Method and system for predicting KPI values, facility status and alarm in an industrial

Abstract

A predetermined key performance indicator (KPI) of the industrial process may be predicted. The plurality of tags each identify a corresponding KPI of the industrial process and a historical value of the KPI identified by the plurality of tags. A KPI forecasting model is trained for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecasting models is trained based at least in part on the received historical values of at least some of the KPIs identified by the plurality of tags. A forecasted KPI value is generated for each of the KPIs identified by the plurality of tags based at least in part on a corresponding KPI forecasting model corresponding to the respective KPI.

Description

Technical Field

[0001] The present disclosure relates to industrial processes, and more particularly to methods and systems for predicting KPI values, facility states, and / or alarms in industrial processes. Background Art

[0002] Industrial processes of a wide range are known. Many industrial processes are large and complex processes that may be difficult to monitor and control, in part because industrial processes may involve a large number of different equipment, each of which may generate a large amount of operating data. For example, in complex process facilities such as refineries or petrochemical complexes, operators may be responsible for monitoring thousands of parameters such as pressure, liquid level, flow and temperature. Current automation systems cannot take all necessary actions when deviations occur, and require human-machine interaction. Experienced operators with many years of service can often manage industrial processes quite well by taking appropriately timed control actions when deviations occur. However, some operators may not have experience and may not have experience in handling every possible deviation that may occur. Taking inappropriate actions may cause deviations to become more serious, or may cause deviations of other parameters. What is desired is a method and system for monitoring and controlling industrial processes. What is desired is a method and system for predicting KPI values, facility status and / or alarms in industrial processes to help operators take appropriately timed actions to properly manage industrial processes. Summary of the invention

[0003] The present disclosure relates to industrial processes, and more particularly to methods and systems for predicting KPI values, facility states, and / or alarms in industrial processes. An example may be found in a method for predicting a predetermined key performance indicator (KPI) of an industrial process. The illustrative method includes: receiving a plurality of tags each identifying a corresponding KPI of an industrial process; and receiving historical values ​​of those KPIs identified by the plurality of tags. A KPI forecast model is trained for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecast models is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags. A predicted KPI value is generated for each of the KPIs identified by the plurality of tags, based at least in part on the corresponding KPI forecast model corresponding to the corresponding KPI.

[0004] Another example may exist in a system for predicting a predetermined key performance indicator (KPI) of an industrial process. The system includes an I / O port, a memory, and a controller operatively coupled to the I / O port and the memory. The controller is configured to receive a plurality of tags that each identify a corresponding KPI of an industrial process via the I / O port, and to receive historical values ​​of those KPIs identified by the plurality of tags via the I / O port. The controller is configured to train a KPI forecast model for each KPI identified by the plurality of tags, wherein each KPI forecast model in the KPI forecast model is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags. The controller is configured to generate a predicted KPI value for each KPI identified by the plurality of tags based at least in part on the corresponding KPI forecast model corresponding to the corresponding KPI.

[0005] Another example may exist in a non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to receive a plurality of tags that each identify a corresponding KPI of an industrial process and receive historical values ​​of those KPIs identified by the plurality of tags. Cause the one or more processors to train a KPI forecast model for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecast models is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags. Cause the one or more processors to generate a forecasted KPI value for each of the KPIs identified by the plurality of tags based at least in part on the corresponding KPI forecast model corresponding to the respective KPI.

[0006] The above summary is provided to facilitate understanding of some innovative features unique to the present disclosure, but is not intended to be a complete description. A comprehensive understanding of the present disclosure can be obtained by viewing the entire specification, claims, drawings and abstract as a whole. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] The present disclosure may be more fully understood by considering the following description of various examples in conjunction with the accompanying drawings, in which:

[0008] Figure 1 is a schematic block diagram of an exemplary system for predicting KPIs for an industrial process;

[0009] Figure 2 is a flow chart illustrating an exemplary method for predicting a KPI;

[0010] Figure 3 is a flow chart illustrating an exemplary method for predicting a KPI;

[0011] Figure 4is a flow chart illustrating an exemplary method for predicting a KPI;

[0012] Figure 5A and Figure 5B are flow charts that collectively illustrate an exemplary method for predicting a KPI;

[0013] Figure 6 is a flow chart illustrating an illustrative method of forecasting a KPI of an industrial process over a specific extended time window;

[0014] Figure 7 is a flow chart illustrating an exemplary method of forecasting a facility status of an industrial process;

[0015] Figure 8 is a flow chart illustrating an exemplary method of forecasting an alarm for an industrial process; and

[0016] Fig. 9 is a flow chart showing an exemplary overall process flow.

[0017] Although the present disclosure is subject to various modifications and alternative forms, the details thereof have been shown by way of example in the accompanying drawings and will be described in detail. However, it should be understood that it is not intended to limit the present disclosure to the specific examples described. On the contrary, it is intended to cover all modifications, equivalents and alternatives that fall within the essence and scope of the present disclosure. DETAILED DESCRIPTION

[0018] The following description should be read with reference to the accompanying drawings, in which similar elements in different drawings are numbered in the same manner. The accompanying drawings are not necessarily drawn to scale, and they depict examples that are not intended to limit the scope of the present disclosure. Although examples of various elements are shown, those skilled in the art will recognize that many of the examples provided have suitable alternatives that can be utilized.

[0019] It is assumed herein that all numbers are modified by the term "about" unless the content clearly dictates otherwise. The recitation of numerical ranges by endpoints includes all numbers subsumed within that range (eg, 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, and 5).

[0020] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the content clearly dictates otherwise. As used in this specification and the appended claims, the term "or" is generally employed in its sense including "and / or" unless the content clearly dictates otherwise.

[0021] It should be noted that references to "one embodiment", "some embodiments", "other embodiments", etc. in the specification indicate that the described embodiments may include a particular feature, structure, or characteristic, but every embodiment may not necessarily include the particular feature, structure, or characteristic. Moreover, these phrases do not necessarily refer to the same embodiment. In addition, when a particular feature, structure, or characteristic is described in conjunction with one embodiment, it is contemplated that the feature, structure, or characteristic is described in conjunction with one embodiment, and it is contemplated that the feature, structure, or characteristic may be applied to other embodiments regardless of whether explicitly described, unless otherwise clearly stated.

[0022] Figure 1 1 is a schematic block diagram showing an exemplary system 10 for predicting predetermined key performance indicators (KPIs) of an industrial process 12. The system 10 includes an I / O port 14, a memory 16, and a controller 18 operatively coupled to the I / O port 14 and the memory 16. The industrial process 12 may generally represent any of a variety of different industrial processes. For example, the industrial process 12 may represent a chemical manufacturing process or a refining process. In some cases, the industrial process 12 may include a plurality of facilities 20, labeled 20a and 20b, respectively. Although two facilities 20 are shown, this is merely illustrative, as the industrial process 12 may include only a single facility 20, or may include three or more facilities 20. Each of the facilities 20 may represent a piece of equipment of the industrial process 12, such as a boiler, a heating furnace, a distillation column, etc. In the example shown, each of the facilities 20 includes a plurality of key performance indicators (KPIs). As shown, facility 20a includes multiple KPIs 22, labeled 22a, 22b, and 22c, and facility 20b includes multiple KPIs 24, labeled 24a, 24b, and 24c, respectively. This is merely illustrative, as facility 20a may include any number of KPIs 22, and may include far more than three KPIs, and facility 20b may include any number of KPIs 24, and may include far more than three KPIs 24.

[0023] The controller 18 is configured to receive a plurality of tags, each identifying a corresponding KPI 22, 24 of the industrial process 12, via the I / O port 14. In some cases, the plurality of tags only mark certain KPIs of the industrial process 12 that are considered critical to the operation and control of the industrial process 12. The controller 18 is configured to receive historical values ​​of the KPIs 22, 24 identified by the plurality of tags via the I / O port 14. The controller 18 is configured to train a KPI prediction model for each of the KPIs 22, 24 identified by the plurality of tags, wherein each of the KPI prediction models is trained based at least in part on the received historical values ​​of at least some of the KPIs 22, 24 identified by the plurality of tags. Each of the KPI prediction models in the KPI prediction models may be a model based on artificial intelligence (AI). In some cases, one or more of the KPI prediction models may be continuously trained using updated current values ​​and historical values ​​of at least some of the KPIs 22, 24, and thus may continuously learn and track the characteristics of the industrial process. In other cases, one or more of the KPI forecast models may be trained during the training phase, and then the trained models may be published for use during the operation phase. In either case, the controller 18 may be configured to generate a predicted KPI value for each of the KPIs 22, 24 identified by the plurality of tags based at least in part on the corresponding KPI forecast model corresponding to the respective KPI 22, 24. In some cases, the controller 18 may be configured to automatically adjust one or more parameters of the industrial process 12 based at least in part on one or more of the predicted KPI values.

[0024] In some cases, each of the KPI prediction models may be trained based at least in part on one or more of the predicted KPI values ​​and received historical values ​​of at least some of the KPIs 22, 24 of the industrial process 12 identified by the plurality of tags. In some cases, when the industrial process 12 includes a plurality of facilities 20, and each facility may include two or more tags of the plurality of tags that each identify a corresponding KPI 22, 24 of the corresponding facility 20, and each facility 20 may have two or more predetermined facility states. The controller 18 may be configured to train a facility state prediction model for each of the plurality of facilities 20, wherein each of the facility state prediction models is trained to predict the facility state of the corresponding facility 20 based at least in part on the received historical values ​​of the KPIs 22, 24 associated with the corresponding facility 20 and / or the predicted KPI values ​​of the KPIs 22, 24 associated with the corresponding facility 20. The controller 18 may be configured to generate a predicted facility state for at least one facility of the plurality of facilities 20 of the industrial process 12 based at least in part on a facility state prediction model corresponding to the respective facility 20 .

[0025] In some cases, the controller 18 may be configured to receive one or more alarm limits for at least some of the KPIs 22, 24 of the industrial process 12, and receive historical alarms and / or events associated with at least some of the KPIs 22, 24 of the industrial process 12. The controller 18 may be configured to train an alarm forecasting model for one or more of the KPIs 22, 24 of the industrial process 12, wherein the alarm forecasting model is trained based on the received historical values ​​and / or one or more of the forecasted KPI values ​​for one or more of the KPIs 22, 24, the historical alarms and / or events associated with at least some of the KPIs 22, 24, and one or more of the alarm limits for at least some of the KPIs 22, 24.

[0026] Figure 22 is a flow chart showing an exemplary method 26 for predicting a predetermined key performance indicator (KPI) of an industrial process, such as an industrial process 12. The exemplary method 26 includes receiving a plurality of tags, each of which identifies a corresponding KPI of the industrial process, as shown at box 28. In some cases, the plurality of tags only mark certain KPIs of the industrial process 12 that are considered critical to the operation and control of the industrial process 12. Receive historical values ​​of the KPIs identified by the plurality of tags, as shown at box 30. Train a KPI forecast model for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecast models is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags, as shown at box 32. Each of the KPI forecast models may be an artificial intelligence (AI) based model. In some cases, one or more of the KPI forecast models may be continuously trained using updated current values ​​and historical values ​​of at least some of the KPIs, and thus may continuously learn and track characteristics of the industrial process. In other cases, one or more of the KPI prediction models may be trained during the training phase, and then the trained models may be released for use during the operation phase. In either case, a predicted KPI value is generated for each of the KPIs identified by the plurality of tags based at least in part on the corresponding KPI prediction model corresponding to the respective KPI, as shown at block 34. In some cases, the method 26 may include automatically adjusting one or more parameters of the industrial process based at least in part on the one or more of the predicted KPI values, as shown at block 36.

[0027] Figure 3is a flow chart showing an exemplary method 38 for predicting a predetermined key performance indicator (KPI) of an industrial process, such as the industrial process 12. The exemplary method 38 includes receiving a plurality of tags, each of which identifies a corresponding KPI of the industrial process, as shown at box 40. In some cases, the plurality of tags only mark certain KPIs of the industrial process that are considered critical to the operation and control of the industrial process. Receive historical values ​​of the KPIs identified by the plurality of tags, as shown at box 42. Train a KPI forecast model for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecast models is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags, as shown at box 44. Each of the KPI forecast models may be an artificial intelligence (AI) based model. In some cases, one or more of the KPI forecast models may be continuously trained using updated current values ​​and historical values ​​of at least some of the KPIs, and thus may continuously learn and track characteristics of the industrial process. In other cases, one or more of the KPI forecast models may be trained during the training phase, and then the trained models are released for use during the operation phase. In either case, and in the example shown, a time for a forecast value is received, as shown at box 46. The time for forecasting represents a future target time for forecasting the forecasted KPI value. A forecasted KPI value is generated for each of the KPIs identified by a plurality of tags to correspond to a future time corresponding to the time for the forecast value, as shown at box 48. In some cases, each of the KPI forecast models may be trained based at least in part on one or more of the forecasted KPI values ​​and received historical values ​​of at least some of the KPIs identified by a plurality of tags for an industrial process.

[0028] Figure 4is a flow chart showing an exemplary method 50 for predicting a predetermined key performance indicator (KPI) of an industrial process, such as the industrial process 12. The exemplary method 50 includes receiving a plurality of tags, each identifying a corresponding KPI of the industrial process, as shown at box 52. In some cases, the plurality of tags only mark certain KPIs of the industrial process that are considered critical to the operation and control of the industrial process. Receive historical values ​​of the KPIs identified by the plurality of tags, as shown at box 54. Train a facility state prediction model for each of a plurality of facilities, wherein each of the facility state prediction models is trained to predict a facility state (e.g., normal, tripped, cold shutdown, hot shutdown) of the corresponding facility based at least in part on the received historical values ​​of the KPIs associated with the corresponding facility and / or the predicted KPI values ​​of the KPIs associated with the corresponding facility, as shown at box 56. Each of the facility state prediction models in the facility state prediction models can be an artificial intelligence (AI) based model. In some cases, one or more of the facility state prediction models can be continuously trained using updated current values ​​and historical values ​​of at least some of the KPIs associated with the corresponding facility, and thus can continuously learn and track the characteristics of the corresponding facility. In other cases, one or more of the facility state prediction models can be trained during the training phase, and then the trained model is released for use during the operation phase. In either case, a predicted facility state is generated for at least one of the multiple facilities of the industrial process based at least in part on the facility state prediction model corresponding to the corresponding facility, as shown in box 58. In some cases, each of the facility state prediction models can be used to predict the facility state of the corresponding facility based at least in part on the received historical values ​​of the KPIs associated with the corresponding facility and one or more of the predicted KPI values ​​of the predicted KPIs of one or more of the KPIs associated with the corresponding facility.

[0029] In some cases, the method 50 may also include receiving one or more alarm limits for at least some of the KPIs associated with each of the plurality of facilities, as shown at block 60. In some cases, each of the facility state prediction models may be trained to predict the facility state of the respective facility based at least in part on the received historical values ​​of the KPIs associated with the respective facility (and / or the predicted KPI values ​​of the KPIs associated with the respective facility) and the one or more alarm limits for at least some of the KPIs associated with the respective facility, as shown at block 62. In some cases, each of the facility state prediction models may be trained to predict the facility state of the respective facility based at least in part on the received historical values ​​of the KPIs associated with the respective facility, the one or more alarm limits for at least some of the KPIs associated with the respective facility, and one or more predicted KPI values ​​of the predicted KPI values ​​of one or more of the KPIs associated with the respective facility.

[0030] Figure 5A and Figure 5B 6 is a flowchart that collectively illustrates an exemplary method 64 for predicting a predetermined key performance indicator (KPI) of an industrial process, such as the industrial process 12. The exemplary method 64 includes receiving a plurality of tags that each identify a corresponding KPI of the industrial process, as shown at box 66. In some cases, the plurality of tags only mark certain KPIs of the industrial process that are considered to be critical to the operation and control of the industrial process. Receive historical values ​​of the KPIs identified by the plurality of tags, as shown at box 68. Train a KPI forecast model for each of the KPIs identified by the plurality of tags, wherein each of the KPI forecast models is trained based at least in part on the received historical values ​​of at least some of the KPIs identified by the plurality of tags, as shown at box 70. Each of the KPI forecast models may be an artificial intelligence (AI) based model. In some cases, one or more of the KPI forecast models may be continuously trained using updated current values ​​and historical values ​​of at least some of the KPIs, and thus may continuously learn and track characteristics of the industrial process. In other cases, one or more of the KPI prediction models may be trained during the training phase, and then the trained models are released for use during the operation phase. In either case, a predicted KPI value is generated for each of the KPIs identified by the plurality of tags based at least in part on the corresponding KPI prediction model corresponding to the respective KPI, as shown at block 72.

[0031] In some cases, the method 64 may include receiving one or more alarm limits for at least some of the KPIs for the industrial process, as shown at block 74. Historical alarms and / or events associated with at least some of the KPIs for the industrial process may be received, as shown at block 76. Figure 5B , the method 64 continues to train an alarm forecast model for one or more of the KPIs of the industrial process, wherein the alarm forecast model is trained based on received historical values ​​of one or more of the KPIs (and / or predicted KPI values ​​of the KPIs associated with the corresponding facilities), historical alarms and / or events associated with at least some of the KPIs, and one or more of the alarm limits of at least some of the KPIs, as shown at box 78. In some cases, the alarm forecast model may also be trained based on one or more of the predicted KPI values. In some cases, the method 64 may include generating a predicted alarm for each of the one or more of the KPIs based on the predicted KPI values ​​of the corresponding PKI, as shown at box 80.

[0032] In some cases, the method 64 may include identifying two or more of the historical alarms and / or events that overlap in time, resulting in one or more overlapping segments, as shown at block 82. The historical alarms and / or events that overlap in time during the one or more overlapping segments may be excluded, as shown at block 84. The alarm forecast model may be trained based on the received historical values ​​of one or more of the KPIs and / or the forecasted KPI values ​​of one or more of the KPIs, the historical alarms and / or events of at least some of the KPIs excluding at least some of the historical alarms and / or events that overlap in time during the one or more overlapping segments, and one or more of the alarm limits of at least some of the KPIs, as shown at block 86. In some cases, each of the historical alarms and / or events associated with at least some of the KPIs may identify a source of the historical alarm and / or event, a category of the historical alarm and / or event, and a condition of the historical alarm and / or event, and an alarm forecasting model may be trained based on the source, category, and / or condition of one or more of the historical alarms and / or events associated with at least some of the KPIs of the industrial process.

[0033] Figure 6is a flow chart showing an exemplary workflow 88 for predicting KPIs over time. The method 88 includes receiving tags and alarm configurations from a distributed control system (DCS) configuration 90 and obtaining historical KPI values ​​from a historical repository 92. The tags are seen at box 94. At box 96, a KPI forecasting model is trained for each tagged KPI, resulting in a plurality of trained KPI forecasting models 98, labeled 98a, 98b, and 98c, respectively. In some cases, there may be any number of trained KPI forecasting models 98, and in some cases, there may be far more than three trained KPI forecasting models 98. The trained KPI forecasting models 98 are provided to MARCK (Multivariate Autoregressive Combined KPI Forecaster) 100, which is a deep learning model using machine learning techniques. MARCK 100 outputs a predicted KPI value for each tagged KPI, as shown at box 102, resulting in a plurality of predicted KPI values ​​104, labeled 104a, 104b, and 104c. In some cases, the predicted KPI values ​​104 are used together with updated historical KPI values ​​from the historical repository 92 to retrain the trained KPI forecast model 98, sometimes continuously.

[0034] Figure 7 is a flow chart showing an exemplary workflow 106, which includes Figure 6 94, but also includes a box 108 that receives an identifier of the tagged KPI 94 from the DCS configuration 90 and receives historical data of the tagged KPI from the historian 92. Box 108 trains a facility state classification model and outputs a facility state classification model 110 that receives input from the trained KPI prediction model 98, as shown. The predicted KPI value 104 can be used to predict a facility state 112 (e.g., normal, tripped, cold shutdown, hot shutdown) using the facility state classification model 110.

[0035] Figure 8 is a flow chart showing an exemplary workflow 114, which includes Figure 6104, but also includes a box 116 that provides historical alarms and events from the archive. At box 118, labels are associated with alarm limits and this information is provided to box 120, where an alarm identification model is trained. The alarms and events from box 116 are provided to box 122, where overlapping and non-overlapping alarms and events are identified. At box 124, overlapping alarms and events are excluded, and the remaining alarms and events are provided to box 120. Information from the historical library 92 is provided to box 120, which trains and outputs the alarm identification model 120. Each of the predicted KPI values ​​104 (labeled 104a, 104b, and 104c, respectively) is provided to a corresponding alarm forecast box 126 (labeled 126a, 126b, and 126c, respectively), as shown. Each of the alarm forecast blocks 126a, 226b and 126c is configured to forecast whether the corresponding forecasted KPI value is expected to fall outside its alarm limit based on the alarm identification model 120 and the corresponding forecasted KPI value 104a, 104b and 104c. Fig. 9 is a flow diagram illustrating an exemplary workflow 128 representing the compilation of other workflows. Fig. 9 One addition is block 130 which outputs an alert forecast.

[0036] Although several exemplary embodiments of the present disclosure have been described in this manner, it will be readily understood by those skilled in the art that other embodiments may be made and used within the scope of the claims appended hereto. However, it should be understood that the present disclosure is in many respects merely illustrative. Details (especially details related to shape, size, arrangement of parts, and exclusion and order of steps) may be changed without exceeding the scope of the present disclosure. Of course, the scope of the present disclosure is defined in the language of the appended claims.

Claims

1. A method for predicting a predetermined key performance indicator (KPI) of an industrial process, the method comprising: receiving a plurality of tags each identifying a corresponding KPI of the industrial process; receiving historical values ​​of the KPI identified by the plurality of tags; training a KPI prediction model for each of the KPIs identified by the plurality of tags, wherein each of the KPI prediction models is trained based at least in part on received historical values ​​of at least some of the KPIs identified by the plurality of tags; as well as A predicted KPI value is generated for each of the KPIs identified by the plurality of tags based at least in part on a corresponding KPI prediction model corresponding to the respective KPI.

2. The method according to claim 1, further comprising: One or more parameters of the industrial process are automatically adjusted based at least in part on one or more of the forecasted KPI values.

3. The method according to claim 1, comprising: receiving the time for the forecast value; as well as Generating the forecasted KPI value for each of the KPIs identified by the plurality of tags includes generating the forecasted KPI value for each of the KPIs identified by at least some of the plurality of tags forecast at a future time corresponding to the time for forecast value.

4. The method according to claim 3, wherein each of the KPI prediction models is trained at least in part based on one or more of the predicted KPI values ​​and received historical values ​​of at least some of the KPIs of the industrial process identified by the multiple tags.

5. The method of claim 1, wherein the industrial process comprises a plurality of facilities, wherein each facility comprises two or more tags of the plurality of tags each identifying a corresponding KPI of the corresponding facility, and each facility has two or more predetermined facility states, the method comprising: training a facility state prediction model for each of the plurality of facilities, wherein each of the facility state prediction models is trained to predict the facility state of the corresponding facility based at least in part on the received historical value of the KPI associated with the corresponding facility and / or the predicted KPI value of the KPI associated with the corresponding facility; as well as A predicted facility state is generated for at least one facility of the plurality of facilities of the industrial process based at least in part on the facility state prediction model corresponding to the respective facility.

6. A method according to claim 5, wherein each of the facility status prediction models is used to predict the facility status of the corresponding facility based at least in part on the received historical values ​​of the KPI associated with the corresponding facility and one or more of the predicted KPI values ​​of one or more of the KPIs associated with the corresponding facility.

7. The method according to claim 5, further comprising: receiving one or more alarm limits for at least some of the KPIs associated with each of the plurality of facilities; and Each of the facility status prediction models is trained to predict the facility status of the corresponding facility based at least in part on received historical values ​​of the KPI associated with the corresponding facility and one or more alarm limits of at least some of the KPIs associated with the corresponding facility.

8. A method according to claim 7, wherein each of the facility status prediction models is trained to predict the facility status of the corresponding facility based at least in part on received historical values ​​of the KPI associated with the corresponding facility, one or more alarm limits of at least some of the KPIs associated with the corresponding facility, and one or more of the predicted KPI values ​​of one or more of the KPIs associated with the corresponding facility.

9. The method according to claim 1, further comprising: receiving one or more alarm limits for at least some of the KPIs for the industrial process; receiving historical alarms and / or events associated with at least some of the KPIs for the industrial process; as well as An alarm forecasting model is trained for one or more of the KPIs of the industrial process, wherein the alarm forecasting model is trained based on received historical values ​​of one or more of the KPIs, the historical alarms and / or events associated with at least some of the KPIs, and one or more of the alarm limits of at least some of the KPIs.

10. A system for predicting a predetermined key performance indicator (KPI) of an industrial process, the system comprising: I / O ports; Memory; a controller operatively coupled to the I / O port and the memory, the controller being configured to: receiving, via the I / O port, a plurality of tags each identifying a corresponding KPI of the industrial process; receiving, via the I / O port, historical values ​​of the KPI identified by the plurality of tags; training a KPI prediction model for each of the KPIs identified by the plurality of tags, wherein each of the KPI prediction models is trained based at least in part on received historical values ​​of at least some of the KPIs identified by the plurality of tags; as well as A predicted KPI value is generated for each of the KPIs identified by the plurality of tags based at least in part on a corresponding KPI prediction model corresponding to the respective KPI.

Images