Systems and methods for mitigating false alarms in building management systems

Abstract

A false alarm artificial intelligence (AI) model is trained using metadata associated with alarms classified as false alarms, and a true alarm artificial intelligence (AI) model is trained using metadata associated with alarms classified as true alarms. An incoming alert is received. Both a false alarm AI model and a true alarm AI model are applied to incoming alarms, and both models classify the incoming alarms as a false alarm classification or a true alarm classification. When the false alarm AI model and the true alarm AI model coincide, incoming alarms are automatically classified accordingly. When the models are inconsistent, the incoming alert is presented to an operator console of the BMS and a manual classification of the incoming alert is received from the operator console.

Description

Technical Field

[0001] This disclosure relates generally to building management systems, and more specifically to mitigating false alarms in building management systems. Background Technology

[0002] A building management system is a system for controlling and / or monitoring a building or other facility. A building management system may include, for example, HVAC systems, security systems, video management systems, access control systems, fire protection systems, and / or any other suitable building control system. In many cases, a building management system issues an alarm when an anomaly is detected in the building and / or during the operation of the building management system. Alarms typically must be acknowledged and / or otherwise resolved by the building's operators or other personnel. In some cases, a building management system may issue an alarm indicating a potential problem or difficulty is occurring, even if no such problem or difficulty is actually occurring in the building. These alarms may be referred to as false alarms. When a false alarm occurs, the operator typically needs to respond, which can waste a significant amount of the operator's time and may divert the operator's attention from actual, genuine alarms. What is desired is a method and system for automatically determining whether an alarm is a false alarm or a genuine alarm. Summary of the Invention

[0003] This disclosure relates generally to building management systems (BMS), and more specifically to mitigating false alarms in building management systems (BMS). An example may exist in a method for mitigating false alarms in a building management system (BMS). This exemplary method includes: storing a false alarm artificial intelligence (AI) model trained using metadata associated with alarms classified as false alarms, and storing a true alarm artificial intelligence (AI) model trained using metadata associated with alarms classified as true alarms. An incoming alarm is received. The false alarm AI model is applied to the incoming alarm, and the false alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. The true alarm AI model is also applied to the incoming alarm, and the true alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a false alarm, the incoming alarm is automatically classified as a false alarm. When the false alarm AI model classifies an incoming alarm as a false alarm and the true alarm AI model classifies an incoming alarm as a true alarm, the incoming alarm is presented to the BMS operator console, and the operator console receives a manual classification of the incoming alarm as either a false alarm or a true alarm. In some cases, when both the false alarm AI model and the true alarm AI model classify an incoming alarm as false alarms, the incoming alarm is presented to the BMS operator console, and the operator console receives a manual classification of the incoming alarm as either a false alarm or a true alarm.

[0004] Another example could exist in an alarm management system. This alarm management system includes: an input terminal; a memory for storing false alarm artificial intelligence (AI) models and true alarm artificial intelligence (AI) models; an operator console including a user interface; and a controller operatively coupled to the input terminal, the memory, and the operator console. The controller is configured to receive incoming alarms from the input terminal. The controller is configured to apply the false alarm AI model to the incoming alarms, which classifies the incoming alarms as either false alarms or true alarms. The controller is also configured to apply the true alarm AI model to the incoming alarms, which classifies the incoming alarms as either false alarms or true alarms. When both the false alarm AI model and the true alarm AI model classify the incoming alarms as false alarms, the controller is configured to automatically classify the incoming alarms as false alarms. When both the false alarm AI model and the true alarm AI model classify the incoming alarms as true alarms, the controller is configured to automatically classify the incoming alarms as true alarms. When both the false alarm AI model and the true alarm AI model classify an incoming alarm as a false alarm, the controller is configured to present the incoming alarm on the operator console, requesting a manual classification of the incoming alarm as either a false alarm or a true alarm. In some cases, when both the false alarm AI model and the true alarm AI model classify an incoming alarm as a false alarm, the controller is also configured to present the incoming alarm on the operator console, requesting a manual classification of the incoming alarm as either a false alarm or a true alarm.

[0005] Another example may exist on a non-transitory computer-readable medium on which instructions are stored. This non-transitory computer-readable medium may include any suitable memory, such as DRAM, RAM, SRAM, flash memory, solid-state memory, hard disk, optical disk, digital video disk, cloud storage device, and / or any other suitable non-transitory computer-readable medium. When the instructions are executed by one or more processors, the processors are caused to receive an incoming alarm. The processors are then caused to apply a false alarm AI model to the incoming alarm, where the false alarm AI model is based on past alarms classified as false alarms. The false alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. The processors are also caused to apply a true alarm AI model to the incoming alarm, where the true alarm AI model is based on past alarms classified as true alarms. The true alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a false alarm, the processors are caused to automatically classify the incoming alarm as a false alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a true alarm, the incoming alarm is automatically classified as a true alarm. When the false alarm AI model classifies an incoming alarm as a false alarm and the true alarm AI model classifies an incoming alarm as a true alarm, the one or more processors present the incoming alarm to the operator console and receive a manual classification of the incoming alarm as either a false alarm or a true alarm from the operator console. In some cases, when both the false alarm AI model and the true alarm AI model classify an incoming alarm as false alarms, the one or more processors present the incoming alarm to the operator console and receive a manual classification of the incoming alarm as either a false alarm or a true alarm from the operator console.

[0006] The foregoing description is provided to facilitate understanding of the innovative features unique to this disclosure and is not intended as a complete description. A full understanding of this disclosure can be obtained by considering the entire specification, claims, drawings, and abstract as a whole. Attached Figure Description

[0007] This disclosure will be more fully understood by taking into account the following description of various examples in conjunction with the accompanying drawings, in which:

[0008] Figure 1 This is a schematic block diagram illustrating an exemplary alarm management system;

[0009] Figure 2A , Figure 2B and Figure 2C It is shown together for use Figure 1 A flowchart illustrating an exemplary alarm management system for mitigating false alarms;

[0010] Figure 3A and Figure 3B It is a flowchart illustrating a series of exemplary steps that can be executed by one or more processors that execute instructions stored on a non-transitory computer-readable medium;

[0011] Figure 4 It is a flowchart illustrating an exemplary workflow; and

[0012] Figure 5 This is a flowchart illustrating an exemplary workflow.

[0013] While this disclosure is subject to various modifications and alternatives, its details have been shown by way of example in the accompanying drawings and will be described in detail. However, it should be understood that this disclosure is not intended to limit it to the specific examples described. Rather, it is intended to cover all modifications, equivalents, and alternatives that fall within the substance and scope of this disclosure. Detailed Implementation

[0014] The following description should be read with reference to the accompanying drawings, in which the same elements in the different drawings are numbered in the same manner. The drawings are not necessarily drawn to scale and depict examples that are not intended to limit the scope of this disclosure. Although examples of various elements are illustrated, those skilled in the art will recognize that many of the examples provided have suitable alternatives that can be utilized.

[0015] This document assumes that all numbers are modified by the term “about” unless otherwise explicitly stated. Expressions of numerical ranges using endpoints include all numbers contained within that range (e.g., 1 to 5 includes 1, 1.5, 2, 2.75, 3, 3.80, 4, and 5).

[0016] As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural references unless otherwise expressly stated. As used in this specification and the appended claims, the term “or” is generally used in its meaning to include “and / or” unless otherwise expressly stated.

[0017] It should be noted that references to "one embodiment," "some embodiments," or "other embodiments" in the specification indicate that the described embodiments may include specific features, structures, or characteristics, but each embodiment need not necessarily include that specific feature, structure, or characteristic. Furthermore, these phrases do not necessarily refer to the same embodiment. Additionally, when a specific feature, structure, or characteristic is described in connection with an embodiment, it is conceivable that, whether explicitly described or not, that feature, structure, or characteristic may be applied to other embodiments, unless otherwise expressly stated otherwise.

[0018] Figure 1This is a schematic block diagram illustrating an exemplary alarm management system 10. The alarm management system 10 can be applied to a building management system (BMS), which may include, for example, HVAC systems, security and / or access control systems, fire protection systems, and / or any other suitable building control systems. In some cases, the alarm management system 10 can be applied as needed to industrial control systems, vehicle control systems (e.g., aircraft, ships), power plant control systems (coal-fired, nuclear, wind power), and / or any other suitable control systems.

[0019] The exemplary alarm management system 10 includes a memory 12 that stores a false alarm artificial intelligence (AI) model 14 and a true alarm artificial intelligence (AI) model 16. In some cases, the memory 12 may also store a false alarm database 18 and a true alarm database 20. The false alarm database 18 may store a history of alarms that have been classified as false alarms, and the true alarm database 18 may store a history of alarms that have been classified as true alarms. The exemplary alarm management system 10 includes an operator console 22 that includes a user interface 24. The user interface 24 may include, for example, a display and a keyboard, or it may be a display and a touchpad. In some cases, the user interface 24 may include or may be a touchscreen display that serves as both a display and a data input mechanism. The exemplary alarm management system 10 includes an input terminal 26. In some cases, the input terminal 26 is configured to receive alarms from, for example, a security panel 28 of a security system or other controller. A security system is used here as an example. The security panel 28 receives signals from a plurality of sensors 30, labeled 30a, 30, and 30c, respectively. There may be dozens, hundreds, or even thousands of sensors 30. Sensor 30 may include various different sensors, such as window opening sensors, door opening sensors, glass breakage detectors, motion detectors, fire sensors, smoke sensors, gas sensors, etc. Sensor 30 may include a camera and associated video analytics algorithms. For example, sensor 30 may be battery powered. In some cases, sensor 30 may communicate directly with input 26, in which case security panel 28 may not be present.

[0020] The alarm management system 10 may include a controller 32 operatively coupled to an input terminal 26, a memory 12, and an operator console 22. The controller 32 is configured to receive incoming alarms from the input terminal 26. For example, an incoming alarm may be emitted by one of a sensor 30 and / or a security panel 28. The controller 32 is configured to apply a false alarm AI model 14 to the incoming alarms. The false alarm AI model 14 is trained based on past alarms that have been classified as false alarms. The false alarm AI model 14 classifies the incoming alarm as either a false alarm or a true alarm. The controller 32 is also configured to apply a true alarm AI model 16 to the incoming alarms. The true alarm AI model 16 is trained based on past alarms that have been classified as true alarms. The true alarm AI model 16 classifies the incoming alarm as either a false alarm or a true alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a false alarm, the incoming alarm is automatically classified as a false alarm. When both the false alarm AI model 14 and the true alarm AI model 16 classify an incoming alarm as a true alarm, the incoming alarm is automatically classified as a true alarm. When both the false alarm AI model 14 and the true alarm AI model 16 classify an incoming alarm as a false alarm, the controller 32 is configured to present the incoming alarm on the operator console 22, thereby requesting a manual classification of the incoming alarm as either a false alarm or a true alarm. In some cases, when both the false alarm AI model 14 and the true alarm AI model 16 classify an incoming alarm as a false alarm, the controller 32 is configured to present the incoming alarm on the operator console 22, thereby requesting a manual classification of the incoming alarm as either a false alarm or a true alarm.

[0021] In some cases, when an incoming alarm is automatically classified as a false alarm, controller 32 may be configured not to display the incoming alarm on operator console 22. When an incoming alarm is classified as a false alarm, controller 32 may be configured to add the incoming alarm to false alarm database 18 and use the updated false alarm database 18 to train and / or retrain the false alarm AI model 14. When an incoming alarm is classified as a true alarm, controller 32 may be configured to add the incoming alarm to true alarm database 20 and use the updated true alarm database 20 to train and / or retrain the true alarm AI model 16.

[0022] In some cases, controller 32 may be configured to receive alarm logs via input 26, which include a log of alarms including metadata associated with each alarm in the log. The alarm log includes alarms classified as both false alarms and true alarms. Controller 32 may generate and / or update a false alarm database 18 based on alarms classified as false alarms in the alarm log, and may generate and / or update a true alarm database 20 based on alarms classified as true alarms in the alarm log. In some cases, when the false alarm database 18 is updated over time, controller 32 may use the false alarm database 18 to retrain the false alarm AI model 14. In some cases, when the true alarm database 20 is updated over time, controller 32 may use the true alarm database 20 to retrain the true alarm AI model 16.

[0023] Figure 2A , Figure 2B and Figure 2C This is a flowchart illustrating an exemplary method 34 for mitigating false alarms in a building management system (BMS). A BMS is used as an example. Method 34 includes storing a false alarm artificial intelligence (AI) model (such as false alarm AI model 14) trained using metadata associated with alarms previously classified as false alarms, as indicated in box 36. It also includes storing a true alarm artificial intelligence (AI) model (such as true alarm AI model 16) trained using metadata associated with alarms previously classified as true alarms, as indicated in box 38. In some cases, the false alarm AI model may include a linear regression-based model or a neural network-based model. In some cases, the true alarm AI model may include a linear regression-based model or a neural network-based model.

[0024] In the case of linear regression-based models, the model creates a simple regression line (sometimes with a threshold) between the input and output, and the coefficients represent the weight of each input. A linear regression-based model can classify an incoming alarm as a false alarm if it falls on one side of the regression line, and as a true alarm if it falls on the opposite side. Therefore, using a linear regression-based model provides a direct relationship between the input and output, which allows for a clear explanation of how each input affects the output. This provides transparency regarding the reasons for classifying an incoming alarm as a true or false alarm. This level of transparency regarding the model's decision-making process is typically not present in neural network-based models.

[0025] Receive incoming alerts, as indicated in box 40. Apply the false alarm AI model to the incoming alerts, where the false alarm AI model classifies the incoming alerts as either false alarms or true alarms, as indicated in box 42. Apply the true alarm AI model to the incoming alerts, where the true alarm AI model classifies the incoming alerts as either false alarms or true alarms, as indicated in box 44. When both the false alarm AI model and the true alarm AI model classify the incoming alerts as false alarms, the incoming alerts are automatically classified as false alarms, as indicated in box 46. When both the false alarm AI model and the true alarm AI model classify the incoming alerts as true alarms, the incoming alerts are automatically classified as true alarms, as indicated in box 48.

[0026] continue Figure 2B When the false alarm AI model classifies an incoming alarm as a false alarm and the true alarm AI model classifies an incoming alarm as a true alarm, the incoming alarm is presented to the BMS operator console, and the operator console receives a manual classification of the incoming alarm as either a false alarm or a true alarm, as indicated in box 50. In some cases, when an incoming alarm is classified as a true alarm, it may be necessary to clear the incoming alarm via operator action on the operator console, as indicated in box 52. In some cases, when an incoming alarm is classified as a false alarm, it may not be necessary to clear the incoming alarm via operator action on the operator console, as indicated in box 54. In some cases, when an incoming alarm is classified as a false alarm, the incoming alarm is not presented on the BMS operator console, as indicated in box 56. In some cases, when an incoming alarm is classified as a false alarm, the incoming alarm may be added to the false alarm database, as indicated in box 58. In some cases, the false alarm database may be used to train and / or retrain the false alarm AI model, as indicated in box 60. In some cases, when an incoming alarm is classified as a true alarm, the incoming alarm can be added to the true alarm database, as indicated in box 62.

[0027] continue Figure 2C In some cases, a true alarm database can be used to train and / or retrain the true alarm AI model, as indicated in box 64. When the false alarm AI model classifies an incoming alarm as a true alarm and the true alarm AI model classifies an incoming alarm as a false alarm, method 34 may include presenting the incoming alarm to the operator console of the BMS and receiving a manual classification of the incoming alarm as either a false alarm or a true alarm from the operator console, as indicated in box 66.

[0028] In some cases, method 34 may include receiving an alarm log, which includes a log of alarms containing metadata associated with each alarm in the log. The alarm log may include alarms classified as both false alarms and true alarms, as indicated in box 68. A false alarm database may be generated and / or updated based on alarms classified as false alarms in the alarm log, rather than based on alarms classified as true alarms in the alarm log, as indicated in box 70. A true alarm database may be generated and / or updated based on alarms classified as true alarms in the alarm log, rather than based on alarms classified as false alarms in the alarm log, as indicated in box 72. The false alarm database may be used to train and / or retrain a false alarm AI model, as indicated in box 74. The true alarm database may be used to train and / or retrain a true alarm AI model, as indicated in box 76.

[0029] Figure 3A and Figure 3B This is a flowchart illustrating a series of exemplary steps 78 that can be executed by one or more processors executing instructions stored on a non-transitory computer-readable medium. For example, one or more processors may be part of controller 32. One or more processors may be made to receive incoming alarms, as indicated at box 80. One or more processors may be made to apply a false alarm AI model to incoming alarms, wherein the false alarm AI model is based on past alarms classified as false alarms. The false alarm AI model classifies incoming alarms as either false alarms or true alarms, as indicated at box 82. One or more processors may be made to apply a true alarm AI model to incoming alarms, wherein the true alarm AI model is based on past alarms classified as true alarms. The true alarm AI model classifies incoming alarms as either false alarms or true alarms, as indicated at box 84. When both the false alarm AI model and the true alarm AI model classify incoming alarms as false alarms, one or more processors may be made to automatically classify incoming alarms as false alarms, as indicated at box 86. When both the false alarm AI model and the true alarm AI model classify incoming alarms as true alarms, one or more processors can be enabled to automatically classify incoming alarms as true alarms, as indicated in box 88. When both the false alarm AI model and the true alarm AI model classify incoming alarms as false alarms, one or more processors can be enabled to present the incoming alarm to the operator console and receive manual classification of the incoming alarm as either a false alarm or a true alarm from the operator console, as indicated in box 90.

[0030] continue Figure 3BWhen the false alarm AI model classifies an incoming alarm as a true alarm and the true alarm AI model classifies an incoming alarm as a false alarm, the incoming alarm is presented to the operator console, and the operator console receives a manual classification of the incoming alarm as either a false alarm or a true alarm, as indicated in box 92. In some cases, when an incoming alarm is classified as a false alarm, one or more processors may add the incoming alarm to the false alarm database, as indicated in box 94. In some cases, when an incoming alarm is classified as a true alarm, one or more processors may add the incoming alarm to the true alarm database, as indicated in box 96. In some cases, when the false alarm database is updated over time, one or more processors may use the false alarm database to train and / or retrain the false alarm AI model, as indicated in box 98. In some cases, when the true alarm database is updated over time, one or more processors may use the true alarm database to train and / or retrain the true alarm AI model, as indicated in box 100. In some cases, when an incoming alarm is classified as a false alarm, one or more processors may be prevented from displaying the incoming alarm on the operator console, as indicated in box 102.

[0031] Figure 4 This is a flowchart illustrating an exemplary workflow 104, in which camera 106 captures a video stream, and video stream 108 detects possible movement within the video stream. An alarm 110 is issued, and alarm 110 is relayed to operator console 112 (which can be considered an example of operator console 22). Initially, the operator at the operator console can manually classify the alarm as a false alarm or a true alarm, as indicated in decision block 114. If it is a true alarm, it is stored in the true alarm database 116. If it is a false alarm, it is stored in the false alarm database 118.

[0032] The true alarm AI model 120 can be trained using the true alarm database 116 (this true alarm AI model can be considered an example of true alarm AI model 16). The false alarm AI model 122 can be trained using the false alarm database 118 (this false alarm AI model can be considered an example of false alarm AI model 14). Alarm 110 is processed by the true alarm AI model 120 and the false alarm AI model 122, and each model classifies alarm 110 as a false alarm or a true alarm and reports the corresponding classification to the inference block 124.

[0033] When inference block 124 determines that both the true alarm AI model 120 and the false alarm AI model 122 classify alarm 110 into the same category (i.e., both classify alarm 110 as a true alarm or both classify alarm 110 as a false alarm), inference block 124 passes control to API block 126. If both classify alarm 110 as a true alarm, the alarm is displayed on operator console 112. If both classify alarm 110 as a false alarm, the alarm is filtered out and not displayed on operator console 112.

[0034] When inference block 124 determines that the true alarm AI model 120 and the false alarm AI model 122 classify alarm 110 into different categories, inference block 124 presents alarm 110 to the operator at operator console 112 and requests the operator to classify alarm 110 as a false alarm or a true alarm, and passes control to decision block 114. In some cases, an alarm of a certain type can be classified as a false alarm at one time and as a true alarm at another time. For example, one operator may classify an alarm as a false alarm, while another operator may classify a similar alarm as a true alarm. Alternatively, one operator may classify an alarm as a false alarm in one week, and the same operator may classify a similar alarm as a true alarm in another week. These situations may result in similar alarms existing in both the true alarm database and the false alarm database, and the true alarm AI model 120 and the false alarm AI model 122 are trained on the true alarm database and the false alarm database, respectively. This may lead to inconsistencies between the true alarm AI model 120 and the false alarm AI model 122. When the real alarm AI model 120 and the false alarm AI model 122 are inconsistent, the system requests the operator at operator console 112 to resolve the conflict and can update the solution in the real alarm database and the false alarm database.

[0035] Figure 5 This is a flowchart illustrating an exemplary workflow 128. The only difference between workflow 128 and workflow 104 is... Figure 5 The workflow 128 transmits both real and false alarms to a public true-false database 130. The true-false database 130 replaces, for example... Figure 4 The diagram shows a separate database of true alarms 116 and a database of false alarms 118. Each alarm in the true-false database 130 is identified as either a false alarm or a true alarm. This allows for the classification of the true-false database 130, enabling the training of a true alarm AI model 120 using those alarms classified as true alarms, and the training of a false alarm AI model 122 using alarms classified as false alarms.

[0036] Although several exemplary embodiments of this disclosure have been described thus, those skilled in the art will readily understand that other embodiments can be made and used within the scope of the appended claims. However, it should be understood that this disclosure is merely illustrative in many respects. Changes may be made to details, particularly those relating to shape, size, arrangement of parts, and exclusion and order of steps, without departing from the scope of this disclosure. The scope of this disclosure is, of course, defined by the language expressed in the appended claims.

Claims

1. A method for mitigating false alarms in a building management system (BMS), the method comprising: The storage uses a fake alarm artificial intelligence (AI) model trained with metadata associated with alarms classified as fake alarms (14); The storage uses a true alarm artificial intelligence (AI) model trained with metadata associated with alarms classified as true alarms (16); Receive incoming alarms; The false alarm AI model is applied to the incoming alarm, wherein the false alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. The true alarm AI model is applied to the incoming alarm, wherein the true alarm AI model classifies the incoming alarm as either the false alarm category or the true alarm category; When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a false alarm, the incoming alarm will be automatically classified as a false alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a true alarm, the incoming alarm will be automatically classified as a true alarm. as well as When the false alarm AI model classifies the incoming alarm as a false alarm and the true alarm AI model classifies the incoming alarm as a true alarm, the incoming alarm is presented to the operator console (22) of the BMS, and the operator console receives a manual classification of the incoming alarm as either a false alarm or a true alarm.

2. The method according to claim 1, further comprising: When an incoming alarm is classified as a true alarm, it needs to be cleared via operator action on the operator console.

3. The method according to claim 1 or 2, further comprising: When the incoming alarm is classified as a false alarm, it is not necessary to clear the incoming alarm via operator action on the operator console.

4. The method according to any one of claims 1 to 3, further comprising: When an incoming alarm is classified as a false alarm, the incoming alarm is not displayed on the operator console of the BMS.

5. The method according to any one of claims 1 to 4, further comprising: When the incoming alarm is classified as a false alarm, the incoming alarm is added to the false alarm database; And using the false alarm database to train and / or retrain the false alarm AI model.

6. The method according to any one of claims 1 to 5, further comprising: When the incoming alarm is classified as a true alarm category, the incoming alarm is added to the true alarm database; And using the real alarm database to train and / or retrain the real alarm AI model.

7. The method according to any one of claims 1 to 6, wherein the false alarm AI model and the true alarm AI model each comprise one of a linear regression-based model or a neural network-based model.

8. The method according to any one of claims 1 to 7, further comprising: When the false alarm AI model classifies the incoming alarm as the true alarm category and the true alarm AI model classifies the incoming alarm as the false alarm category, the incoming alarm is presented to the operator console of the BMS, and the operator console receives a manual classification of the incoming alarm as either the false alarm category or the true alarm category.

9. The method according to any one of claims 1 to 8, further comprising: Receive alarm logs, the alarm logs including logs of alarms including metadata associated with each of the alarms, the alarm logs including alarms classified into both the false alarm category and the true alarm category; The false alarm database is generated and / or updated based on alarms in the alarm log that are classified as false alarms, rather than based on alarms in the alarm log that are classified as true alarms. The true alarm database is generated and / or updated based on alarms in the alarm log that are classified as true alarms, rather than alarms in the alarm log that are classified as false alarms. Use the false alarm database to train and / or retrain the false alarm AI model; and The true alarm database is used to train and / or retrain the true alarm AI model.

10. An alarm management system, the alarm management system comprising: Input terminal (26); A memory (12) for storing a false alarm artificial intelligence (AI) model (14) and a true alarm artificial intelligence (AI) model (16); An operator console (22) includes a user interface (24); A controller (32), operatively coupled to the input, the memory, and the operator console, is configured to: Receive incoming alarms from the input terminal; The false alarm AI model is applied to the incoming alarm, wherein the false alarm AI model classifies the incoming alarm as either a false alarm or a true alarm. The true alarm AI model is applied to the incoming alarm, wherein the true alarm AI model classifies the incoming alarm as either the false alarm category or the true alarm category; When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a false alarm, the incoming alarm will be automatically classified as a false alarm. When both the false alarm AI model and the true alarm AI model classify the incoming alarm as a true alarm, the incoming alarm will be automatically classified as a true alarm. as well as When the false alarm AI model classifies the incoming alarm as a false alarm and the true alarm AI model classifies the incoming alarm as a true alarm, the incoming alarm is presented on the operator console, thereby requesting manual classification of the incoming alarm as either a false alarm or a true alarm.

Images