Operation and maintenance alarm method and device of power system, computer device and storage medium

Abstract

The application relates to an operation and maintenance alarm method and device of a power system, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring abnormal alarm information generated by each device in a new power system; determining the preference degree of each operation and maintenance personnel to each piece of abnormal alarm information by using a pre-trained preference model; determining a target operation and maintenance personnel list corresponding to each piece of abnormal alarm information according to the preference degree; and distributing the abnormal alarm information to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list. The method distributes abnormal alarm information to different operation and maintenance personnel based on the preference of the operation and maintenance personnel to the alarm information, and the abnormal alarm information distributed to each operation and maintenance personnel corresponds to the operation and maintenance work that the operation and maintenance personnel are more proficient in, so that the work efficiency of the operation and maintenance personnel of the new power system is improved.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a power system operation and maintenance alarm method, device, computer equipment, storage medium and computer program product. Background Technology

[0002] The safe operation and maintenance of new power systems is crucial for ensuring their stable operation. Rapidly locating and handling various safety incidents can reduce losses in abnormal situations and guarantee safe and efficient power production. Safety alarms are key to ensuring the normal operation of the power system. If a power system fault or malicious attack occurs during the normal operation of a new power system, maintenance personnel must promptly locate and resolve the fault to ensure the safe and stable operation of the system. Anomaly alarms in new power systems serve as crucial reference information to assist maintenance personnel in effectively carrying out their work.

[0003] To minimize losses caused by sudden anomalies, maintenance personnel need to quickly locate and resolve anomalies. Therefore, the list of anomaly alarms provided to maintenance personnel should facilitate rapid identification and analysis of abnormal states. Currently, some processing technologies for anomaly alarm information in new power systems can classify alarms by type and sort them according to their importance. Maintenance personnel can view the categorized and sorted anomaly alarm information. This categorized and sorted anomaly alarm information can, to some extent, improve the work efficiency of maintenance personnel and ensure the stable operation of the power system.

[0004] Anomaly alarms in new power systems are often complex and massive. Current new power systems are highly information-based and intelligent, containing numerous subsystems, devices, and sensors, among other security entities, and are heterogeneous in type with business coupling between different parts. Given the massive scale of these new power systems, system failures and security incidents occur frequently. A single failure or intrusion can trigger anomaly alarms from various functions of numerous devices. An excessive number of anomaly alarms can severely disrupt maintenance personnel and hinder their work. Therefore, given the complexity and sheer volume of anomaly alarms, maintenance personnel need to sift through the vast amount of alarms to find those within their expertise, while alarms outside their expertise are easily overlooked. Consequently, existing processing technologies for the massive amounts of anomaly alarms generated by new power systems present a challenge for maintenance personnel in efficiently handling such large volumes. Summary of the Invention

[0005] Therefore, it is necessary to provide a power system operation and maintenance alarm method, device, computer equipment, computer-readable storage medium, and computer program product that can improve the processing efficiency of operation and maintenance personnel, in response to the above-mentioned technical problems.

[0006] Firstly, this application provides a method for operation and maintenance alarms in a power system. The method includes:

[0007] Acquire abnormal alarm information generated by various devices in the new power system;

[0008] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each of the aforementioned abnormal alarm messages;

[0009] Based on the preference level, determine the target maintenance personnel list corresponding to each of the above abnormal alarm messages;

[0010] The abnormal alarm information is distributed to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0011] In one embodiment, the method further includes:

[0012] When the preference model has not been trained, the abnormal alarm information will be pushed to all operation and maintenance personnel;

[0013] Obtain feedback from operations and maintenance personnel regarding their preference for the aforementioned abnormal alarm information;

[0014] Based on the maintenance personnel's preference for the abnormal alarm information, a training sample set for the preference model is constructed.

[0015] The preference model is trained based on the training sample set.

[0016] In one embodiment, training the preference model based on the training sample set includes:

[0017] The training sample set is decomposed into a user model and an alarm model using a matrix factorization collaborative filtering algorithm.

[0018] The user model and the alarm model are multiplied together to obtain the predicted degree of preference.

[0019] The user model and the alarm model are iteratively trained using the squared error of the preference degree of the training samples in the training sample set and the predicted preference degree as the minimum loss function.

[0020] Based on the trained user model and alarm model, a preference model is obtained.

[0021] In one embodiment, a training sample set for the preference model is constructed based on the maintenance personnel's preference for the abnormal alarm information, including:

[0022] If the operations and maintenance personnel report a preference for useful anomaly alarm information, then increase the preference value; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then decrease the preference value.

[0023] A training sample set for the model is constructed based on the abnormal alarm information that adjusts the preference level value.

[0024] In one embodiment, if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value is decreased, including:

[0025] If the operation and maintenance personnel report a preference for useful abnormal alarm information, then increasing the preference value of the abnormal alarm information will increase it to half of the current preference value plus 1 / 2.

[0026] If the maintenance personnel report a preference for useless abnormal alarm information, then the preference value for the abnormal alarm information will be reduced by half of the current preference value.

[0027] In one embodiment, the loss function is the sum of the squared error between the preference level of the training samples in the training sample set and the predicted preference level, plus a regularization term.

[0028] In one embodiment, determining the target maintenance personnel list corresponding to each abnormal alarm message based on the preference level includes:

[0029] The maintenance personnel are ranked according to their preference for each abnormal alarm message.

[0030] Based on the sorting results, select the maintenance personnel whose preference level is greater than the preset value to obtain the first list;

[0031] If the proportion of maintenance personnel in the first list is greater than the preset proportion, then the personnel with the preset proportion are selected from the first list in sorted order to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0032] Secondly, this application also provides a power system operation and maintenance alarm device. The device includes:

[0033] The abnormal alarm information acquisition module is used to acquire abnormal alarm information generated by various devices in the new power system;

[0034] The preference degree calculation module is used to determine the preference degree of each operation and maintenance personnel for each of the above-mentioned abnormal alarm messages using a pre-trained preference model;

[0035] The maintenance personnel list determination module is used to determine the target maintenance personnel list corresponding to each abnormal alarm message based on the preference level.

[0036] The abnormal alarm information distribution module is used to distribute the abnormal alarm information to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0037] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to perform the following steps:

[0038] Acquire abnormal alarm information generated by various devices in the new power system;

[0039] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each of the aforementioned abnormal alarm messages;

[0040] Based on the preference level, determine the target maintenance personnel list corresponding to each of the above abnormal alarm messages;

[0041] The abnormal alarm information is distributed to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0042] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, performs the following steps:

[0043] Acquire abnormal alarm information generated by various devices in the new power system;

[0044] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each of the aforementioned abnormal alarm messages;

[0045] Based on the preference level, determine the target maintenance personnel list corresponding to each of the above abnormal alarm messages;

[0046] The abnormal alarm information is distributed to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0047] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, performs the following steps:

[0048] Acquire abnormal alarm information generated by various devices in the new power system;

[0049] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each of the aforementioned abnormal alarm messages;

[0050] Based on the preference level, determine the target maintenance personnel list corresponding to each of the above abnormal alarm messages;

[0051] The abnormal alarm information is distributed to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0052] The aforementioned power system operation and maintenance alarm method, device, computer equipment, storage medium, and computer program product acquire abnormal alarm information generated by various devices in the new power system. Using a pre-trained preference model, it determines the preference level of each operation and maintenance personnel for each abnormal alarm message. Based on the preference level, it determines the target operation and maintenance personnel list corresponding to each abnormal alarm message and distributes the abnormal alarm information to the operation terminals of each operation and maintenance personnel in the corresponding target personnel list. This method, based on the operation and maintenance personnel's preferences for alarm information, personalizes the distribution of abnormal alarm information to different operation and maintenance personnel. The abnormal alarm information distributed to each operation and maintenance personnel corresponds to the operation and maintenance work they are more proficient in, thereby improving the work efficiency of operation and maintenance personnel in the new power system. Attached Figure Description

[0053] Figure 1 This is an application environment diagram of a power system operation and maintenance alarm method in one embodiment;

[0054] Figure 2 This is a flowchart illustrating a power system operation and maintenance alarm method in one embodiment;

[0055] Figure 3 This is a flowchart illustrating matrix decomposition in one embodiment;

[0056] Figure 4 This is a flowchart illustrating a power system operation and maintenance alarm method in another embodiment;

[0057] Figure 5 This is a schematic diagram illustrating the process of calculating and generating a target maintenance personnel list based on a single alarm message in one embodiment.

[0058] Figure 6 This is a structural block diagram of a power system operation and maintenance alarm device in one embodiment;

[0059] Figure 7 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0060] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0061] The power system operation and maintenance alarm method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, the new power system equipment 102 communicates with the server 104 via a network. A data storage system can store the data that the server 104 needs to process. The data storage system can be integrated onto the server 104 or placed in the cloud or on other network servers. The new power system equipment 102 sends abnormal alarm information generated by various devices to the server 104 via the network. The server 104 obtains the abnormal alarm information generated by various devices in the new power system, uses a pre-trained preference model to determine the preference level of each maintenance personnel for each abnormal alarm information, determines the target maintenance personnel list corresponding to each abnormal alarm information based on the preference level, and distributes the abnormal alarm information to each maintenance personnel in the corresponding target maintenance personnel list. The new power system equipment 102 can be various devices in the new power system, and the server 104 can be a standalone server or a server cluster composed of multiple servers.

[0062] In one embodiment, such as Figure 2 The diagram shows a flowchart of a power system operation and maintenance alarm method, which is applied to... Figure 1 The server shown includes the following steps:

[0063] Step 202: Obtain abnormal alarm information generated by various devices in the new power system.

[0064] In the context of new power systems, due to the large number of connected devices and the heterogeneous types of participating entities, many security services are applied in new power systems to ensure the safe and stable operation of the system. However, these multiple security services generate a large number of abnormal alarm messages.

[0065] Specifically, when various devices in the new power system generate abnormal alarm information, the server receives the abnormal alarm information generated by these devices.

[0066] Step 204: Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each abnormal alarm message.

[0067] Because new power systems contain various types of equipment, and even the same type of equipment can be produced by different manufacturers, abnormal alarm information is complex and numerous. In this situation, although alarm information is classified and sorted, individual maintenance personnel have limited capabilities and cannot be proficient in the maintenance of all equipment. Furthermore, different maintenance personnel may be skilled in different maintenance tasks. For example, maintenance personnel A may be skilled in operating equipment from manufacturer A, while maintenance personnel B may be skilled in handling alarm information of type a. Therefore, maintenance personnel need time to find abnormal alarm information that is suitable for their handling, which prevents them from quickly processing abnormal alarms and causes economic losses to the power system.

[0068] Among them, the preference model can reflect the degree of preference of any operation and maintenance personnel for any abnormal alarm information. The preference model is a probability model, and the higher the probability, the more likely the operation and maintenance personnel are to adopt this abnormal alarm information.

[0069] Specifically, upon receiving an anomaly alarm, a pre-trained preference model is used to calculate the preference level for that alarm for each operations and maintenance personnel. Similarly, for each received anomaly alarm, the pre-trained preference model is used to calculate the preference level for each anomaly alarm for each operations and maintenance personnel.

[0070] Step 206: Determine the list of target maintenance personnel corresponding to each abnormal alarm message based on the preference level.

[0071] The target operations and maintenance (O&M) personnel list corresponding to the abnormal alarm information is a list of O&M personnel to whom the abnormal alarm information needs to be distributed. O&M personnel are ranked according to their preference for a particular abnormal alarm information, resulting in a ranked list of O&M personnel for each abnormal alarm information. Based on this ranked list, a truncation operation is performed to obtain the target O&M personnel list for each abnormal alarm information. The truncation length must meet two conditions: first, the length must not be less than the specified minimum distribution amount, which can be 60% of the total number of O&M personnel, meaning that a majority of O&M personnel will receive each alarm information. The proportion of O&M personnel receiving each alarm information can be adjusted according to specific circumstances; second, the preference level must be greater than a certain probability to require distribution, meaning the higher the probability that an O&M personnel will accept the abnormal alarm information, the more likely that abnormal alarm information should be pushed to that O&M personnel. The probability can be 50% or 60%, and this value can be adjusted according to actual conditions.

[0072] Specifically, based on the calculated preference of each operations and maintenance (O&M) personnel for each abnormal alarm message, a sorted list of O&M personnel corresponding to each abnormal alarm message is obtained. The sorted list of O&M personnel is then truncated to obtain a target O&M personnel list corresponding to each abnormal alarm message, thus providing conditions for personalized distribution of abnormal alarm messages to each O&M personnel in the target O&M personnel list.

[0073] Step 208: Distribute the abnormal alarm information to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list.

[0074] Specifically, based on the target maintenance personnel list corresponding to each abnormal alarm message, each abnormal alarm message is distributed to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

[0075] The aforementioned power system operation and maintenance alarm method acquires abnormal alarm information generated by various devices in the new power system. Using a pre-trained preference model, it determines the preference level of each operation and maintenance personnel for each abnormal alarm message. Based on the preference level, it determines the target operation and maintenance personnel list corresponding to each abnormal alarm message and distributes the abnormal alarm information to the operation terminals of each operation and maintenance personnel in the corresponding target personnel list. This method, based on the operation and maintenance personnel's preferences for alarm information, personalizes the distribution of abnormal alarm information to different operation and maintenance personnel. The abnormal alarm information distributed to each operation and maintenance personnel corresponds to the operation and maintenance work they are more proficient in, thus improving the work efficiency of operation and maintenance personnel in the new power system.

[0076] In one embodiment, the power system operation and maintenance alarm method further includes: pushing abnormal alarm information to all operation and maintenance personnel when the preference model has not been trained; obtaining the preference degree of the operation and maintenance personnel for the abnormal alarm information; constructing a training sample set for the preference model based on the preference degree of the operation and maintenance personnel for the abnormal alarm information; and training the preference model based on the training sample set.

[0077] When the preference model is untrained, i.e., in the initial state, the training sample set is empty. At this point, it's assumed that all elements in the training sample set have a value of 0.5, meaning the adoption probability is 50%. That is, in the initial state, any maintenance worker has a 50% probability of adopting any given abnormal alarm. Following the principle of "pushing only those with a probability of at least 50%," every abnormal alarm generated by the new power equipment will be sent to all maintenance workers. Therefore, collaborative filtering in the initial state cannot personalize the distribution of abnormal alarms; the result is the same as the original classification and ranking results, avoiding the cold start problem caused by a lack of historical data or insufficient historical data. Furthermore, to ensure the accuracy of the user preference model used for collaborative filtering, maintenance workers need to provide feedback on alarm adoption after maintenance, allowing for a more accurate calculation of their preference for subsequent abnormal alarms.

[0078] Specifically, after each maintenance worker completes their maintenance work, the system collects their preferences for abnormal alarm information. The abnormal alarm information includes some useful alarm information and some useless alarm information. Based on the feedback from the maintenance workers regarding their preferences for abnormal alarm information, the system improves the content of the training sample set for the preference model and trains the preference model based on the training sample set.

[0079] In this embodiment, the preference model is obtained by the feedback of the operation and maintenance personnel on their preference for abnormal alarm information. The preference model is then trained based on the training sample set to obtain the preference model.

[0080] In one embodiment, training the preference model based on the training sample set includes: decomposing the training sample set into a user model and an alarm model using a collaborative filtering algorithm with matrix factorization; multiplying the user model and the alarm model to obtain the predicted preference level; using the minimum squared error between the preference level of the training samples and the predicted preference level in the training sample set as the loss function, iteratively training the user model and the alarm model; and obtaining the preference model based on the trained user model and the alarm model.

[0081] Collaborative filtering algorithms based on matrix factorization use a matrix model to reflect the degree of user preference; in this application, the user is the operations and maintenance personnel. For example... Figure 3As shown, the matrix factorization-based collaborative filtering algorithm mainly consists of three parts of data: a training sample set, a user model, and an alarm model. The user model and alarm model are two smaller matrices obtained from the training sample set through matrix factorization. Multiplying these two models yields a larger preference model, which should have approximately the same values ​​as those in the training sample set. While the training sample set and preference model are matrices of the same size, the training sample set contains only a small subset of the matrix values, while the calculated preference model is the complete matrix—that is, the preference model is a prediction model calculated based on the training sample set. In actual computation, it is not necessary to use matrix multiplication to obtain the complete preference model; only the vector dot product method is needed to calculate any specified element in the preference model. Therefore, matrix factorization-based collaborative filtering only requires storing a small training sample set, user model, and alarm model to calculate the value of a specified element in the preference model in constant time.

[0082] Specifically, if there are M maintenance personnel in a new power system, and the abnormal alarm information is divided into N types according to different attributes of the alarm equipment (such as the location of the alarm equipment, the manufacturer of the equipment, the type of equipment, the type of alarm, etc.), then there should exist an M-row N-column matrix P. M×N Let represent the probability of M operations and maintenance personnel adopting N different abnormal alarm messages. The element in the i-th row and j-th column of the matrix represents the degree of preference of the i-th operations and maintenance personnel for the j-th alarm message.

[0083] Matrix factorization of the training sample set yields the user model and the alarm model. User model X M×k It is an M x k matrix, which can also be viewed as M k-dimensional feature vectors, meaning that each of the M users corresponds to a k-dimensional feature; Alarm model Y N×k It is an N x k matrix, which can also be viewed as N k-dimensional feature vectors, meaning that each of the N alarms corresponds to a k-dimensional feature. The value of k needs to be analyzed specifically through feature engineering, and it generally will not exceed 10, much smaller than M or N. The user model X obtained through matrix factorization... M×k And alarm model Y N×k The matrix product should approximate the training sample set P as closely as possible. M×N That is, it satisfies the following formula:

[0084] P M×N ≈X M×k ·Y T N×k

[0085] The obtained user model X M×k And alarm model Y N×kTo calculate the preference of a specified user (i∈[1,M]) for a specified alarm message (j∈[1,N]), the feature vector x of the i-th row of the user model should be taken. i The feature vector y in the j-th row of the alarm model j The inner product of the two vectors is the element in the i-th row and j-th column of the desired preference model, which is the predicted value of the preference level.

[0086] Alternating least squares is used to perform matrix factorization on the training sample set. To make the product of the user matrix and the alarm matrix approximate the training sample set, a loss function, which minimizes the loss, needs to be defined.

[0087]

[0088] The loss function contains the sum of two terms: the sum of squared errors and the regularization term. Here, K represents the value of the row and column (i,j) corresponding to the element in the training sample set, and λ is the regularization coefficient used to avoid overfitting. The value of λ needs to be adjusted based on the data characteristics and training results.

[0089] The key calculation formula for alternating least squares can be obtained from the loss function Loss.

[0090] Assuming the values ​​of matrix Y are known, and Y is fixed, taking the partial derivative of the loss function with respect to X and setting it equal to zero yields matrix X:

[0091]

[0092] Assuming the values ​​of matrix X are known, and X is fixed, taking the partial derivative of the loss function with respect to Y and setting it equal to zero yields matrix Y:

[0093]

[0094] Based on the above two formulas, the steps for designing a matrix factorization learning algorithm are as follows:

[0095] 1. Randomly initialize matrix Y.

[0096] 2. Let the loss function be applied to x. i Find the partial derivative to zero, and obtain matrix X from matrix Y.

[0097] 3. Let the loss function be applied to y j Find the partial derivative to be zero, and obtain matrix Y from matrix X.

[0098] 4. Repeat steps 2 and 3 to update matrices X and Y alternately until the loss function converges.

[0099] In this embodiment, matrix P is subjected to alternating least squares method. M×NMatrix factorization is performed to obtain the user model and alarm model, namely the operations and maintenance personnel model and the alarm information model. A preference model is then derived based on the user model and alarm model.

[0100] In one embodiment, a training sample set for a preference model is constructed based on the preference level of operations and maintenance personnel for abnormal alarm information. This includes: increasing the preference level value if the operations and maintenance personnel report a preference level for useful abnormal alarm information, and decreasing the preference level value if the operations and maintenance personnel report a preference level for useless abnormal alarm information; and constructing a training sample set for the model based on the abnormal alarm information for which the preference level value has been adjusted.

[0101] Matrix factorization-based collaborative filtering algorithms only provide a prediction method and do not adaptively update the training sample set. Therefore, it is necessary to design an iterative scheme for the training sample set for matrix factorization-based collaborative filtering algorithms. The data in the training sample set should be dynamically updated according to the iterative scheme.

[0102] Initially, the training sample set is empty. At this point, the training sample set cannot fully reflect the actual preferences of operations and maintenance personnel for abnormal alarm information. Therefore, the training sample set needs continuous iteration to achieve more effective distribution of abnormal alarm information. Specifically, abnormal alarm information can be divided into useful and useless alarm information. For a useful alarm, its probability needs to be increased, i.e., the preference level needs to be increased; similarly, for a useless alarm, its probability needs to be decreased, i.e., the preference level needs to be decreased.

[0103] In this embodiment, by continuously iterating the training sample set in the initial state, the error correction capability of the model is improved, so as to obtain the true preference of operation and maintenance personnel for abnormal alarm information, thereby obtaining a more accurate preference model.

[0104] In one embodiment, if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value is decreased. This includes: if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value for the anomaly alarm information is increased to half of the current preference value plus 1 / 2; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value for the anomaly alarm information is decreased to half of the current preference value.

[0105] Specifically, the iterative formula for the training sample set is as follows:

[0106]

[0107] The above iterative scheme uses a 50% increase approach. For a useful anomaly alarm, its probability needs to be increased, which can be done by increasing the current probability to (1+P). i，j Similarly, for a useless alarm, its probability needs to be reduced, which can be done by reducing it to half of its current probability value, i.e., P. i，j / 2.

[0108] In this embodiment, the probability of useful abnormal alarm information is increased and the probability of useless abnormal alarm information is reduced through an iterative scheme. On the one hand, the iterative formula can ensure that the probability will not be lower than 0% or higher than 100%; on the other hand, it has a large step size for increasing the probability of low probability and decreasing the probability of high probability, which can demonstrate a strong model error correction capability.

[0109] In one embodiment, the loss function is the sum of the squared error between the preference level of the training samples and the predicted preference level in the training sample set, and a regularization term.

[0110] To make the product of the user matrix and the alarm matrix approximate the training sample set, we need to define a loss function, Loss, which is expressed as follows:

[0111]

[0112] The first term in the loss function is the squared error, and the second term is the regularization term. Here, K represents the row and column (i,j) values ​​corresponding to the elements in the training sample set, and λ is the regularization coefficient used to avoid overfitting. The value of λ needs to be adjusted based on the data characteristics and training results.

[0113] In this embodiment, a loss function is defined to minimize the multiplication of the user matrix and the alarm matrix to approximate the training sample set.

[0114] In one embodiment, determining the target maintenance personnel list corresponding to each abnormal alarm message based on preference level includes: sorting the maintenance personnel for each abnormal alarm message according to their preference level; selecting maintenance personnel with a preference level greater than a preset value based on the sorting result to obtain a first list; if the proportion of maintenance personnel in the first list is greater than a preset proportion, then selecting the personnel with the preset proportion from the first list according to the sorting result to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0115] The preset value is a pre-set value, which can be 50% and can be adjusted according to actual conditions. The preset percentage is a preset ratio, which can be 60% and can also be adjusted according to actual conditions.

[0116] Specifically, a preference model is used to calculate the preference level of each operations and maintenance (O&M) personnel for each abnormal alarm message. The O&M personnel corresponding to each abnormal alarm message are then sorted according to their preference level, resulting in a sorted O&M personnel list. This list is then truncated based on a preset value, with O&M personnel whose preference level is greater than the preset value forming the first list. If the ratio of the number of O&M personnel in the first list to the total number of O&M personnel is greater than a preset percentage, the O&M personnel in the first list are truncated sequentially according to the preset percentage, resulting in the target O&M personnel list corresponding to each abnormal alarm message.

[0117] In this embodiment, selecting maintenance personnel based on preset values ​​ensures that their preference for abnormal alarm information exceeds a certain probability, guaranteeing that the alarm will be distributed to that personnel and ensuring the effectiveness of alarm information distribution. Selecting maintenance personnel based on a preset proportion ensures the broad reach of alarm information distribution. Distributing alarm information based on both effectiveness and broad reach improves the efficiency of maintenance personnel in processing abnormal alarm information.

[0118] In one embodiment, such as Figure 4 As shown, a method for issuing operation and maintenance alarms in a power system is provided, including the following steps:

[0119] Step 402: Obtain abnormal alarm information generated by various devices in the new power system.

[0120] Step 404: Using the collaborative filtering algorithm of matrix factorization and the training sample set, a pre-trained preference model is obtained.

[0121] Step 406: Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each abnormal alarm message.

[0122] Step 408: Determine the list of target maintenance personnel corresponding to each abnormal alarm message based on the preference level.

[0123] Step 410: Based on the target maintenance personnel list corresponding to each abnormal alarm message, distribute each abnormal alarm message to the operation terminal of each maintenance personnel in the corresponding target maintenance personnel list.

[0124] Step 412: Sort the multiple abnormal alarm messages received by each operations and maintenance personnel to obtain a list of abnormal alarm messages.

[0125] Specifically, because the new power system generates a large number of abnormal alarm messages, each maintenance personnel will receive a large number of abnormal alarm messages. These abnormal alarm messages are classified and sorted according to their type and importance to obtain an abnormal alarm message list. Thus, each maintenance personnel will have their own abnormal alarm message list.

[0126] Step 414: Each maintenance personnel performs maintenance work according to the list of abnormal alarm information.

[0127] Specifically, maintenance personnel can clearly understand the priority order of various maintenance tasks through the abnormal alarm information list, and then perform corresponding maintenance work on each alarm device according to the priority order.

[0128] Taking the distribution of an abnormal alarm message as an example, such as Figure 5 As shown, a collaborative filtering algorithm is used to calculate the preference level of each operations and maintenance (O&M) personnel for the specific anomaly alert, resulting in an O&M personnel-preference level list. This list is then sorted in descending order of preference level, with higher preference levels appearing earlier in the list. Finally, the sorted O&M personnel-preference level list is truncated according to two truncation criteria to obtain a target O&M personnel list corresponding to each anomaly alert.

[0129] In this embodiment, a pre-trained preference model is used to calculate the preference level of each maintenance personnel for abnormal alarm information generated by each device. Based on this preference level, a list of maintenance personnel for each abnormal alarm is obtained. Then, a target maintenance personnel list for each abnormal alarm is derived from this list. The abnormal alarm information is distributed according to the target maintenance personnel list. The abnormal alarm information received by the maintenance personnel is sorted, and the maintenance personnel perform their maintenance work according to the sorted list. Distributing abnormal alarm information based on the preference model ensures that each maintenance personnel receives maintenance work that they can handle and are relatively proficient in, thereby helping to improve the efficiency of their maintenance work.

[0130] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0131] Based on the same inventive concept, this application also provides a power system operation and maintenance alarm device for implementing the power system operation and maintenance alarm method described above. The solution provided by this device is similar to the solution described in the above method; therefore, the specific limitations of one or more power system operation and maintenance alarm device embodiments provided below can be found in the limitations of the power system operation and maintenance alarm method described above, and will not be repeated here.

[0132] In one embodiment, such as Figure 6 As shown, a power system operation and maintenance alarm device is provided, including: an abnormal alarm information acquisition module 602, a preference degree calculation module 604, an operation and maintenance personnel list determination module 606, and an abnormal alarm information distribution module 608, wherein:

[0133] The abnormal alarm information acquisition module 602 is used to acquire abnormal alarm information generated by various devices in the new power system.

[0134] The preference calculation module 604 is used to determine the preference level of each operation and maintenance personnel for each abnormal alarm message using a pre-trained preference model.

[0135] The operations and maintenance personnel list determination module 606 is used to determine the target operations and maintenance personnel list corresponding to each abnormal alarm message based on preference levels.

[0136] The abnormal alarm information distribution module 608 is used to distribute abnormal alarm information to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list.

[0137] In another embodiment, the preference degree calculation module is used to push abnormal alarm information to all operation and maintenance personnel when the preference model has not been trained; obtain the preference degree of the operation and maintenance personnel on the abnormal alarm information; construct a training sample set of the preference model based on the preference degree of the operation and maintenance personnel on the abnormal alarm information; and train the preference model based on the training sample set.

[0138] In another embodiment, the preference degree calculation module is further used to train the preference model based on the training sample set, including: decomposing the training sample set into a user model and an alarm model using a matrix factorization collaborative filtering algorithm; multiplying the user model and the alarm model to obtain the predicted preference degree; iteratively training the user model and the alarm model using the squared error between the preference degree of the training samples and the predicted preference degree in the training sample set as the loss function; and obtaining the preference model based on the trained user model and alarm model.

[0139] In another embodiment, the preference degree calculation module is further configured to construct a training sample set for the preference model based on the preference degree of the operation and maintenance personnel for abnormal alarm information, including: if the operation and maintenance personnel report a preference degree for useful abnormal alarm information, then increase the preference degree value; if the operation and maintenance personnel report a preference degree for useless abnormal alarm information, then decrease the preference degree value; and construct a training sample set for the model based on the abnormal alarm information with adjusted preference degree values.

[0140] In another embodiment, the preference degree calculation module is further configured to increase the preference degree value if the operation and maintenance personnel report a preference degree for useful abnormal alarm information, and decrease the preference degree value if the operation and maintenance personnel report a preference degree for useless abnormal alarm information. This includes: if the operation and maintenance personnel report a preference degree for useful abnormal alarm information, increasing the preference value of the abnormal alarm information to half of the current preference value plus 1 / 2; if the operation and maintenance personnel report a preference degree for useless abnormal alarm information, decreasing the preference value of the abnormal alarm information to half of the current preference value.

[0141] In another embodiment, the preference calculation module is further configured to use a loss function that is the sum of the squared error between the preference of the training samples in the training sample set and the predicted preference, and a regularization term.

[0142] In another embodiment, the preference degree calculation module is further configured to determine the target maintenance personnel list corresponding to each abnormal alarm message based on the preference degree, including: sorting the maintenance personnel for each abnormal alarm message according to their preference degree for each abnormal alarm message; selecting maintenance personnel with a preference degree greater than a preset value based on the sorting result to obtain a first list; if the proportion of maintenance personnel in the first list is greater than a preset proportion, then selecting the personnel with the preset proportion from the first list according to the sorting to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0143] Each module in the aforementioned power system operation and maintenance alarm device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0144] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data related to training sample sets, user models, alarm models, and preference models. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a power system operation and maintenance alarm method.

[0145] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0146] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0147] Acquire abnormal alarm information generated by various devices in the new power system;

[0148] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each abnormal alarm message;

[0149] Determine the target maintenance personnel list corresponding to each abnormal alarm message based on the degree of preference;

[0150] The abnormal alarm information is distributed to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list.

[0151] In one embodiment, the operation and maintenance alarm method further includes:

[0152] When the preference model has not been trained, anomaly alerts will be pushed to all operations and maintenance personnel.

[0153] Obtain feedback from operations and maintenance personnel regarding their preferences for abnormal alarm information;

[0154] Based on the preference of operations and maintenance personnel for abnormal alarm information, a training sample set for the preference model is constructed.

[0155] The preference model is trained based on the training sample set.

[0156] In one embodiment, training the preference model based on a training sample set includes:

[0157] The training sample set is decomposed into a user model and an alarm model using a collaborative filtering algorithm with matrix factorization.

[0158] Multiply the user model and the alarm model to obtain the predicted degree of preference;

[0159] The user model and the alarm model are iteratively trained by minimizing the squared error between the preference degree of the training samples and the predicted preference degree in the training sample set.

[0160] Based on the trained user model and alarm model, a preference model is obtained.

[0161] In one embodiment, a training sample set for the preference model is constructed based on the maintenance personnel's preference for abnormal alarm information, including:

[0162] If the operations and maintenance personnel report a preference for useful anomaly alarm information, then increase the preference value; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then decrease the preference value.

[0163] The training sample set of the model is constructed based on the abnormal alarm information of the adjusted preference degree value.

[0164] In one embodiment, if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value is decreased, including:

[0165] If the operation and maintenance personnel report a preference for useful abnormal alarm information, then increasing the preference value for abnormal alarm information will increase it to half of the current preference value plus 1 / 2.

[0166] If the operations and maintenance personnel report a preference for useless abnormal alarm information, then the preference value for abnormal alarm information will be reduced by half of the current preference value.

[0167] In one embodiment, the loss function is the sum of the squared error between the preference level of the training samples and the predicted preference level in the training sample set, and a regularization term.

[0168] In one embodiment, determining the target maintenance personnel list corresponding to each abnormal alarm message based on preference level includes:

[0169] Based on the preference of the operation and maintenance personnel for each abnormal alarm message, the operation and maintenance personnel for each abnormal alarm message are ranked.

[0170] Based on the sorting results, select the maintenance personnel whose preference level is greater than the preset value to obtain the first list;

[0171] If the proportion of maintenance personnel in the first list is greater than the preset proportion, then the personnel with the preset proportion are selected from the first list in sorted order to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0172] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0173] Acquire abnormal alarm information generated by various devices in the new power system;

[0174] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each abnormal alarm message;

[0175] Determine the target maintenance personnel list corresponding to each abnormal alarm message based on the degree of preference;

[0176] The abnormal alarm information is distributed to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list.

[0177] In one embodiment, the operation and maintenance alarm method further includes:

[0178] When the preference model has not been trained, anomaly alerts will be pushed to all operations and maintenance personnel.

[0179] Obtain feedback from operations and maintenance personnel regarding their preferences for abnormal alarm information;

[0180] Based on the preference of operations and maintenance personnel for abnormal alarm information, a training sample set for the preference model is constructed.

[0181] The preference model is trained based on the training sample set.

[0182] In one embodiment, training the preference model based on a training sample set includes:

[0183] The training sample set is decomposed into a user model and an alarm model using a collaborative filtering algorithm with matrix factorization.

[0184] Multiply the user model and the alarm model to obtain the predicted degree of preference;

[0185] The user model and the alarm model are iteratively trained by minimizing the squared error between the preference degree of the training samples and the predicted preference degree in the training sample set.

[0186] Based on the trained user model and alarm model, a preference model is obtained.

[0187] In one embodiment, a training sample set for the preference model is constructed based on the maintenance personnel's preference for abnormal alarm information, including:

[0188] If the operations and maintenance personnel report a preference for useful anomaly alarm information, then increase the preference value; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then decrease the preference value.

[0189] The training sample set of the model is constructed based on the abnormal alarm information of the adjusted preference degree value.

[0190] In one embodiment, if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value is decreased, including:

[0191] If the operation and maintenance personnel report a preference for useful abnormal alarm information, then increasing the preference value for abnormal alarm information will increase it to half of the current preference value plus 1 / 2.

[0192] If the operations and maintenance personnel report a preference for useless abnormal alarm information, then the preference value for abnormal alarm information will be reduced by half of the current preference value.

[0193] In one embodiment, the loss function is the sum of the squared error between the preference level of the training samples and the predicted preference level in the training sample set, and a regularization term.

[0194] In one embodiment, determining the target maintenance personnel list corresponding to each abnormal alarm message based on preference level includes:

[0195] Based on the preference of the operation and maintenance personnel for each abnormal alarm message, the operation and maintenance personnel for each abnormal alarm message are ranked.

[0196] Based on the sorting results, select the maintenance personnel whose preference level is greater than the preset value to obtain the first list;

[0197] If the proportion of maintenance personnel in the first list is greater than the preset proportion, then the personnel with the preset proportion are selected from the first list in sorted order to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0198] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps:

[0199] Acquire abnormal alarm information generated by various devices in the new power system;

[0200] Using a pre-trained preference model, determine the degree of preference each operations and maintenance personnel has for each abnormal alarm message;

[0201] Determine the target maintenance personnel list corresponding to each abnormal alarm message based on the degree of preference;

[0202] The abnormal alarm information is distributed to the operation terminals of each operation and maintenance personnel in the corresponding target operation and maintenance personnel list.

[0203] In one embodiment, the operation and maintenance alarm method further includes:

[0204] When the preference model has not been trained, anomaly alerts will be pushed to all operations and maintenance personnel.

[0205] Obtain feedback from operations and maintenance personnel regarding their preferences for abnormal alarm information;

[0206] Based on the preference of operations and maintenance personnel for abnormal alarm information, a training sample set for the preference model is constructed.

[0207] The preference model is trained based on the training sample set.

[0208] In one embodiment, training the preference model based on a training sample set includes:

[0209] The training sample set is decomposed into a user model and an alarm model using a collaborative filtering algorithm with matrix factorization.

[0210] Multiply the user model and the alarm model to obtain the predicted degree of preference;

[0211] The user model and the alarm model are iteratively trained by minimizing the squared error between the preference degree of the training samples and the predicted preference degree in the training sample set.

[0212] Based on the trained user model and alarm model, a preference model is obtained.

[0213] In one embodiment, a training sample set for the preference model is constructed based on the maintenance personnel's preference for abnormal alarm information, including:

[0214] If the operations and maintenance personnel report a preference for useful anomaly alarm information, then increase the preference value; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then decrease the preference value.

[0215] The training sample set of the model is constructed based on the abnormal alarm information of the adjusted preference degree value.

[0216] In one embodiment, if the operations and maintenance personnel report a preference for useful anomaly alarm information, the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, the preference value is decreased, including:

[0217] If the operation and maintenance personnel report a preference for useful abnormal alarm information, then increasing the preference value for abnormal alarm information will increase it to half of the current preference value plus 1 / 2.

[0218] If the operations and maintenance personnel report a preference for useless abnormal alarm information, then the preference value for abnormal alarm information will be reduced by half of the current preference value.

[0219] In one embodiment, the loss function is the sum of the squared error between the preference level of the training samples and the predicted preference level in the training sample set, and a regularization term.

[0220] In one embodiment, determining the target maintenance personnel list corresponding to each abnormal alarm message based on preference level includes:

[0221] Based on the preference of the operation and maintenance personnel for each abnormal alarm message, the operation and maintenance personnel for each abnormal alarm message are ranked.

[0222] Based on the sorting results, select the maintenance personnel whose preference level is greater than the preset value to obtain the first list;

[0223] If the proportion of maintenance personnel in the first list is greater than the preset proportion, then the personnel with the preset proportion are selected from the first list in sorted order to obtain the target maintenance personnel list corresponding to each abnormal alarm message.

[0224] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0225] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0226] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for operation and maintenance alarm in a power system, characterized in that, The method includes: Acquire abnormal alarm information generated by various devices in the new power system; Using a pre-trained preference model, the preference level of each operations and maintenance personnel for each of the aforementioned abnormal alarm messages is determined; the preference model and the training sample set are matrices of the same size, and the training sample set is composed of the preference levels of each of the aforementioned operations and maintenance personnel for each of the aforementioned abnormal alarm messages; The preference model is obtained as follows: A collaborative filtering algorithm based on matrix factorization is used to perform matrix factorization on the training sample set to obtain a user model and an alarm model; the user model and the alarm model are multiplied to obtain the predicted preference level; the user model and the alarm model are iteratively trained using the squared error between the preference level of the training samples and the predicted preference level in the training sample set as the minimum loss function; and the preference model is obtained based on the trained user model and the alarm model. The maintenance personnel are ranked according to their preference for each abnormal alarm message. Based on the sorting results, select the maintenance personnel whose preference level is greater than the preset value to obtain the first list; If the proportion of maintenance personnel in the first list is greater than the preset proportion, then the personnel with the preset proportion are selected from the first list in sorted order to obtain the target maintenance personnel list corresponding to each abnormal alarm message; the abnormal alarm message is distributed to the operation terminal of each maintenance personnel in the corresponding target maintenance personnel list.

2. The method according to claim 1, characterized in that, The method further includes: When the preference model has not been trained, the abnormal alarm information will be pushed to all operation and maintenance personnel; Obtain feedback from operations and maintenance personnel regarding their preference for the aforementioned abnormal alarm information; Based on the maintenance personnel's preference for the abnormal alarm information, a training sample set for the preference model is constructed. The preference model is trained based on the training sample set.

3. The method according to claim 2, characterized in that, Based on the maintenance personnel's preference for the abnormal alarm information, a training sample set for the preference model is constructed, including: If the operations and maintenance personnel report a preference for useful anomaly alarm information, then increase the preference value; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then decrease the preference value. A training sample set for the model is constructed based on the abnormal alarm information that adjusts the preference level value.

4. The method according to claim 3, characterized in that, If the operations and maintenance personnel report a preference for useful anomaly alarm information, increase the preference value; if they report a preference for useless anomaly alarm information, decrease the preference value, including: If the operation and maintenance personnel report a preference for useful abnormal alarm information, then increasing the preference value of the abnormal alarm information will increase it to half of the current preference value plus 1 / 2. If the maintenance personnel report a preference for useless abnormal alarm information, then the preference value for the abnormal alarm information will be reduced by half of the current preference value.

5. The method according to claim 1, characterized in that, The loss function is the sum of the squared error between the preference level of the training samples and the predicted preference level in the training sample set, and the regularization term.

6. A power system operation and maintenance alarm device, characterized in that, The device includes: The abnormal alarm information acquisition module is used to acquire abnormal alarm information generated by various devices in the new power system; The preference degree calculation module is used to determine the preference degree of each operation and maintenance personnel for each of the above-mentioned abnormal alarm messages using a pre-trained preference model; The maintenance personnel list determination module is used to determine the target maintenance personnel list corresponding to each abnormal alarm message based on the preference level. The abnormal alarm information distribution module is used to distribute the abnormal alarm information to the operation terminals of each maintenance personnel in the corresponding target maintenance personnel list.

7. The apparatus according to claim 6, characterized in that, The preference degree calculation module is also used for: When the preference model has not been trained, the abnormal alarm information is pushed to all operation and maintenance personnel; the preference level of the operation and maintenance personnel for the abnormal alarm information is obtained. Based on the maintenance personnel's preference for the abnormal alarm information, a training sample set for the preference model is constructed. The preference model is trained based on the training sample set.

8. The apparatus according to claim 7, characterized in that, The preference degree calculation module is also used for: If the operations and maintenance personnel report a preference for useful anomaly alarm information, then the preference value is increased; if the operations and maintenance personnel report a preference for useless anomaly alarm information, then the preference value is decreased; a training sample set for the model is constructed based on the anomaly alarm information with the adjusted preference values.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images