understood broadly to include hardware, in some embodiments, the terms “circuitry” and “engine” may in addition refer to software instructions that configure the hardware components of the apparatus 200 to perform the various functions described herein.

[0078] Although the preprocessing circuitry 208, the machine learning model circuitry 210, the first principle modeling circuitry 212, the health index circuitry 214, and the visualization circuitry 216 may utilize processing circuitry 202, memory 204, or communications circuitry 206 as described above, it will be understood that any of these elements of apparatus 200 may include one or more dedicated processors, specially configured field programmable gate arrays (FPGA), or application specific interface circuits (ASIC) to perform its corresponding functions, and may accordingly utilize processing circuitry 202 executing software stored in a memory or memory 204, communications circuitry 206 for enabling any functions not performed by special-purpose hardware elements. In all embodiments, however, it will be understood that the preprocessing circuitry 208, the machine learning model circuitry 210, the first principle modeling circuitry 212, the health index circuitry 214, and the visualization circuitry 216 are implemented via particular machinery designed for performing the functions described herein in connection with such elements of apparatus 200.

[0079] In some embodiments, various components of the apparatus 200 may be hosted remotely (e.g., by one or more cloud servers) and thus need not physically reside on the corresponding apparatus 200. Thus, some or all of the functionality described herein may be provided by third party circuitry. For example, a given apparatus 200 may access one or more third party circuitries via any sort of networked connection that facilitates transmission of data and electronic information between the apparatus 200 and the third party circuitries. In turn, that apparatus 200 may be in remote communication with one or more of the other components describe above as comprising the apparatus 200.

[0080] As will be appreciated based on this disclosure, example embodiments contemplated herein may be implemented by an apparatus 200 (or by a computing device 302). Furthermore, some example embodiments (such as the embodiments described for FIGS. 1 and 3) may take the form of a computer program product comprising software instructions stored on at least one non-transitory computer-readable storage medium (such as memory 204). Any suitable non-transitory computer- readable storage medium may be utilized in such embodiments, some examples of which are non- transitory hard disks, CD-ROMs, flash memory, optical storage devices, and magnetic storage devices. It should be appreciated, with respect to certain devices embodied by apparatus 200 as described in FIG. 2, that loading the software instructions onto a computing device or apparatus produces a special-purpose machine comprising the means for implementing various functions described herein.

[0081] FIG. 3 is a schematic diagram of a computing device 300 to determine a multi-level health index, in accordance with certain embodiments of the present disclosure. The computing device 300 may be or include a control system, such as a controller, one or more controllers, a PLC, a SCADA system, and/or other components to determine a multi-level health index and/or initiate maintenance and/or generate or transmit alerts based on the multi-level health index. The computing device 302 may include one or more processors (e.g., processor 304) to execute instructions stored in memory 306. In an example, the memory 306 may be a machine-readable storage medium.

[0082] As used herein, “signal communication” refers to electric communication such as hard wiring two components together or wireless communication, as understood by those skilled in the art. For example, wireless communication may be or include Wi-Fi®, Bluetooth®, ZigBee, forms of near field communications, or other wireless communication methods as will be understood by those skilled in the art. In addition, signal communication may include one or more intermediate controllers, relays, or switches disposed between elements that are in signal communication with one another.

[0083] As noted, the memory 306 may store instructions executable by the processor 304, to preprocess data, such as preprocessing instructions 308. The computing device 302 may connect to and/or receive real time data from one or more of a plants 324A, 324B and up to 324N (including sensors, components, equipment, and/or controllers at the plants 324A, 324B and up to 324N). The preprocessing instructions 308 may, in response to reception of such data, be executed to preprocess such data. The preprocessing instructions 308 may, upon execution, determine remove noise and/or outlying data points form the real time data.

[0084] The computing device 302 may include machine learning models 312 and first principle models 314. The machine learning models may include trained machine learning models configured to provide an output upon application of data thereto. The output may include a number, probability, percentage, graph, chart, and/or other indicator. The first principle model m314 may be based on a selected question and may produce a series of data points and/or chart/curves. Each machine learning model 312 may be trained as described herein. Each machine learning model 312 and/or first principle model 314 may correspond to a selected type of equipment and/or component.

[0085] The computing device 302 may include health index instructions 316 to generate the multilevel health index. The health index instructions 316 may apply data to the machine learning models 312 and/or first principle models 314 to generate the multi-level health index. The health index instructions 316 may first generate component level health indices and then proceed to generate equipment level health indices, followed by plant level health indices, and finally an all plant health index.

[0086] In an embodiment, the computing device 302 may include data visualization instructions 318. The data visualization instructions 318 may generate one or more interactive user interfaces 322, each illustrating a level of the multi-level health index and enabling a user to move up or down the multilevel health index. The computing device 302 may include alert and corrective action instructions 320. Such an alert and corrective action may be generated if any level within the multi-level health index falls below a selected threshold or is outside a range. Such an alert and corrective action may include instructions detailing how to resolve any potential issues involving components and/or equipment. In an embodiment, the threshold may be based on historical conditions or historical optimal conditions.

[0087] FIG. 4 is a flow diagram to determine a multi-level health index, in accordance with certain embodiments of the present disclosure. Unless otherwise specified, the actions of method 500 may be completed within system 100, apparatus 200, and/or computing device 302. Specifically, method 400 may be included in one or more programs, protocols, or instructions loaded into the memory 106 of the equipment health care system 102 and executed on the processor 104 or one or more processors. The order in which the operations are described is not intended to be construed as a limitation, and any number of the described blocks may be combined in any order and/or in parallel to implement the methods.

[0088] At block 402, the equipment health care system 102 may receive historical data. The historical data may include a desired outcome and/or first principle model outputs. Further, the equipment health care system 102 may receive the historical data from one or more sources, such as a plant, a controller, a database, and/or other source. At block 404, the equipment health care system 102 may train one or more machine learning models using the historical data. In an embodiment, the historical data may include (a) historical raw sensor data from equipment at the plurality of industrial plants, (b) historical first principle model outputs associated with the plurality of industrial plants, (c) maintenance records of the equipment at the plurality of industrial plants, (d) failure records of the equipment in the plurality of industrial plants, or (e) one or more identifiers to indicate one or more selected equipment, selected chemicals, an equipment type, or a chemical type. [0089] Once one or more machine learning models have been trained, at block 406, the equipment health care system 102 may receive data in real-time from one or more plants. At block 408, the equipment health care system 102 may apply the real-time data to first principle model. At block 410, the equipment health care system 102 may generate a multi-level health index based on application of the real-time data and/or the first principle model output to one or more of the trained machine learning models.

[0090] At block 412, the equipment health care system 102 may determine whether any level within the multi-level health index is outside of historical conditions, historical optimal conditions, or a selected range. In an embodiment, the selected range may be a value or values based on the historical conditions. The selected range and/or historical conditions may be determined by an algorithm or instructions stored in the equipment health care system 102 or, in another example, by a user. If an algorithm or instructions are utilized to determine the selected range or historical conditions, then, in an embodiment, the equipment health care system 102 may wait until a response is received from another user verifying the selected range prior to use of the selected range. In yet another embodiment, the historical conditions may include conditions from a period of time when a component or equipment is operating at a healthy condition, for example, after repair or overhaul of a component or equipment.

[0091] In a non -limiting example, a compressor’s bearing temperature may be between about 30 degrees Celsius to about 40 degrees Celsius during operation. Such a range may be the selected range, historical conditions, or historical optimal conditions. However, during the compressor’s operation, the bearing temperature may increase past that range. If the temperature increases past the selected range, then the equipment health care system 102 may generate an alert. In another example, if the temperature increases gradually over time, but remains within the range, then the equipment health care system 102 may generate another alert indicating such a behavior.

[0092] If the multi-level health index is outside of the selected range for a component, equipment, and/or plant, then the equipment health care system 102, at block 414, may generate an alert for the applicable component, equipment, and/or plant. The alert may indicate that the health index from any level of the multi-level health index is outside of a selected range or falls below a selected threshold. In some embodiments, at block 416, the equipment health care system 102 may determine whether a corrective action is available. The equipment health care system 102 may select one or more corrective actions from a predefined set of different corrective actions. In some embodiments, the type of corrective action is selected based on the multi-level health index and/or a particular operation, behavior, and/or condition exhibited by the equipment (e.g., whether the equipment’s operation, behavior, and/or condition is within or outside of historical conditions, historical optimal conditions, or the selected range). If one or more corrective actions are available, then the equipment health care system 102 may include one or more corrective actions in the alert. For example, the corrective action may include guidance to (i) inspect the equipment or component(s) of the equipment, (ii) replace the equipment or the component(s) and/or (iii) perform preventative maintenance on such equipment or component s). The equipment health care system 102 may also include additional information within the alert for the corrective action, such as part availability and/or time needed to address the issue.

[0093] In some embodiments, the equipment health care system 102 may, at block 418, perform such a corrective action and/or indicate performing the corrective action. Some corrective actions may be able to be performed by the equipment health care system 102, in which case, the equipment health care system 102 performs such a corrective action. Such corrective actions that can be performed by the equipment health care system 102 may include (i) disabling the equipment/components and relying on back-up equipment/components and/or (ii) operating the equipment/components at different operating conditions (e.g., different, lower or higher settings, power, pressure, and/or temperature). In some embodiments, the equipment health care system 102 may automatically initiate the corrective actions. In other embodiments, the equipment health care system 102 may initiate them only after input and direction by a user through a user interface.

[0094] In yet another embodiment, for each component, and in other examples in addition to sensors associated with such components, all potential failures may be listed in an alert. Further, the alert may include general remedy actions associate with each failure mode. For example, if a compressor bearing temperature alert is generated, such an alert may include that a potential issue is insufficient cooling, as well as the resolution for such a potential issue.

[0095] In another embodiment, at block 420, the equipment health care system 102 may determine whether a new plant has been detected. The equipment health care system 102 may make such a determination based on an indicator provided by a user or a signal received from the corresponding new plant. If a new plant is detected, and if no corresponding machine learning models are available, then the equipment health care system 102 may request historical data and train new machine learning models with the historical data. In an embodiment, historical data may include (a) historical raw sensor data from equipment at the plurality of industrial plants, (b) historical first principle model outputs associated with the plurality of industrial plants, (c) maintenance records of the equipment at the plurality of industrial plants, (d) failure records of the equipment in the plurality of industrial plants, or (e) one or more identifiers to indicate one or more selected equipment, selected chemicals, an equipment type, or a chemical type.

[0096] FIG. 5 A, FIG. 5B, FIG. 5C, FIG. 5D, FIG. 5E, FIG. 5F, FIG. 5G, FIG. 5H, FIG. 51, and FIG. 5J are illustrations of user interfaces displaying varying levels of a multi-level health index, in accordance with certain embodiments of the present disclosure. FIGS. 5A-5J illustrate graphical user interfaces (GUIs) corresponding to one or more different levels of the multi-level health index. In FIG. 5A, a first GUI 502 illustrates a map with multiple plants located globally. Each region in such an example is selectable, allowing a user to view plants within that selected region. Further, the overall health index is available for the entire scope of monitored plants.

[0097] If a user selects a particular region, as illustrated in FIG. 5B, the GUI 504 may illustrate particular details for that selected region. Additional data may be shown as well, such as active alerts, overdue investigations, and/or compliance data, among other data points.

[0098] FIG. 5C illustrates another view (see GUI 506) of multiple plants, which may be grouped regionally and/or globally, or, in other embodiments, based on a filter selected by the user. If a user selects a plant, than the GUI 508 may illustrate the next level of the multi-level health index. For example, in FIG. 5D, various equipment for a particular plant that was selected is listed. The GUI 506 may show various data points and alerts for that particular plant, as well as identifying characteristics (for example, type of products produced from that plant). A user may also select different views for a plant, such as a view with charts illustrating performances and/or health for that plant, as shown in FIG. 5E (see GUI 510). FIG. 5F and FIG. 5G, are yet other plant alert GUIs 512, 514, which lists plant alert statuses and/or incidents for a selected plant.

[0099] If a user selects particular equipment, the GUI 516, 520 may transition to an equipment view, as shown in FIG. 5H and FIG. 5J. The equipment may be shown as a two or three dimensional schematic drawings. Further, various health indices may be overlay ed throughout the GUI 516. A user may also select a view that illustrates received sensor data, as shown in GUI 518 in FIG. 51.

[0100] While particular terms and concepts are incorporated in the present disclosure, Applicant notes that the disclosed terms and concepts are exclusively utilized in a descriptive capacity and should not therefore be construed or interpreted as limiting in any way. Certain embodiments and aspects of the disclosed systems, processes and methods have been described in detail with particular reference to the illustrated embodiments. However, it will be apparent that numerous and various modifications and alterations may be made within the spirit and scope of the embodiments of systems, processes and methods described herein, and such modifications and changes are to be considered equivalents and within the breadth and scope of the disclosure.

Claims

CLAIMS What is claimed is:

1. A method to detect equipment anomalies or equipment failure based on a multi-level health index for equipment positioned at one of a plurality of industrial plants, the method comprising: training a plurality of machine learning models with historical data associated with a plurality of industrial plants to form a plurality of trained machine learning models, the historical data to include one or more of (a) historical raw sensor data from equipment at the plurality of industrial plants, (b) historical first principle model outputs associated with the plurality of industrial plants, (c) maintenance records of the equipment at the plurality of industrial plants, (d) failure records of the equipment in the plurality of industrial plants, or (e) one or more identifiers to indicate one or more selected equipment, selected chemicals, an equipment type, or a chemical type; receiving, in real time, a plurality of raw sensor data points from the equipment at the plurality of industrial plants applying data points to one or more first principle models to produce a plurality of outputs from one or more first principle models associated with the plurality of industrial plants; determining the multi-level health index based on application of one or more of (a) one or more of the plurality of raw sensor data points or (b) one or more of the plurality of outputs from the one or more first principal models to a selected one or more of the plurality of trained machine learning models; comparing, in real-time, each level of the multi-level health index to historical conditions to detect one or more of equipment anomalies or equipment failure; and in response to one level of the multi-level health index being outside of the historical conditions, generating an alert.

2. The method of claim 1, further comprising: monitoring for detection of a new industrial plant; and in response to detection of the new industrial plant, determining, based on a type of equipment positioned at the new industrial plant, if the plurality of trained machine learning models correspond to equipment positioned at the new industrial plant.

3. The method of claim 2, wherein, if equipment at the new industrial plant does not correspond to the plurality of trained machine learning models: obtaining, via the new industrial plant, new plant historical data, the new plant historical data to include one or more of (a) historical raw sensor data from the equipment at the new industrial plant with no corresponding trained machine learning model, (b) historical first principle model outputs associated with the equipment at the new industrial plant with no corresponding trained machine learning model, (c) maintenance records of equipment at the new industrial plant with no corresponding trained machine learning model, (d) failure records of the equipment in the new industrial plant with no corresponding trained machine learning model, or (e) the one or more identifiers to indicate one or more selected equipment, selected chemicals, the equipment type, or the chemical type at the new industrial plant; and training one or more new machine learning models with the new plant historical data.

4. The method of claim 1, wherein the multi-level health index comprises a plurality of health indices that indicate a status of each of the plurality of the industrial plants, one or more equipment at each of the plurality of the one or more of the industrial plants, and one or more component of each equipment at each of the plurality of the one or more of the industrial plants.

5. The method of claim 1, further comprising: generating a user interface, the user interface to include an interactive display and to initially display a top level of the multi-level health index for each of the plurality of the industrial plants; and in response to selection of one of the plurality of the industrial plants on the interactive display, displaying a next level of the multi-level health index for each equipment at the one of the plurality of the industrial plants.

6. The method of claim 5, further comprising: in response to selection of one of each equipment at the one of the plurality of the industrial plants, generating a schematic view of the one of each equipment, wherein the schematic view includes each component of the one of each equipment and a corresponding level of the multi-level health index.

7. The method of claim 1, wherein the alert includes a corrective action selected from a predefined set of different corrective actions.

8. A method to detect equipment anomalies or equipment failure based on a multi-level health index for a plurality of plants, the method comprising: obtaining real-time component level data from a plurality of equipment positioned at one of the plurality of plants; determining a plurality of subsets of the real-time component level data based on a type of component and a type of equipment; determining, based on application of each of the plurality of subsets to one or more machine learning models or first principle models that correspond to each of the plurality of subsets, each level of the multi-level health index for each of the plurality of plants; generating a top level of the multi-level health index for each of the plurality of plants based on each determined level of the multi-level health index for each of the plurality of plants; comparing, in real-time, each level of the multi-level health index to corresponding historical conditions to detect one or more of equipment anomalies or equipment failure; and in response to a comparison that is not within the corresponding selected threshold range of each level of the multi-level health index for each of the plurality of plants with a historical optimal condition, generating an alert.

9. The method of claim 8, wherein each level of the multi-level health index for each of the plurality of plants is based on one or more of a corresponding type of equipment, a weight for the corresponding type of equipment, a corresponding type of component, or a weight for the corresponding type of component.

10. The method of claim 8, wherein the historical condition comprises one or more of previous maintenance, previous pauses in production, previous equipment failures, previous chemical output, or use of redundant equipment.

11. The method of claim 8, wherein a top level of the multi-level health index of a selected plant indicates a current operational status of the selected plants.

12. The method of claim 8, wherein the multi-level health index indicates a failure prediction time frame of selected equipment, and wherein the alert includes a next action to prevent failure of the selected equipment.

13. A system to detect equipment anomalies or equipment failure based on a multi-level health index for a plurality of plants, the system comprising: a communications circuitry configured to: obtain, in real-time, component level data corresponding to a plurality of equipment positioned at one of a plurality of plants, obtaining real-time component level data from a plurality of equipment positioned at one of the plurality of plants; a preprocessing circuitry configured to: determine a plurality of subsets of the component level data based on a type of component, a type of equipment, or a type of plant; a modeling circuitry configured to: generate a plurality of probabilities based on application of each of the plurality of subsets to one or more corresponding machine learning models; and a health index circuitry configured to: determine a health index for each component based on the plurality of probabilities, generate a multi-level health index for each of the plurality of plants based on the health index for each component, and monitor for one or more of equipment anomalies or equipment failure based on a comparison of each level of the multi-level health index to a corresponding selected range.

14. The system of claim 13, wherein the multi-level health index comprises a component level health index, an equipment level health index, and a plant level health index.

15. The system of claim 14, wherein the component level health index comprises each health index for each component, wherein the equipment level health index comprises a second health index for each equipment based on the health index of each component associated with corresponding equipment, and wherein the plant level health index comprises a third health index for each plant based on the second health index of each equipment associated with a corresponding plant and wherein the second health index for each equipment is further based on a weight associated with each component of each equipment, and the third health index for each plant is further based on a weight associated with each equipment positioned at the corresponding plant.