AI-based alarm prediction in process control system

Abstract

The embodiment of the invention relates to AI-based alarm prediction in a process control system. A method (200) for managing an alarm in a distributed control system, where the method (200) comprises the steps of monitoring (S1) at least one process parameter (1), activating (S3) an alarm if the process parameter (1) exceeds an alarm threshold (3), releasing (S4) the alarm if the process parameter (1) no longer exceeds the alarm threshold (3), and determining (S2), by the AI-based alarm control system before the alarm is activated, at least one piece of predicted alarm information comprising information on whether the alarm threshold (3) is to be exceeded.

Description

Technical Field

[0001] The present invention relates to systems and methods, computer program products, and computer readable media for managing alarms in a distributed control system. Background Art

[0002] In industrial process control systems, an "alarm system" is a standard feature that defines various alarms to alert operators to take corrective actions when predefined critical thresholds are exceeded. During plant engineering, alarms are defined for different sensor measurements (such as temperature, flow, pressure or level sensors) based on thresholds that must not be exceeded. During plant operation, if any of these thresholds are exceeded, an alarm will be issued. The engineering and operation of a plant can be done through a distributed control system, which allows the implementation and control of various parts and systems of the plant.

[0003] When an alarm is raised, the operator must take corrective action in the control system, such as opening or closing valves, to direct production back on track. In many alarm situations, operators have limited time to react, requiring them to remain highly aware and focused at all times. Alarms that are not promptly addressed by operators may escalate to more serious plant conditions, resulting in possible production loss, system damage, and downtime.

[0004] The complex interactions between different parts and systems of a plant often make it difficult for operators who control parts of the plant or more systems to decide on actions to properly handle and resolve alarms. In addition, when a user takes countermeasures to react to an alarm, the user action may have other undesirable effects on other systems that the operator could not have anticipated.

[0005] Therefore, there is a need for improved systems and methods for managing alarms in distributed control systems. Summary of the invention

[0006] This object is achieved by the invention according to the independent claims. Preferred embodiments of the invention are provided in the dependent claims, the description and the drawings.

[0007] The present invention relates to a method for managing alarms in a distributed control system.

[0008] The method comprises the step of monitoring at least one process parameter.

[0009] The method further comprises the step of activating an alarm if the process parameter exceeds an alarm threshold.

[0010] The method further comprises the step of clearing the alarm if the process parameter no longer exceeds the alarm threshold.

[0011] The method further comprises the step of determining, by the AI-based alarm control system before the alarm is activated, at least one predicted alarm information comprising information as to whether an alarm threshold will be exceeded.

[0012] Thus, an improved alarm management method is provided, wherein the adoption of AI-based alarm prediction can help operators predict alarm situations even before the alarm occurs.

[0013] A distributed control system may be understood as a control system for a process or plant that typically includes one or more control loops, in which automatic controllers are distributed throughout the system. Such a distributed control system may be supported by a computer. Even though the method of the present invention is described herein as a distributed control system, the advantages of the method and system according to the present invention may also be applied to any other suitable industrial process control system known to those skilled in the art.

[0014] The monitoring of at least one process parameter can be performed continuously, for example without interruption, or at a given one or more time instances, for example for a certain duration. In addition, the monitoring can be tied to other conditions. For example, the monitoring of the process parameter can only be started when a certain threshold value of the process parameter is exceeded.

[0015] Activation of an alarm may include a single activation once an alarm threshold is exceeded, continuous activation, and / or repeated activation at one or more time instances, such as activation within a certain time. The alarm may accordingly present an "ACT" state, indicating the "activation state" of the alarm. The state may accordingly be recorded in the alarm list and alarm log of the distributed control system.

[0016] Once a certain alarm threshold is exceeded, an alarm may be activated. This may include a situation where a process parameter rises over time and the alarm threshold defines an upper limit to which normal operating conditions for the process parameter are assumed to be able to reach.

[0017] A process parameter may be understood as any entity relevant to the operation of a system, such as a physical or electrical condition. For example, a process parameter may be a system pressure when a system of a plant is operating, and an alarm threshold may correspond to an upper limit of pressure at which normal operation of the system of the plant can be ensured.

[0018] Alternatively, if the process parameter decreases over time, and the threshold value defines a lower limit at which normal operating conditions of the process parameter are assumed to be reached, an alarm may be activated. For example, the process parameter may be the current at which a system of the plant operates, and the alarm threshold value may correspond to a lower limit at which normal operation of the system of the plant can be ensured.

[0019] An alarm threshold may be a predefined value for a process parameter above or below which an alarm may be sounded. In other words, an alarm may be activated if a process parameter is outside a predefined "normal operating condition" zone. It is also possible to provide more than one alarm threshold for one or more process parameters. Furthermore, multiple upper and / or lower alarm thresholds may be provided and connected to different states of the alarm. Thus, different alarms may be activated, such as a "low level alarm", "medium level alarm" and / or "high level alarm", depending on the criticality of the monitored process parameter value.

[0020] The disarming of an alarm may include a single disarming once the alarm no longer exceeds a threshold, a continuous disarming, and / or a repeated disarming at one or more time instances, such as a disarming within a certain duration. When an alarm is disarmed, the alarm may present an "RTN" status, indicating "return to normal" operation. This status may be recorded in the alarm list and alarm log of the distributed control system accordingly.

[0021] The determination of at least one predicted alarm information may be performed continuously or at specific one or more predetermined time instances. In addition, this determination may be tied to other conditions. For example, once a certain threshold of a process parameter is exceeded, this determination may begin. The alarm may present a "PRE" state, indicating a "predicted state", in which the predicted information is determined by the AI-based alarm control system. This state may be recorded in the alarm list and alarm log of the distributed control system accordingly.

[0022] The AI-based alarm control system may include any suitable AI-based software or algorithm and / or machine learning device, which is suitable for providing information for prediction or forecast of the further development of the monitored process parameter. The AI-based alarm control system may be trained by historical data and / or simulation data. In addition, the AI-based alarm control system may take into account any suitable parameters that may affect the future development of the monitored process parameter for the prediction. Similarly, the AI-based alarm control system may take into account in its forecast any interaction with one or more other systems of the distributed control system and corresponding further process parameters that may have an impact on the future development of the monitored process parameter.

[0023] The predicted alarm information may include a prediction or forecast about the timely development of the process parameter and may include a determination of the likelihood or probability if an alarm threshold is exceeded. In this regard, the predicted alarm information may also include time information indicating the time at which the process parameter is expected to exceed the alarm threshold. Furthermore, the predicted information may include a forecast about the level or value that the process parameter may have at a future time, preferably taking into account certain conditions.

[0024] In a preferred embodiment, the method further comprises the step of providing predicted alarm information to the user, preferably before the alarm is activated.

[0025] Therefore, users or operators who control and operate one or more processes of a plant via a distributed control system can have more time to prepare for upcoming situations. Thus, operators can be allowed to run their plants more proactively rather than reacting to different plant situations.

[0026] Predicted alarm information may be provided to the operator continuously, or only at certain times or under certain conditions, for example if a process parameter exceeds a certain predefined threshold. Such a predefined threshold may be different from the alarm threshold. For example, this may be a level at which the process parameter is still operating under normal conditions, but exceeding this level may indicate that the process parameter may increase or decrease towards the alarm threshold. Such a predefined threshold may be set, for example, in such a way that the operator may have enough time to react to the changing process parameter.

[0027] In another example, an operator may be warned when it is predicted that an operator's action will result in a critical alarm. For example, each time an operator changes a set point for a process value, the determination and provision of predicted alarm information may be triggered to predict whether the operator's action will result in a critical alarm. In some cases, due to complex process dynamics, it may not be easy for an operator to foresee such an alarm. However, an AI-based alarm control system may be able to make this prediction based on a deep understanding of the process dynamics (e.g., from historical data). If an operator's action is predicted to result in a critical alarm, the operator may be warned accordingly as to whether the action should be set in the system.

[0028] In a preferred embodiment, the method further comprises the step of determining, preferably by an AI based alarm control system, at least one control parameter of the distributed control system, the at least one control parameter being adapted to reduce the likelihood that an alarm threshold will be exceeded.

[0029] In another preferred embodiment, the method further comprises the step of providing at least one control parameter information to a user, wherein the control parameter information comprises information relating to the control parameter and / or wherein the control parameter information comprises information relating to at least one appropriate adaptation of the control parameter to reduce the likelihood that the alarm threshold will be exceeded.

[0030] Thus, by determining appropriate control parameters, the method according to the invention may enable the user to react appropriately to a forecasted alarm. This may include timely response to predicted developments in monitored process parameters in order to avoid future alarms.

[0031] The determination of the control parameter and / or the appropriate adaptation of said control parameter may be performed by an AI-based alarm control system or a different system. The control parameter may be a predefined control or setting which may prevent the determined process parameter from exceeding an alarm threshold. Thus, activation of an alarm may be prevented.

[0032] For example, where pressure can be monitored, and where the AI-based alarm control system can predict that a certain threshold is exceeded in the future, a selected valve can be determined to release the pressure. This information can be provided to the user accordingly, so that the user can therefore react appropriately, i.e., open the selected valve to release the pressure before the pressure exceeds a certain alarm threshold, thereby preventing the activation of the alarm.

[0033] The providing of at least one control parameter information may include, for example, providing a visual or audible indication of said control parameter information to the user.

[0034] The user may not only be informed of the respective control parameters, but also how to adapt the control parameters to reduce the likelihood that an alarm threshold will be exceeded. This may also include the provision of one or more alternative options of how to proceed further. Thus, the operator may be enabled to optimally adapt one or more control parameters to positively influence the predicted timely development of the process parameter. Considering the above pressure example, not only the selected valve may be provided, but also the selected optimal opening angle of the selected valve.

[0035] Additionally, the user may be provided with a sequence of multiple actions.

[0036] In a preferred embodiment, the method further comprises the step of providing at least one alarm information to the user when the alarm is activated, wherein the alarm information and the predicted alarm information are provided to the user together or separately.

[0037] Therefore, the method of the present invention can allow for effective control and supervision of a distributed control system with respect to both alarm information and predicted alarm information.

[0038] The alarm information may include information that a process parameter has exceeded a predefined threshold. Additionally, predicted alarm information may include information that a process parameter is expected to exceed a predefined threshold. Providing one or more alarm information and / or predicted alarm information to a user may include a visual indication, an audible indication, or any other suitable method or device for providing alarm information to a user.

[0039] For example, information about the monitored process parameter and the alarm threshold at which an alarm will be indicated may be provided to the user. Alternatively or additionally, predicted alarm information may also be provided to the user on the same or different screen or display, where the user may oversee or monitor the development of the process parameter and its prediction. Furthermore, both information may be provided simultaneously or at different times and / or on the same or different screens. The alarm information and / or predicted alarm information may be explained to the user by "Explainable AI (XAI)".

[0040] Thus, for example, by means of an on-screen warning, the user may be notified that the process parameter has exceeded the alarm threshold and / or will exceed the alarm threshold in the future. This may also include the provision of information about the probability or likelihood that the alarm threshold will be exceeded.

[0041] In one example, an alarm may be displayed, for example, as a red alarm icon next to a process display where the alarm occurred. This may include a label indicating the alarm, for example, "HH" for a very severe "high-high" alarm. Similarly, where a process value is too low or too high, a "low" or "high" alarm may appear. In addition, the alarm may also be shown in an alarm list that shows more information about the alarm, such as its title and description and the recorded time when the alarm occurred. In order for the operator to distinguish between an AI-based predicted alarm and a "real" alarm, the AI-based predicted alarm may be integrated into the operator screen and presented to the user, for example, by adding a leading "P" to "predicted" before the label to indicate that this is a predicted alarm. For example, "PH" would mean a "predicted high" alarm and "PLL" would mean a "predicted low-low" alarm. Optionally, additional information next to the alarm label may indicate, for example, the probability that the predicted alarm will actually occur and / or the time of occurrence. In addition, the alarm list may provide additional information about the predicted alarm, such as indicating in the alarm title text that this is a "predicted" alarm rather than a "real" alarm.

[0042] In a preferred embodiment, the method further comprises the step of automatically adapting control parameters to reduce the likelihood that an alarm threshold will be exceeded.

[0043] Thus, an alarm can be prevented before it occurs, without requiring operator intervention. Automatic adaptation can be understood as changing one or more control parameters substantially without user interaction. In this way, assistance to the operator can be ensured by convenient control of the distributed control system, especially in emergency situations.

[0044] In another embodiment, the adaptation of the control parameters may be performed at least semi-manually by an operator.

[0045] The invention also relates to a system for managing alarms in a distributed control system.

[0046] The system includes a process parameter monitoring device configured to monitor at least one process parameter.

[0047] The system also includes an alarm control configured to activate an alarm if the process parameter exceeds an alarm threshold and to deactivate the alarm if the process parameter no longer exceeds the alarm threshold.

[0048] The system also includes an AI-based alarm control system configured to determine at least one predicted alarm information before the alarm is activated, the at least one predicted alarm information including information about whether an alarm threshold will be exceeded.

[0049] The same explanations of the above-mentioned features for the method of the invention also apply to the system of the invention and in this regard reference is made to the above-mentioned parts which describe said features in detail.

[0050] According to the present invention, an improved alarm management system is provided, wherein the adoption of AI-based alarm prediction can help operators predict alarm situations even before they occur.

[0051] The system of the present invention may be a part of a distributed control system or may be a separate system.

[0052] The process parameter monitoring apparatus is not limited to a single device, but may also include multiple hardware-based and / or software-based devices, such as one or more sensors, detectors, etc., and corresponding (multiple) controllers configured to allow detection of the condition or state of the process parameter.

[0053] The alarm control means may include a plurality of hardware-based and / or software-based means adapted to activate an alarm when a process parameter exceeds an alarm threshold and to deactivate the alarm when the process parameter no longer exceeds the alarm threshold. For example, this may include any suitable alarm device(s) and their corresponding control(s), such as the provision and control of an alarm whistle and a visual alarm, etc.

[0054] One or more alarm thresholds for one or more process parameters may be predefined and stored in a corresponding storage device of the system.

[0055] Predicted alarm information may be visual or audible information related to forecasted information, for example, whether a process parameter will exceed an alarm threshold.

[0056] The one or more process parameters may be any suitable parameters, such as fill level, mass flow, pressure, current, temperature, etc., which may be monitored by corresponding detection devices. For each of these parameters, an appropriate normal operating range may be defined, which is respectively bounded by a corresponding upper and / or lower limit, indicating an operating range that should preferably be prevented. For example, the pressure in a tank may be defined as operating normally between an upper and lower limit. When the pressure exceeds the upper and / or lower limit, an alarm may be issued. Thus, when the AI-based control system determines that a threshold value is about to be exceeded, corresponding predicted alarm information may be provided.

[0057] Predicted alarm information may preferably be provided before a threshold is exceeded and an alarm is sounded. This may allow the user to take early countermeasures to avoid activating an alarm.

[0058] The predicted alarm information may include information about the likelihood or probability, such as a percentage, which indicates the likelihood that the threshold will be exceeded and the alarm will be activated. In addition, the predicted alarm information may include time information which indicates when the process parameter is predicted to exceed the threshold.

[0059] The system according to the invention may comprise a control unit configured to adapt the control parameters to reduce the likelihood that the alarm threshold will be exceeded. The control unit may comprise corresponding hardware and / or software means to allow a suitable adaptation of the control parameters.

[0060] The system according to the invention may be particularly valuable for rare severe alarms. In these severe but rare situations, the operator may wish to receive an early warning that can help the operator prevent such severe situations. Thus, the system according to the invention enables the prevention of severe situations at an early stage, thereby avoiding severe alarms and facilitating operation via a distributed control system.

[0061] By adopting an AI-based alarm control system, the system according to the present invention may also allow future situations to be properly handled, which have not yet occurred but can be anticipated through the prediction / forecast of the AI-based alarm control system (e.g., historical data that the AI-based alarm control system has been trained on).

[0062] In a preferred embodiment, the system further comprises an information device configured to provide predicted alarm information to the user, preferably before the alarm is activated.

[0063] Thus, the user or operator can be appropriately informed of the risk that a future alarm will be issued. The information device may include one or more monitors or displays that visually indicate the predicted alarm information. Alternatively or additionally, the information device may include one or more audio devices that audibly indicate the predicted alarm information.

[0064] Predicted alarm information may preferably be provided before an alarm is activated. Thus, a user controlling or supervising one or more systems of a distributed control system may be allowed to take early appropriate countermeasures to avoid future alarms.

[0065] In a preferred embodiment, predictive alarm information is provided when the likelihood that an alarm threshold will be exceeded exceeds a predefined likelihood threshold.

[0066] Thus, when the likelihood that an alarm threshold will be exceeded is very low, the system may avoid providing predicted alarm information. Thus, the operator may only be provided with relevant predictions about future alarms, thereby facilitating the operation of the system according to the present invention. In other words, alarms that are unlikely to occur (i.e., have a low likelihood of occurring) may not exceed the predefined likelihood threshold and therefore may not be provided to the user.

[0067] The likelihood or probability may be a percentage and may depend on other variables or conditions. In one example, when the likelihood that the monitored process parameter will exceed the alarm threshold is higher than 5%, the predicted alarm information may be provided to the user.

[0068] In a preferred embodiment, the system is also configured to determine, preferably by an AI-based alarm control system, at least one control parameter of the distributed control system, the at least one control parameter can be adapted to reduce the likelihood that an alarm threshold will be exceeded, and preferably provide at least one control parameter information to a user, wherein the control parameter information includes information related to the control parameter and / or wherein the control parameter information includes information related to at least one appropriate adaptation of the control parameter to reduce the likelihood that the alarm threshold will be exceeded.

[0069] Therefore, the system according to the present invention can allow the user to react appropriately to predicted alarms by determining appropriate control parameters.Thus, the system can enable timely responses to predicted developments in monitored process parameters, thereby potentially avoiding future alarms.

[0070] In a preferred embodiment, the system is further configured to provide at least one alarm information to the user when the alarm is activated, wherein the alarm information and the predicted alarm information are provided to the user together or separately.

[0071] Therefore, the system of the present invention can allow for effective control and supervision of distributed control systems.

[0072] In a preferred embodiment, the system includes a plurality of control subsystems, wherein each of the plurality of control subsystems is addable to and / or removable from a distributed control system, wherein the AI-based alarm control system is an AI-based alarm control subsystem.

[0073] Therefore, a modular and flexible system can be provided, in which certain control subsystems can be added or deleted according to the user's requirements. The distributed control system can organize the entire system functionality in a modular manner. Therefore, each control subsystem can be understood as a "feature" or "service" that can be selectively included in the system. Depending on the services that the user may require, each control subsystem can be individually licensed, installed / uninstalled, and run / stopped. For example, the corresponding subsystem can be selected as an option for installation in a system installation tool. Multiple control subsystems may include basic or mandatory services that are always included and run, as well as additional optional services. An AI-based alarm control subsystem can be provided as an optional service to allow the user to license and use the feature to enhance the control of the distributed control system.

[0074] The AI-based alarm control subsystem may be provided in the installation tool in the form of an "AI Predictive Alarms" installer feature as part of a larger installer feature that may also include various other additional AI-based features, such as anomaly detection. When the "AI Predictive Alarms" feature is selected in the installation tool, the corresponding service may be listed as part of the service status viewer, indicating the selected / licensed and installed services. This may provide the user with improved monitoring and management of the selected services.

[0075] In a preferred embodiment, the system also includes a non-AI-based alarm control subsystem, which is configured to add or remove one or more first alarm rules for managing alarms of a distributed control system, wherein the system is also configured to allow adding or removing one or more second alarm rules, the one or more second alarm rules being used to manage alarms of a distributed control system defined by the AI-based alarm control subsystem.

[0076] A non-AI-based alarm control subsystem may be understood as a control system that manages alarms without the assistance of AI. In particular, a non-AI-based alarm control subsystem may not employ forecast information of process parameters based on AI prediction.

[0077] Alarm rules may be understood as control rules for controlling alarms, such as conditions for activating and / or deactivating alarms, such as alarm thresholds for more process parameters. Thus, one or more first alarm rules may be created and controlled by a non-AI-based alarm control subsystem to allow general control of alarms in a distributed control system.

[0078] One or more second alarm rules can be correspondingly created and controlled by the AI-based alarm control subsystem, in particular to supplement the first alarm rule. Thus, an improved complex alarm control system can be obtained.

[0079] In a preferred embodiment, the system is further configured to allow modification of the first alarm rules by the AI-based alarm control subsystem to create one or more modified first alarm rules for managing alarms of the distributed control system.

[0080] Therefore, the AI-based alarm control subsystem can not only provide a second alarm rule that is complementary to the first alarm rule, but can also be used to modify the first alarm rule, thereby allowing for further improvement of the alarm management of the distributed control system. In other words, AI-based alarms can be created not only as new alarms, but also based on existing alarms.

[0081] Thus, the user may be allowed to configure alert rules for existing alerts accordingly. For example, this may include filters, hiding rules, alert grouping rules, etc. In one embodiment of the system of the present invention, the user may be allowed to include existing alert rules in the AI-based alert control subsystem. When alert rules have been selected to be included, the AI-based alert prediction service may train and / or update the corresponding AI-based prediction model that runs to predict alerts. This may include using a trainable machine learning model, such as an LSTM-based time series forecasting model for multivariate time series data.

[0082] The user can correspondingly choose whether to create additional alarm rules and / or whether to modify existing alarm rules based on the AI-based alarm control subsystem. Therefore, the system according to the present invention provides a sophisticated method for managing alarm rules in a distributed control system.

[0083] The present invention also relates to a computer program product comprising instructions which, when executed by a computer, cause the computer to perform and / or control any of the methods disclosed herein.

[0084] The features of the system according to the invention may be implemented by corresponding suitable digital or computing means, which may comprise, for example, one or more computers, applications and / or a network.

[0085] The method may be at least partially implemented by a computer and may be implemented in software or hardware or in software and hardware. In addition, the method may be performed by computer program instructions running on a device providing data processing functionality.

[0086] The data processing device may be a suitable computing device, such as an electronic control module, etc., which may also be a distributed computer system. The data processing device or computer may include one or more of a processor, a memory, a data interface, etc., respectively.

[0087] The present invention also relates to a computer-readable medium comprising instructions which, when executed by a computer, cause the computer to perform and / or control any of the methods disclosed herein.

[0088] The features and advantages outlined above in the context of the system and method also apply to the computer program product and computer readable medium described herein. Likewise, any features and advantages described with respect to the method of the present invention are correspondingly applicable to the system of the present invention, and vice versa.

[0089] The computer program may be stored / distributed on an appropriate medium, such as an optical storage medium or a solid-state medium, provided together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0090] Other features, examples and advantages will become apparent from the following detailed description of the preferred embodiments and the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] In order to better understand the present invention and illustrate its practicality, the following drawings are provided and referenced. It should be understood that the drawings represent only exemplary embodiments and therefore do not limit the scope of the claimed invention in any way. Elements with the same or similar functions are always represented by the same reference symbols. Any reference mark in the claims should not be interpreted as limiting the scope of the claims.

[0092] In the accompanying drawings,

[0093] Figure 1 Schematically illustrates a method for managing alarms in a distributed control system according to the present invention;

[0094] Figure 2 is a flow chart illustrating a method for managing alarms in a distributed control system according to the present invention;

[0095] Figure 3 Schematically illustrates a system for managing alarms in a distributed control system according to the present invention;

[0096] Figure 4 A system for managing alarms in a distributed control system according to the present invention is schematically illustrated. DETAILED DESCRIPTION

[0097] Figure 1 An embodiment of a method 200 for managing alarms in a distributed control system according to the invention is described. Here, a process parameter 1 is monitored. The time "t" is depicted on the horizontal axis and the amplitude "A" of the process parameter 1 is depicted on the vertical axis.

[0098] As shown, process parameter 1 starts at a level below alarm threshold 3 and begins to rise over time.

[0099] According to the present invention, in step S2, it is predicted by the AI-based alarm control system 9 whether the alarm threshold 3 will be exceeded. In the figure, the state of the alarm is represented as "PRE", indicating that the corresponding predicted alarm information is determined by the AI-based alarm control system 9 before the alarm is activated. Therefore, based on the predicted alarm information, it can be indicated to the user that the alarm threshold 3 will be exceeded. Therefore, the corresponding control can be performed or adapted with the goal of avoiding the process parameter 1 from exceeding the alarm threshold 3.

[0100] When process parameter 1 exceeds alarm threshold 3, the alarm is activated, as shown in step S3. In the figure, the alarm state is marked as "ACT", indicating that the alarm is in the "activated" state. The alarm can remain activated as long as process parameter 1 exceeds alarm threshold 3. After the control parameters of process parameter 1 are adapted, process parameter 1 starts to decrease.

[0101] When the process parameter 1 no longer exceeds the alarm threshold 3, the alarm is cleared, as shown in step S4. In the figure, the alarm state is marked as "RTN", indicating that the system returns to normal conditions.

[0102] As will be appreciated by a person skilled in the art, the described embodiments do not define the claimed method, but any timely progression of the process parameter 1 may also be envisaged. It will therefore be appreciated that the process parameter 1 may also decrease over time, and that the alarm threshold 3 is defined as a lower threshold below which an alarm is sounded. In this case, the alarm will be activated as soon as the process parameter 1 falls below the alarm threshold 3. The method according to the invention therefore correspondingly allows the determination of corresponding predicted alarm information by the AI-based alarm control system 9 in a "pre" state, preferably before the process parameter 1 falls below the alarm threshold 3.

[0103] Figure 2 A flow chart illustrating a method according to the invention for managing alarms in a distributed control system is described. The flow chart described is based on monitoring of at least one process parameter 1 (see Figure 1 Step S1).

[0104] In step S2, at least one predicted alarm information is determined by the AI-based alarm control system 9 based on the monitoring of at least one process parameter 1. The determination is preferably performed before the alarm is activated. The predicted alarm information includes information about whether and / or when an alarm threshold will be exceeded.

[0105] In step S3, the process parameter 1 exceeds the alarm threshold 3 and the alarm is activated. As long as the process parameter 1 exceeds the alarm threshold 3, the alarm may remain activated.

[0106] At step S4, the process parameter 1 no longer exceeds the alarm threshold 3 and the alarm is cleared. The system 100 then returns to normal operation.

[0107] In step S5, the predicted alarm information is provided to the user. This can be performed before the alarm is activated, ie before step S3, as indicated by the dashed arrow.

[0108] At step S6, at least one control parameter of the distributed control system is determined, the at least one control parameter being adapted to reduce the likelihood that the alarm threshold 3 will be exceeded. This may be done by the AI ​​based alarm control system 9.

[0109] In step S7, at least one control parameter information is provided to the user. The control parameter information includes information about the relevant control parameter. The control parameter information may also include information about at least one appropriate adaptation of the control parameter to reduce the likelihood that the alarm threshold will be exceeded. This may help the user take correct countermeasures to avoid future alarms.

[0110] In step S8, when an alarm is activated, alarm information is provided to the user. This may include, for example, information about the alarm type and the relevant process parameters 1. Here, the alarm information and the predicted alarm information may be provided to the user together or separately.

[0111] In step S9, the control parameters are automatically adapted to reduce the likelihood that the alarm threshold will be exceeded. Thus, future alarms can be avoided before activation.

[0112] It should be understood that the method according to the present invention is not limited to the order of the above method steps. On the contrary, the method steps can also be provided in a different order, and one or more of the above method steps can be removed or further method steps can be added as needed.

[0113] Furthermore, a computer program product 300 and a computer readable medium 500 are schematically depicted, each comprising instructions which, when executed by the computer 400, cause the computer 400 to perform the method of the present disclosure, in particular Figure 1 and Figure 2 Method 200 is shown.

[0114] Figure 3 A system 100 for managing alarms in a distributed control system according to the present invention is schematically depicted.

[0115] The system 100 comprises a process parameter monitoring device 5 configured to monitor at least one process parameter 1. The process parameter monitoring device 5 may comprise a corresponding detector or other measuring device suitable for measuring the amplitude, level etc. of the process parameter 1 to be controlled.

[0116] The system 100 further comprises an alarm control device 7, which is configured to activate an alarm when the process parameter 1 exceeds the alarm threshold value 3. Therefore, the alarm control device 7 may comprise one or more suitable devices for issuing an alarm. Furthermore, if the process parameter 1 no longer exceeds the alarm threshold value 3, the alarm may be deactivated by the alarm control device 7. In this case, the system may return to a normal operating state.

[0117] The system 100 also includes an AI-based alarm control system 9, which is configured to determine predicted alarm information. This determination is preferably performed before the alarm is activated. Thus, forecast information about potential future alarms can be provided. The predicted alarm information correspondingly includes information about whether and / or when the alarm threshold 3 will be exceeded.

[0118] The system 100 also includes an information device 11, which is configured to provide the predicted alarm information to the user. This may include one or more monitors or displays that visually indicate the predicted alarm information. Alternatively or additionally, the information device 11 may include one or more audio devices that audibly indicate the predicted alarm information.

[0119] Figure 4 A system 100 for managing alarms in a distributed control system according to the present invention is schematically described. The system 100 includes a plurality of control subsystems 100a, 100b, 100c, 100d, which can be integrated in a distributed control system. Each of the plurality of control subsystems 100a, 100b, 100c, and 100d can be selectively added to and / or removed from a distributed control system. Other control subsystems can be provided, as indicated by the "..." in the lowest box of the figure.

[0120] The AI-based alarm control system 9 is configured as a control subsystem, namely, an AI-based alarm control subsystem 100a. The system 100 also includes a non-AI-based alarm control subsystem 100b, which is configured to add or remove one or more first alarm rules for managing alarms of a distributed control system. The system 100 is also configured to allow one or more second alarm rules defined by the AI-based alarm control subsystem 100a to be added or removed for managing alarms of a distributed control system. The system 100 can also be configured to allow the first alarm rule to be modified by the AI-based alarm control subsystem 100a to create one or more modified first alarm rules for managing alarms of a distributed control system.

[0121] When the system 100 is implemented as software or a computer program (product), the control subsystems 100a, 100b, 100c, 100d may be implemented as respective software features that may be added or removed by a user as needed to properly control the distributed control system.

[0122] List of Reference Numerals

[0123] 1 Process parameters

[0124] 3 Alarm Thresholds

[0125] 5. Process parameter monitoring device

[0126] 7 Alarm control device

[0127] 9 AI-based alarm control system

[0128] 11 Information Devices

[0129] 100 Systems for managing alarms

[0130] 100a, b, c, d Control subsystem

[0131] 100a AI-based alarm control subsystem

[0132] 100b Non-AI based alarm control subsystem

[0133] 200 Methods for managing alerts

[0134] 300 Computer program products

[0135] 400 Computer

[0136] 500 Computer readable medium

[0137] S1 Process Parameter Monitoring

[0138] S2 Predicted Alarm OK

[0139] S3 alarm activated

[0140] S4 Alarm cleared

[0141] S5 Forecast Alert Information Provision

[0142] S6 Control Parameter Determination

[0143] S7 control parameter information provision

[0144] S8 Alarm Information Provided

[0145] S9 control parameter adaptation

Claims

1. A method (200) for managing alarms in a distributed control system, wherein the method (200) comprises the following steps: monitoring (S1) at least one process parameter (1), If the process parameter (1) exceeds an alarm threshold (3), an alarm is activated (S3), If the process parameter (1) no longer exceeds the alarm threshold (3), the alarm is deactivated (S4), and Before the alarm is activated, at least one predicted alarm information is determined (S2) by the AI-based alarm control system, the at least one predicted alarm information comprising information as to whether the alarm threshold (3) will be exceeded.

2. The method according to the preceding claim, wherein the method further comprises the following steps: The predicted alarm information is provided (S5) to a user, preferably before the alarm is activated.

3. The method according to any one of the preceding claims, wherein the method further comprises the following steps: determining (S6), preferably by the AI-based alarm control system (9), at least one control parameter of the distributed control system, the at least one control parameter being adaptable to reduce the likelihood that the alarm threshold will be exceeded, and Preferably, at least one control parameter information is provided (S7) to the user, wherein the control parameter information comprises information relating to the control parameter, and / or wherein the control parameter information comprises information relating to at least one appropriate adaptation of the control parameter to reduce the likelihood that the alarm threshold will be exceeded.

4. The method according to any one of the preceding claims 2 or 3, wherein the method further comprises the following steps: When the alarm is activated, at least one alarm information is provided (S8) to the user, wherein the alarm information and the predicted alarm information are provided to the user together or separately.

5. The method according to any one of the preceding claims 3 or 4, wherein the method further comprises the following steps: The control parameters are automatically adapted (S9) to reduce the likelihood that the alarm threshold will be exceeded.

6. A system (100) for managing alarms in a distributed control system, wherein the system comprises: A process parameter monitoring device (5) is configured to monitor at least one process parameter (1), an alarm control device (7) configured to activate an alarm if the process parameter (1) exceeds an alarm threshold (3), and to deactivate the alarm if the process parameter (1) no longer exceeds the alarm threshold (3), An AI-based alarm control system (9) is configured to determine at least one predicted alarm information before the alarm is activated, the at least one predicted alarm information comprising information whether the alarm threshold (3) will be exceeded.

7. System according to the preceding claim, wherein the system further comprises an information device (11) configured to provide the predicted alarm information to a user, preferably before the alarm is activated.

8. System according to the preceding claim, wherein the predicted alarm information is provided when the likelihood that the alarm threshold (3) will be exceeded exceeds a predefined likelihood threshold.

9. The system according to any one of the preceding claims 6 to 8, wherein the system is further configured to: Preferably, the AI-based alarm control system (9) determines (S6) at least one control parameter of the distributed control system, the at least one control parameter being adaptable to reduce the likelihood that the alarm threshold will be exceeded, and preferably At least one control parameter information is provided (S7) to the user, wherein the control parameter information comprises information relating to the control parameter and / or wherein the control parameter information comprises information relating to at least one appropriate adaptation of the control parameter to reduce the likelihood that the alarm threshold (3) will be exceeded.

10. The system according to any one of the preceding claims 7 to 9, wherein the system is further configured to: When the alarm is activated, at least one alarm information is provided (S8) to the user, wherein the alarm information and the predicted alarm information are provided to the user together or separately.

11. The system according to any one of the preceding claims 6 to 10, wherein the system (100) comprises a plurality of control subsystems (100a, 100b, 100c, 100d), wherein each of the plurality of control subsystems (100a, 100b, 100c, 100d) is addable to and / or removable from the distributed control system, The AI-based alarm control system is an AI-based alarm control subsystem (100a).

12. The system according to the preceding claim, wherein the system further comprises: a non-AI-based alarm control subsystem (100b) configured to add or remove one or more first alarm rules for managing alarms of the distributed control system, The system (100) is further configured to allow the addition or removal of one or more second alarm rules for managing alarms of the distributed control system, wherein the one or more second alarm rules are defined by the AI-based alarm control subsystem (100a).

13. The system according to the preceding claim, wherein the system (100) is further configured to allow the first alarm rule to be modified by the AI-based alarm control subsystem (100a) to create one or more modified first alarm rules for managing alarms of the distributed control system.

14. A computer program product (300) comprising instructions which, when executed by a computer (400), cause the computer (400) to perform and / or control the method (200) according to any one of claims 1 to 5.

15. A computer readable medium (500) comprising instructions which, when executed by a computer (400), cause the computer (400) to perform and / or control the method (200) according to any one of claims 1 to 5.

Images