Method and apparatus for generating maintenance strategy for pumped storage device, and computer device, readable storage medium and program product

Abstract

The present application relates to a method and apparatus for generating a maintenance strategy for a pumped storage device, and a computer device, a computer-readable storage medium and a computer program product. The method comprises: on the basis of a device maintenance manual of a pumped storage device, constructing an initial maintenance item list; inputting current device data of the pumped storage device into a device state analysis model, so as to obtain a current device state of the pumped storage device, and on the basis of the current device state and historical maintenance data of the pumped storage device, updating the initial maintenance item list, so as to obtain a candidate maintenance item list; inputting the current device data and the current device state of the pumped storage device into a device fault prediction model, so as to obtain fault information of the pumped storage device, and on the basis of the fault information, updating the candidate maintenance item list again, so as to obtain a target maintenance item list; and on the basis of the target maintenance item list, generating a maintenance strategy for the pumped storage device. By means of the method, the accuracy of determining a maintenance strategy can be improved.

Description

Methods, apparatus, computer equipment, readable storage media, and program products for generating maintenance strategies for pumped storage hydroelectric power plants. Technical Field

[0001] This application relates to the field of power grid technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for generating maintenance strategies for pumped storage equipment. Background Technology

[0002] In the field of power grid technology, in order to maintain pumped storage equipment and ensure its normal operation, it is necessary to carry out regular maintenance of pumped storage equipment through relevant maintenance strategies.

[0003] In traditional technologies, maintenance strategies for pumped storage facilities are generally determined based on human experience. However, relying solely on human experience when formulating maintenance strategies for pumped storage facilities can easily lead to limitations in the resulting strategies and a low accuracy rate in determining the maintenance strategies. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for generating maintenance strategies for pumped storage equipment, which can improve the accuracy of determining maintenance strategies, in response to the above-mentioned technical problems.

[0005] Firstly, this application provides a method for generating maintenance strategies for pumped storage equipment, including:

[0006] Obtain the equipment maintenance manual for the pumped storage equipment;

[0007] Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment;

[0008] The current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list.

[0009] The current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0010] Based on the target maintenance item list, a maintenance strategy for the pumped storage equipment is generated.

[0011] In one embodiment, the pre-trained device state analysis model includes multiple sub-device state analysis models;

[0012] The step of inputting the current equipment data of the pumped storage equipment into a pre-trained equipment state analysis model to obtain the current equipment state of the pumped storage equipment includes:

[0013] The current equipment data of the pumped storage device is processed by feature extraction to obtain the target feature vector;

[0014] The target feature vector is input into the multiple sub-equipment state analysis models respectively to obtain the equipment state of the pumped storage equipment output by each sub-equipment state analysis model, and the state prediction probability corresponding to the equipment state.

[0015] From each sub-device state analysis model, select the target sub-device state analysis model whose output device state prediction probability is greater than the first preset probability.

[0016] From the equipment states of the pumped storage equipment output by the status analysis model of each target sub-equipment, the equipment state that appears most frequently is selected as the current equipment state of the pumped storage equipment.

[0017] In one embodiment, the step of performing feature extraction processing on the current equipment data of the pumped storage device to obtain a target feature vector includes:

[0018] The current equipment data of the pumped storage device is input into a pre-trained importance prediction model to obtain the importance prediction probability of the current equipment data under various preset importance levels.

[0019] From the various preset importance values, the preset importance value with the highest predicted probability is selected as the importance value of the current device data;

[0020] From the current device data, select device data with an importance greater than a preset importance level and designate them as key device data;

[0021] The key equipment data is input into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key equipment data;

[0022] Multiple feature vectors of the key equipment data are fused to obtain a fused feature vector, which is used as the target feature vector.

[0023] In one embodiment, the initial maintenance item list is updated based on the current equipment status and historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list, including:

[0024] The current equipment status, the historical maintenance data of the pumped storage equipment, and the initial maintenance item list are input into a pre-trained maintenance item prediction model to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list.

[0025] From the various maintenance items, the maintenance items with a predicted maintenance probability greater than the second preset probability are identified as target maintenance items.

[0026] The maintenance items other than the target maintenance item are deleted from the initial maintenance item list to obtain the candidate maintenance item list.

[0027] In one embodiment, the pre-trained equipment failure prediction model includes a first equipment failure prediction model and a second equipment failure prediction model;

[0028] The step of inputting the current equipment data and current equipment status of the pumped storage equipment into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment includes:

[0029] The current device data and the current device state are subjected to feature extraction processing to obtain the feature vector corresponding to the current device data and the feature vector corresponding to the current device state;

[0030] The feature vector corresponding to the current device data and the feature vector corresponding to the current device state are concatenated to obtain a concatenated feature vector.

[0031] The spliced ​​feature vector is input into the first equipment fault prediction model to obtain the current fault information of the pumped storage equipment, and the spliced ​​feature vector is input into the second equipment fault prediction model to obtain the future fault information of the pumped storage equipment.

[0032] The current fault information and the future fault information are fused together to obtain the fault information of the pumped storage equipment.

[0033] In one embodiment, updating the candidate maintenance item list again based on the fault information to obtain the target maintenance item list includes:

[0034] Based on the current fault information of the pumped storage equipment, the correspondence between the current fault information and the maintenance items is queried to obtain the first maintenance item corresponding to the current fault information.

[0035] Based on the future fault information of the pumped storage equipment, the correspondence between the future fault information and the maintenance items is queried to obtain the second maintenance item corresponding to the future fault information.

[0036] Add the first maintenance item and the second maintenance item to the candidate maintenance item list to obtain the target maintenance item list.

[0037] Secondly, this application also provides a pumped storage equipment maintenance strategy generation device, comprising:

[0038] The information acquisition module is used to acquire the equipment maintenance manual of the pumped storage equipment;

[0039] The list construction module is used to construct an initial maintenance item list for the pumped storage equipment based on the equipment maintenance manual.

[0040] The first update module is used to input the current equipment data of the pumped storage equipment into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment, and update the initial maintenance item list based on the current equipment status and the historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list.

[0041] The second update module is used to input the current equipment data and current equipment status of the pumped storage equipment into a pre-trained equipment fault prediction model to obtain the fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0042] The strategy generation module is used to generate a maintenance strategy for the pumped storage equipment based on the target maintenance item list.

[0043] Thirdly, this application also provides a computer device, 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:

[0044] Obtain the equipment maintenance manual for the pumped storage equipment;

[0045] Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment;

[0046] The current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list.

[0047] The current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0048] Based on the target maintenance item list, a maintenance strategy for the pumped storage equipment is generated.

[0049] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0050] Obtain the equipment maintenance manual for the pumped storage equipment;

[0051] Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment;

[0052] The current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list.

[0053] The current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0054] Based on the target maintenance item list, a maintenance strategy for the pumped storage equipment is generated.

[0055] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, performs the following steps:

[0056] Obtain the equipment maintenance manual for the pumped storage equipment;

[0057] Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment;

[0058] The current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list.

[0059] The current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0060] Based on the target maintenance item list, a maintenance strategy for the pumped storage equipment is generated.

[0061] The aforementioned method, apparatus, computer equipment, computer-readable storage medium, and computer program product for generating maintenance strategies for pumped storage equipment first obtain the equipment maintenance manual for the pumped storage equipment and construct an initial maintenance item list based on the manual. Next, the current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status. Based on the current equipment status and historical maintenance data, the initial maintenance item list is updated to obtain a candidate maintenance item list. Then, the current equipment data and current equipment status are input into a pre-trained equipment fault prediction model to obtain fault information. Based on the fault information, the candidate maintenance item list is updated again to obtain a target maintenance item list. The fault information includes current fault information and future fault information. Finally, a maintenance strategy for the pumped storage equipment is generated based on the target maintenance item list. In this way, when generating maintenance strategies for pumped storage equipment, the equipment maintenance manual, current equipment status, historical maintenance data, current fault information, and future fault information are comprehensively considered. This approach helps to determine the target maintenance item list, which improves the accuracy of the target maintenance item list. Consequently, the maintenance strategies generated based on the target maintenance item list are more accurate, thus increasing the accuracy of maintenance strategy determination. At the same time, it avoids the limitations of maintenance strategies developed based on human experience, which result in lower accuracy of maintenance strategy determination. Attached Figure Description

[0062] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0063] Figure 1 is a flowchart illustrating a method for generating maintenance strategies for pumped storage equipment in one embodiment;

[0064] Figure 2 is a flowchart illustrating the steps of inputting the current equipment data of the pumped storage equipment into a pre-trained equipment state analysis model to obtain the current equipment state of the pumped storage equipment in one embodiment.

[0065] Figure 3 is a flowchart illustrating the method for generating maintenance strategies for pumped storage equipment in another embodiment;

[0066] Figure 4 is a flowchart illustrating a method for determining maintenance strategies for pumped storage equipment in one embodiment.

[0067] Figure 5 is a structural block diagram of a pumped storage equipment maintenance strategy generation device in one embodiment;

[0068] Figure 6 is an internal structure diagram of a computer device in one embodiment. Detailed Implementation

[0069] 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.

[0070] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0071] Figure 1 is a flowchart illustrating a method for generating a maintenance strategy for pumped storage equipment according to an exemplary embodiment. As shown in Figure 1, this method for generating a maintenance strategy for pumped storage equipment is applied to a server. It is understood that this method can also be applied to a terminal, and further to a system including a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets; the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. In this exemplary embodiment, the method includes the following steps S101 to S105. Wherein:

[0072] Step S101: Obtain the equipment maintenance manual for the pumped storage equipment.

[0073] Among them, the equipment maintenance manual refers to one manual per type. One manual per type means that a separate booklet or manual is written for pumped storage equipment. Such booklets or manuals usually contain detailed parameters, operation guidelines, maintenance suggestions, troubleshooting methods and other relevant information for this type of pumped storage equipment, aiming to help staff better understand and use the pumped storage equipment.

[0074] For example, the server retrieves the equipment maintenance manual for the pumped storage equipment from a local database or from the network.

[0075] Step S102: Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment.

[0076] The initial maintenance item list includes multiple maintenance items, as well as the maintenance type, maintenance time, maintenance cycle, and maintenance requirements for each maintenance item.

[0077] For example, the server determines the maintenance items corresponding to the pumped storage equipment based on the equipment maintenance manual, and constructs an initial maintenance item list for the pumped storage equipment based on the maintenance items corresponding to the pumped storage equipment.

[0078] Step S103: Input the current equipment data of the pumped storage equipment into the pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, update the initial maintenance item list to obtain the candidate maintenance item list.

[0079] The current equipment data refers to basic parameters, electric generator parameters, reversible pump-turbine parameters, and other key parameters. Basic parameters include total installed capacity, annual power generation, maximum load, rated frequency, and highest voltage level; electric generator parameters include generator type and ventilation / cooling method; reversible pump-turbine parameters include head, flow rate, speed, efficiency, power, and runner diameter; other key parameters include available power system capacity, power plant unit startup method, response time, and lifespan.

[0080] Among them, the pre-trained equipment status analysis model refers to the network model used to output the current equipment status of pumped storage equipment, such as neural network model, deep learning model, large model, etc.

[0081] Among them, historical maintenance data is used to characterize the historical maintenance status of pumped storage equipment, such as historical maintenance time and historical maintenance items.

[0082] Based on the current equipment status and historical maintenance data of pumped storage equipment, it is possible to determine which maintenance items in the initial maintenance item list need to be retained and which need to be deleted, thereby obtaining a candidate maintenance item list.

[0083] The candidate maintenance item list includes maintenance items that need to be retained (i.e., candidate maintenance items), as well as the maintenance type, maintenance time, maintenance cycle, and maintenance requirements for each maintenance item that needs to be retained.

[0084] For example, the server retrieves the current equipment data and historical maintenance data of the pumped storage equipment from the local database, performs feature extraction processing on the current equipment data to obtain a feature vector, and inputs the feature vector into a pre-trained equipment status analysis model. The equipment status analysis model performs status prediction processing on the feature vector to obtain the predicted probability of the pumped storage equipment in various preset equipment states. From the preset equipment states, the preset equipment state with the highest predicted probability is selected as the current equipment state of the pumped storage equipment. Next, based on the current equipment state and the historical maintenance data of the pumped storage equipment, maintenance items that match the current equipment state and historical maintenance data of the pumped storage equipment are selected from the initial maintenance item list as candidate maintenance items. Finally, a candidate maintenance item list is obtained based on the candidate maintenance items.

[0085] Furthermore, the pre-trained equipment state analysis model is trained in the following manner: the server obtains the current sample equipment data of the sample pumped storage equipment and inputs the current sample equipment data of the sample pumped storage equipment into the equipment state analysis model to be trained to obtain the predicted equipment state of the sample pumped storage equipment; the actual equipment state of the sample pumped storage equipment is obtained, and the equipment state analysis model to be trained is iteratively trained according to the difference between the predicted equipment state and the actual equipment state of the sample pumped storage equipment to obtain the trained equipment state analysis model, which serves as the pre-trained equipment state analysis model.

[0086] Step S104: Input the current equipment data and current equipment status of the pumped storage equipment into the pre-trained equipment fault prediction model to obtain the fault information of the pumped storage equipment. Based on the fault information, update the candidate maintenance item list again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0087] Among them, the pre-trained equipment fault prediction model refers to the network model used to output fault information of pumped storage equipment, such as neural network model, deep learning model, large model, etc.

[0088] The current fault information refers to the fault information currently existing in the pumped storage equipment.

[0089] Among them, future fault information refers to the fault information that may occur in the pumped storage equipment in the future.

[0090] Based on the fault information, the necessary new maintenance items can be identified and added to the candidate maintenance item list, thus obtaining the target maintenance item list. The newly added maintenance items correspond to the fault information; for example, if the fault information is A, then maintenance item a is added.

[0091] The target maintenance item list includes newly added maintenance items as well as a candidate maintenance item list.

[0092] For example, the server performs feature extraction processing on the current equipment data and current equipment status of the pumped storage equipment to obtain feature vectors corresponding to the current equipment data and current equipment status. These feature vectors are then input into a pre-trained equipment fault prediction model. The model performs fault prediction processing on these feature vectors to obtain current and future fault information of the pumped storage equipment. This current and future fault information is then combined to obtain the pumped storage equipment's fault information. Finally, the server queries the correspondence between the fault information and maintenance items to obtain the maintenance items corresponding to the pumped storage equipment's fault information. These new maintenance items are then added to the candidate maintenance item list to obtain the target maintenance item list for the pumped storage equipment.

[0093] Furthermore, the pre-trained equipment failure prediction model is trained in the following manner: The server obtains the current sample equipment data and current sample equipment status of the sample pumped storage equipment, and inputs the current sample equipment data and current sample equipment status of the sample pumped storage equipment into the equipment failure prediction model to be trained to obtain the predicted failure information of the sample pumped storage equipment; the server obtains the actual failure information of the sample pumped storage equipment, and iteratively trains the equipment failure prediction model to be trained based on the difference between the predicted failure information and the actual failure information of the sample pumped storage equipment to obtain the trained equipment failure prediction model, which serves as the pre-trained equipment failure prediction model.

[0094] Step S105: Generate a maintenance strategy for pumped storage equipment based on the target maintenance item list.

[0095] Among them, the maintenance strategy refers to the strategy for maintaining pumped storage equipment, specifically the list of target maintenance items.

[0096] For example, the server will identify the target maintenance item list as a maintenance strategy for pumped storage equipment.

[0097] In the above-mentioned method for generating maintenance strategies for pumped storage equipment, the equipment maintenance manual of the pumped storage equipment is first obtained, and an initial maintenance item list for the pumped storage equipment is constructed based on the equipment maintenance manual. Then, the current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list. Then, the current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain the fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain a target maintenance item list. The fault information includes current fault information and future fault information. Finally, the maintenance strategy for the pumped storage equipment is generated based on the target maintenance item list. In this way, when generating maintenance strategies for pumped storage equipment, the equipment maintenance manual, current equipment status, historical maintenance data, current fault information, and future fault information are comprehensively considered. This approach helps to determine the target maintenance item list, which improves the accuracy of the target maintenance item list. Consequently, the maintenance strategies generated based on the target maintenance item list are more accurate, thus increasing the accuracy of maintenance strategy determination. At the same time, it avoids the limitations of maintenance strategies developed based on human experience, which result in lower accuracy of maintenance strategy determination.

[0098] In an exemplary embodiment, as shown in FIG2, the pre-trained equipment state analysis model includes multiple sub-equipment state analysis models; then, step S103 above, which inputs the current equipment data of the pumped storage equipment into the pre-trained equipment state analysis model to obtain the current equipment state of the pumped storage equipment, includes the following steps S201 to S204. Wherein:

[0099] Step S201: Perform feature extraction processing on the current equipment data of the pumped storage equipment to obtain the target feature vector.

[0100] Step S202: Input the target feature vector into multiple sub-equipment state analysis models respectively to obtain the equipment state of the pumped storage equipment output by each sub-equipment state analysis model, as well as the state prediction probability corresponding to the equipment state.

[0101] Step S203: From each sub-equipment state analysis model, select the target sub-equipment state analysis model whose output pumped storage equipment state has a state prediction probability greater than the first preset probability.

[0102] Step S204: From the equipment states of the pumped storage equipment output by the status analysis model of each target sub-equipment, select the equipment state that appears most frequently and use it as the current equipment state of the pumped storage equipment.

[0103] Among them, the sub-equipment status analysis model refers to the network model used to output the equipment status of pumped storage equipment, such as neural network models, deep learning models, etc. The model structure of different sub-equipment status analysis models is different.

[0104] The target feature vector refers to the feature vector corresponding to the current equipment data of the pumped storage equipment.

[0105] For example, the server determines a feature extraction model corresponding to the current equipment data of the pumped storage device, and performs feature extraction processing on the current equipment data of the pumped storage device using this feature extraction model to obtain a target feature vector. Then, the target feature vector is input into multiple sub-device state analysis models to obtain the state prediction probability of the pumped storage device under each preset equipment state output by each sub-device state analysis model. For each sub-device state analysis model, the preset equipment state with the highest state prediction probability is selected from the preset equipment states and used as the equipment state of the pumped storage device output by each sub-device state analysis model. The state prediction probability corresponding to the preset equipment state with the highest state prediction probability is then used as the output state of each sub-device state analysis model. The process involves several steps: First, determining the predicted state probability of the pumped storage equipment's state. Then, selecting the sub-equipment state analysis models from each sub-equipment state analysis model whose predicted state probability is greater than a first preset probability, and using these as target sub-equipment state analysis models. Finally, selecting the equipment state that appears most frequently from the output of each target sub-equipment state analysis model, and using this as the current equipment state of the pumped storage equipment. For example, if the equipment states of the pumped storage equipment output by the target sub-equipment state analysis models M1, M2, M3, and M4 are S1, S2, S1, and S3 respectively, then equipment state S1 appears most frequently, and therefore, equipment state S1 is taken as the current equipment state of the pumped storage equipment.

[0106] In this embodiment, the device status of the pumped storage device output by each sub-device status analysis model and the corresponding state prediction probability are first obtained. Then, from each sub-device status analysis model, target sub-device status analysis models whose output state prediction probability of the pumped storage device is greater than a first preset probability are selected. Finally, from the device status of the pumped storage device output by each target sub-device status analysis model, the device status that appears most frequently is selected as the current device status of the pumped storage device. This helps to improve the accuracy of determining the current device status of the pumped storage device.

[0107] In an exemplary embodiment, step S201 above, which involves performing feature extraction processing on the current equipment data of the pumped storage equipment to obtain a target feature vector, specifically includes the following: inputting the current equipment data of the pumped storage equipment into a pre-trained importance prediction model to obtain the importance prediction probability of the current equipment data under various preset importance levels; selecting the preset importance level with the highest importance prediction probability from among the preset importance levels as the importance level of the current equipment data; selecting equipment data with an importance greater than the preset importance level from the current equipment data as key equipment data; inputting the key equipment data into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key equipment data; and fusing the multiple feature vectors of the key equipment data to obtain a fused feature vector as the target feature vector.

[0108] Among them, the pre-trained importance prediction model refers to the network model used to output the importance of the current device data, such as the attention mechanism model.

[0109] Among them, the model structures of different pre-trained feature extraction models are different.

[0110] For example, the server inputs the current equipment data of the pumped storage equipment into a pre-trained importance prediction model. The pre-trained importance prediction model performs importance prediction processing on the current equipment data to obtain the importance prediction probability of the current equipment data under various preset importance levels. Then, from the various preset importance levels, the preset importance level with the highest importance prediction probability is selected as the importance level of the current equipment data. Next, from the current equipment data, equipment data with an importance greater than the preset importance level is selected as key equipment data. Finally, the key equipment data is input into multiple pre-trained feature extraction models. Through each pre-trained feature extraction model, feature extraction processing is performed on the key equipment data to obtain the feature vector of the key equipment data output by each pre-trained feature extraction model. These feature vectors are then fused to obtain a fused feature vector, which is used as the target feature vector.

[0111] In this embodiment, the importance of the current device data is determined by a pre-trained importance prediction model. Then, based on the importance of the current device data, key device data is selected from the current device data. Finally, the key device data is input into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key device data, and these feature vectors are fused to obtain the target feature vector, which helps to improve the accuracy of the target feature vector determination.

[0112] In an exemplary embodiment, step S103, which updates the initial maintenance item list based on the current equipment status and historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list, specifically includes the following: inputting the current equipment status, historical maintenance data of the pumped storage equipment, and the initial maintenance item list into a pre-trained maintenance item prediction model to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list; identifying maintenance items whose predicted maintenance probability is greater than a second preset probability from among the maintenance items, and using them as target maintenance items; deleting maintenance items from the initial maintenance item list other than the target items to obtain the candidate maintenance item list.

[0113] Among them, the pre-trained maintenance project prediction model refers to the network model used to output the predicted maintenance probability of maintenance projects, such as neural network models, deep learning models, etc.

[0114] Among them, the predicted maintenance probability of a maintenance project is used to characterize the necessity of maintenance for the maintenance project.

[0115] Among them, the target maintenance item refers to the maintenance item in the initial maintenance item list whose predicted maintenance probability is greater than the second preset probability, and is used to indicate the maintenance item that needs to be retained.

[0116] For example, the server inputs the current equipment status, historical maintenance data of the pumped storage equipment, and an initial maintenance item list into a pre-trained maintenance item prediction model. The pre-trained maintenance item prediction model performs feature extraction processing on the current equipment status and historical maintenance data to obtain feature vectors corresponding to the current equipment status and historical maintenance data. Then, it performs maintenance prediction processing on the feature vectors to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list. Next, from each maintenance item, the maintenance items with a predicted maintenance probability greater than a second preset probability are identified as target maintenance items. Finally, the maintenance items in the initial maintenance item list other than the target items are deleted to obtain a processed maintenance item list, which serves as a candidate maintenance item list.

[0117] In this embodiment, the initial maintenance item list is updated based on the current equipment status and historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list. This makes the determined candidate maintenance item list more accurate, thereby improving the accuracy of the candidate maintenance item list determination.

[0118] In an exemplary embodiment, the pre-trained equipment fault prediction model includes a first equipment fault prediction model and a second equipment fault prediction model. Therefore, step S104, which inputs the current equipment data and current equipment state of the pumped storage equipment into the pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment, specifically includes the following: performing feature extraction processing on the current equipment data and current equipment state to obtain feature vectors corresponding to the current equipment data and the current equipment state; concatenating the feature vectors corresponding to the current equipment data and the current equipment state to obtain a concatenated feature vector; inputting the concatenated feature vector into the first equipment fault prediction model to obtain the current fault information of the pumped storage equipment, and inputting the concatenated feature vector into the second equipment fault prediction model to obtain future fault information of the pumped storage equipment; and fusing the current fault information and future fault information to obtain the final fault information of the pumped storage equipment.

[0119] Among them, the first equipment fault prediction model refers to the network model used to output the current fault information of pumped storage equipment, such as neural network model, deep learning model, etc.

[0120] The second equipment failure prediction model refers to a network model used to output future failure information of pumped storage equipment, such as a neural network model or a deep learning model.

[0121] For example, the server uses a feature extraction model to extract features from the current device data and the current device state, obtaining feature vectors corresponding to the current device data and the current device state. Then, the feature vectors corresponding to the current device data and the current device state are concatenated to obtain a concatenated feature vector. Next, the concatenated feature vector is input into a first device fault prediction model, which performs fault prediction processing on the concatenated feature vector to obtain the probability of the pumped storage equipment under various first preset fault information. The first preset fault information with the highest probability is taken as the current fault information for the pumped storage equipment. Simultaneously, the concatenated feature vector is input into a second device fault prediction model, which performs fault prediction processing on the concatenated feature vector to obtain the probability of the pumped storage equipment under various second preset fault information. The second preset fault information with the highest probability is taken as the future fault information for the pumped storage equipment. Finally, the current fault information and the future fault information are combined to obtain the fault information for the pumped storage equipment.

[0122] In this embodiment, the current fault information of the pumped storage equipment is predicted by the first equipment fault prediction model, and the future fault information of the pumped storage equipment is predicted by the second equipment fault prediction model. The current fault information and the future fault information are fused and processed to obtain the fault information of the pumped storage equipment, which helps to improve the accuracy of determining the fault information of the pumped storage equipment.

[0123] In an exemplary embodiment, step S104 above, which updates the candidate maintenance item list again based on the fault information to obtain the target maintenance item list, specifically includes the following: based on the current fault information of the pumped storage equipment, querying the correspondence between the current fault information and maintenance items to obtain the first maintenance item corresponding to the current fault information; based on the future fault information of the pumped storage equipment, querying the correspondence between the future fault information and maintenance items to obtain the second maintenance item corresponding to the future fault information; adding the first maintenance item and the second maintenance item to the candidate maintenance item list to obtain the target maintenance item list.

[0124] The correspondence between current fault information and maintenance items is used to indicate that different current fault information corresponds to different maintenance items.

[0125] The correspondence between future fault information and maintenance items is used to indicate that different future fault information corresponds to different maintenance items.

[0126] The target maintenance item list includes the first maintenance item, the second maintenance item, and a list of candidate maintenance items.

[0127] For example, the server first retrieves the correspondence between current fault information and maintenance items, as well as the correspondence between future fault information and maintenance items, from the local database. Then, based on the current fault information of the pumped storage equipment, it queries the correspondence between the current fault information and maintenance items to obtain the maintenance items corresponding to the current fault information of the pumped storage equipment, which are designated as the first maintenance item. Simultaneously, based on the future fault information of the pumped storage equipment, it queries the correspondence between the future fault information and maintenance items to obtain the maintenance items corresponding to the future fault information of the pumped storage equipment, which are designated as the second maintenance item. Finally, the first and second maintenance items are added to the candidate maintenance item list to obtain the processed candidate maintenance item list, which serves as the target maintenance item list. That is, the target maintenance item list includes the first maintenance item, the second maintenance item, and the candidate maintenance item list.

[0128] In this embodiment, the candidate maintenance item list is updated again based on the current and future fault information of the pumped storage equipment to obtain the target maintenance item list. This makes the final target maintenance item list more accurate, thereby improving the accuracy of the target maintenance item list determination.

[0129] In an exemplary embodiment, as shown in Figure 3, another method for generating maintenance strategies for pumped storage equipment is provided. Taking the application of this method to a server as an example, the method includes the following steps S301 to S312. Wherein:

[0130] Step S301: Obtain the equipment maintenance manual for the pumped storage equipment.

[0131] Step S302: Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment.

[0132] Step S303: Input the current equipment data of the pumped storage equipment into the pre-trained importance prediction model to obtain the importance prediction probability of the current equipment data under each preset importance level; from each preset importance level, select the preset importance level with the highest importance prediction probability as the importance level of the current equipment data.

[0133] Step S304: Select device data with an importance greater than a preset importance from the current device data and use it as key device data; input the key device data into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key device data.

[0134] Step S305: The multiple feature vectors of the key equipment data are fused to obtain a fused feature vector, which is used as the target feature vector.

[0135] Step S306: Input the current equipment status, historical maintenance data of pumped storage equipment, and initial maintenance item list into the pre-trained maintenance item prediction model to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list.

[0136] Step S307: From all maintenance items, identify maintenance items with a predicted maintenance probability greater than the second preset probability as target maintenance items; delete maintenance items other than target maintenance items from the initial maintenance item list to obtain a candidate maintenance item list.

[0137] Step S308: Perform feature extraction processing on the current device data and the current device state to obtain the feature vector corresponding to the current device data and the feature vector corresponding to the current device state; concatenate the feature vector corresponding to the current device data and the feature vector corresponding to the current device state to obtain the concatenated feature vector.

[0138] Step S309: Input the spliced ​​feature vector into the first equipment fault prediction model to obtain the current fault information of the pumped storage equipment, and input the spliced ​​feature vector into the second equipment fault prediction model to obtain the future fault information of the pumped storage equipment; fuse the current fault information and the future fault information to obtain the fault information of the pumped storage equipment.

[0139] Step S310: Based on the current fault information of the pumped storage equipment, query the correspondence between the current fault information and the maintenance items to obtain the first maintenance item corresponding to the current fault information; based on the future fault information of the pumped storage equipment, query the correspondence between the future fault information and the maintenance items to obtain the second maintenance item corresponding to the future fault information.

[0140] Step S311: Add the first maintenance item and the second maintenance item to the candidate maintenance item list to obtain the target maintenance item list.

[0141] Step S312: Generate a maintenance strategy for pumped storage equipment based on the target maintenance item list.

[0142] In the above-mentioned method for generating maintenance strategies for pumped storage equipment, the maintenance strategy is generated by comprehensively considering the equipment maintenance manual, current equipment status, historical maintenance data, current fault information, and future fault information of the pumped storage equipment. This is used to determine the target maintenance item list, which helps to improve the accuracy of the target maintenance item list. As a result, the maintenance strategy generated based on the target maintenance item list is more accurate, thereby improving the accuracy of the maintenance strategy determination. At the same time, it avoids the shortcomings of maintenance strategies formulated based on human experience, which have limitations and result in a low accuracy of maintenance strategy determination.

[0143] In an exemplary embodiment, to more clearly illustrate the method for generating maintenance strategies for pumped storage equipment provided in this application, the following specific embodiment will be used to describe the method in detail. In an exemplary embodiment, as shown in FIG4, this application also proposes a method for determining maintenance strategies for pumped storage equipment. Based on the standard maintenance test configuration and equipment condition evaluation results of the pumped storage equipment, the method analyzes and automatically calculates the maintenance plan, generating a list of maintenance plan items; specifically including the following:

[0144] (1) Maintenance project plan for newly built pumped storage equipment.

[0145] (2) Obtain historical maintenance data of pumped storage equipment from the power grid management platform, and summarize maintenance project information of pumped storage equipment based on historical maintenance data and one type of equipment per manual.

[0146] (3) Obtain the health and importance evaluation results of the pumped storage equipment, and determine the control level of the pumped storage equipment according to the equipment risk matrix based on the health and importance evaluation results. The control levels of the pumped storage equipment are divided into "Level I, Level II, Level III and Level IV" from high to low. Based on the RCM (Reliability Centered Maintenance) maintenance strategy library and the control level, determine the maintenance decision results of the pumped storage equipment. The maintenance decision results should include the maintenance items, maintenance time and maintenance cycle of a single unit. Adjust the regular maintenance plan according to the maintenance decision results. The principle of differentiation of maintenance strategies for different control levels is as follows: (a) Level I control equipment: shorten the maintenance cycle and carry out maintenance in a timely manner (within 1 year). (b) Level II and III control equipment: carry out maintenance according to the maintenance cycle determined by RCM analysis, that is, the baseline cycle. (c) Level IV control equipment: for equipment with normal status evaluation, the maintenance cycle can be appropriately extended according to the equipment failure pattern, and the extension time shall not exceed its baseline cycle.

[0147] For example, based on historical maintenance data and equipment status evaluation results, the following strategy analysis algorithm is executed: (1) Regular maintenance items, maintenance upon expiration: For items where “maintenance strategy” = “regular maintenance”, maintenance is performed upon expiration; (2) Condition-based maintenance items, healthy equipment is exempt from maintenance: For items where “maintenance strategy” = “condition-based maintenance”, maintenance is determined based on the equipment status evaluation results, but the maximum cycle shall not be exceeded; (3) Defect elimination and technical improvement countermeasures, additional maintenance as needed: Maintenance items for equipment defect elimination, technical improvement or functional optimization are supplemented manually or by a third-party system.

[0148] The maintenance strategy formulation may also include the following steps: ① Based on component nodes, users configure standard maintenance lists; ② Users associate one or more completed equipment status evaluation tasks, complete maintenance strategy analysis tasks according to system guidance and calculation, and intelligently output maintenance plans; wherein, evaluation standards, evaluation tasks, and maintenance strategies are linked together by components; ③ (Power grid management platform) links defect, countermeasure, and other information to complete maintenance plan formulation and automatically generate plan forms, etc.

[0149] (4) Add a list of maintenance projects based on the needs of defect elimination, countermeasures, technical upgrades, etc., and support importing from the power grid management platform.

[0150] (5) Generate a list of maintenance items for pumped storage equipment based on the increased maintenance plan.

[0151] The above embodiments help improve the accuracy of determining equipment maintenance strategies, while avoiding the limitations of maintenance strategies developed based on human experience, which result in a low accuracy of determining maintenance strategies.

[0152] It should be understood that although the steps in the flowcharts of the embodiments described above 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 embodiments described above 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.

[0153] Based on the same inventive concept, this application also provides a pumped storage equipment maintenance strategy generation device for implementing the above-mentioned pumped storage equipment maintenance strategy generation method. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the pumped storage equipment maintenance strategy generation device provided below can be found in the limitations of the pumped storage equipment maintenance strategy generation method described above, and will not be repeated here.

[0154] In an exemplary embodiment, as shown in FIG5, a pumped storage equipment maintenance strategy generation device is provided, comprising: an information acquisition module 510, a list construction module 520, a first update module 530, a second update module 540, and a strategy generation module 550, wherein:

[0155] Information acquisition module 510 is used to acquire the equipment maintenance manual of pumped storage equipment.

[0156] The list building module 520 is used to build an initial maintenance item list for pumped storage equipment based on the equipment maintenance manual.

[0157] The first update module 530 is used to input the current equipment data of the pumped storage equipment into the pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain the candidate maintenance item list.

[0158] The second update module 540 is used to input the current equipment data and current equipment status of the pumped storage equipment into the pre-trained equipment fault prediction model to obtain the fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information.

[0159] The strategy generation module 550 is used to generate maintenance strategies for pumped storage equipment based on the target maintenance item list.

[0160] In one exemplary embodiment, the pre-trained device state analysis model includes multiple sub-device state analysis models;

[0161] The first update module 530 is also used to perform feature extraction processing on the current equipment data of the pumped storage equipment to obtain a target feature vector; input the target feature vector into multiple sub-equipment state analysis models respectively to obtain the equipment state of the pumped storage equipment output by each sub-equipment state analysis model, and the state prediction probability corresponding to the equipment state; from each sub-equipment state analysis model, select the target sub-equipment state analysis model whose output state prediction probability of the pumped storage equipment is greater than a first preset probability; from the equipment states of the pumped storage equipment output by each target sub-equipment state analysis model, select the equipment state that appears most frequently as the current equipment state of the pumped storage equipment.

[0162] In an exemplary embodiment, the first update module 530 is further configured to input the current equipment data of the pumped storage equipment into a pre-trained importance prediction model to obtain the importance prediction probability of the current equipment data under various preset importance levels; select the preset importance with the highest importance prediction probability from the various preset importance levels as the importance of the current equipment data; select equipment data with an importance greater than the preset importance from the current equipment data as key equipment data; input the key equipment data into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key equipment data; and fuse the multiple feature vectors of the key equipment data to obtain a fused feature vector as the target feature vector.

[0163] In an exemplary embodiment, the first update module 530 is further configured to input the current equipment status, historical maintenance data of the pumped storage equipment, and the initial maintenance item list into a pre-trained maintenance item prediction model to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list; determine the maintenance items whose predicted maintenance probability is greater than a second preset probability from each maintenance item as target maintenance items; and delete the maintenance items in the initial maintenance item list other than the target maintenance items to obtain a candidate maintenance item list.

[0164] In one exemplary embodiment, the pre-trained device failure prediction model includes a first device failure prediction model and a second device failure prediction model;

[0165] The second update module 540 is also used to perform feature extraction processing on the current device data and the current device status to obtain the feature vector corresponding to the current device data and the feature vector corresponding to the current device status; to concatenate the feature vector corresponding to the current device data and the feature vector corresponding to the current device status to obtain the concatenated feature vector; to input the concatenated feature vector into the first device fault prediction model to obtain the current fault information of the pumped storage equipment, and to input the concatenated feature vector into the second device fault prediction model to obtain the future fault information of the pumped storage equipment; and to fuse the current fault information and the future fault information to obtain the fault information of the pumped storage equipment.

[0166] In an exemplary embodiment, the second update module 540 is further configured to: query the correspondence between the current fault information and maintenance items based on the current fault information of the pumped storage equipment to obtain the first maintenance item corresponding to the current fault information; query the correspondence between the future fault information and maintenance items based on the future fault information of the pumped storage equipment to obtain the second maintenance item corresponding to the future fault information; and add the first maintenance item and the second maintenance item to the candidate maintenance item list to obtain the target maintenance item list.

[0167] Each module in the aforementioned pumped storage equipment maintenance strategy generation device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the operations corresponding to each module.

[0168] In an exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram is shown in Figure 6. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data such as equipment maintenance manuals and troubleshooting strategies. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for generating a maintenance strategy for pumped storage equipment.

[0169] Those skilled in the art will understand that the structure shown in Figure 6 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.

[0170] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0171] In one exemplary embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above-described method embodiments.

[0172] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0173] 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, and when executed, it 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 memory 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, artificial intelligence (AI) processors, etc., and are not limited to these.

[0174] 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 application.

[0175] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent 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 generating maintenance strategies for pumped storage equipment, characterized in that, The method includes: Obtain the equipment maintenance manual for the pumped storage equipment; Based on the equipment maintenance manual, construct an initial maintenance item list for the pumped storage equipment; The current equipment data of the pumped storage equipment is input into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment. Based on the current equipment status and the historical maintenance data of the pumped storage equipment, the initial maintenance item list is updated to obtain a candidate maintenance item list. The current equipment data and current equipment status of the pumped storage equipment are input into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information. Based on the target maintenance item list, a maintenance strategy for the pumped storage equipment is generated.

2. The method according to claim 1, characterized in that, The pre-trained device state analysis model includes multiple sub-device state analysis models; The step of inputting the current equipment data of the pumped storage equipment into a pre-trained equipment state analysis model to obtain the current equipment state of the pumped storage equipment includes: The current equipment data of the pumped storage device is processed by feature extraction to obtain the target feature vector; The target feature vector is input into the multiple sub-equipment state analysis models respectively to obtain the equipment state of the pumped storage equipment output by each sub-equipment state analysis model, and the state prediction probability corresponding to the equipment state. From each sub-device state analysis model, select the target sub-device state analysis model whose output device state prediction probability is greater than the first preset probability. From the equipment states of the pumped storage equipment output by the status analysis model of each target sub-equipment, the equipment state that appears most frequently is selected as the current equipment state of the pumped storage equipment.

3. The method according to claim 2, characterized in that, The step of performing feature extraction processing on the current equipment data of the pumped storage equipment to obtain the target feature vector includes: The current equipment data of the pumped storage device is input into a pre-trained importance prediction model to obtain the importance prediction probability of the current equipment data under various preset importance levels. From the various preset importance values, the preset importance value with the highest predicted probability is selected as the importance value of the current device data; From the current device data, select device data with an importance greater than a preset importance level and designate them as key device data; The key equipment data is input into multiple pre-trained feature extraction models to obtain multiple feature vectors of the key equipment data; Multiple feature vectors of the key equipment data are fused to obtain a fused feature vector, which is used as the target feature vector.

4. The method according to claim 1, characterized in that, The initial maintenance item list is updated based on the current equipment status and the historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list, including: The current equipment status, the historical maintenance data of the pumped storage equipment, and the initial maintenance item list are input into a pre-trained maintenance item prediction model to obtain the predicted maintenance probability of each maintenance item in the initial maintenance item list. From the various maintenance items, the maintenance items with a predicted maintenance probability greater than the second preset probability are identified as target maintenance items. The maintenance items other than the target maintenance item are deleted from the initial maintenance item list to obtain the candidate maintenance item list.

5. The method according to claim 1, characterized in that, The pre-trained equipment failure prediction model includes a first equipment failure prediction model and a second equipment failure prediction model. The step of inputting the current equipment data and current equipment status of the pumped storage equipment into a pre-trained equipment fault prediction model to obtain fault information of the pumped storage equipment includes: The current device data and the current device state are subjected to feature extraction processing to obtain the feature vector corresponding to the current device data and the feature vector corresponding to the current device state; The feature vector corresponding to the current device data and the feature vector corresponding to the current device state are concatenated to obtain a concatenated feature vector. The spliced ​​feature vector is input into the first equipment fault prediction model to obtain the current fault information of the pumped storage equipment, and the spliced ​​feature vector is input into the second equipment fault prediction model to obtain the future fault information of the pumped storage equipment. The current fault information and the future fault information are fused together to obtain the fault information of the pumped storage equipment.

6. The method according to any one of claims 1 to 5, characterized in that, The step involves updating the candidate maintenance item list based on the fault information to obtain the target maintenance item list, including: Based on the current fault information of the pumped storage equipment, the correspondence between the current fault information and the maintenance items is queried to obtain the first maintenance item corresponding to the current fault information. Based on the future fault information of the pumped storage equipment, the correspondence between the future fault information and the maintenance items is queried to obtain the second maintenance item corresponding to the future fault information. Add the first maintenance item and the second maintenance item to the candidate maintenance item list to obtain the target maintenance item list.

7. A device for generating maintenance strategies for pumped storage equipment, characterized in that, The device includes: The information acquisition module is used to acquire the equipment maintenance manual of the pumped storage equipment; The list construction module is used to construct an initial maintenance item list for the pumped storage equipment based on the equipment maintenance manual. The first update module is used to input the current equipment data of the pumped storage equipment into a pre-trained equipment status analysis model to obtain the current equipment status of the pumped storage equipment, and update the initial maintenance item list based on the current equipment status and the historical maintenance data of the pumped storage equipment to obtain a candidate maintenance item list. The second update module is used to input the current equipment data and current equipment status of the pumped storage equipment into a pre-trained equipment fault prediction model to obtain the fault information of the pumped storage equipment. Based on the fault information, the candidate maintenance item list is updated again to obtain the target maintenance item list. The fault information includes current fault information and future fault information. The strategy generation module is used to generate a maintenance strategy for the pumped storage equipment based on the target maintenance item list.

8. 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 6.

9. 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 6.

10. A computer program product, comprising a computer program, 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 6.

Images