Optical module operation and maintenance method and device and electronic equipment

Abstract

The embodiment of the invention provides an optical module operation and maintenance method and device and electronic equipment, and relates to the technical field of communication. According to the method, target optical module information and operation and maintenance information are received in response to a performance optimization request or a fault diagnosis request of a user for a target optical module, and it is ensured that differentiation processing can be carried out according to an actual operation and maintenance scene; and determining a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information, thereby realizing accurate mapping between the specific equipment instance and the structured knowledge system. And taking the target optical module node as a starting point, executing graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, and generating at least one group of operation and maintenance schemes. According to the method, search and expansion are carried out from a single node along a semantic relation path by utilizing the association reasoning capability of a knowledge graph, personalized and interpretable operation and maintenance suggestion generation aiming at a specific equipment state and an operation and maintenance target is realized, and the response precision and decision transparency in the operation and maintenance process of the optical module are enhanced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of communication, in particular to an optical module operation and maintenance method, device and electronic equipment. BACKGROUND

[0002] In the research and development, debugging and production practice of optical modules, the following technical bottlenecks are generally faced: first, the parameter configuration and fault diagnosis process excessively relies on the personal experience accumulation of engineers, resulting in low overall work efficiency. Second, various types of data related to optical modules (including performance monitoring data, configuration parameters, alarm information and production test data, etc.) are scattered and stored in multiple independent systems, forming a data island phenomenon, and there is a lack of effective data association mechanism between systems.

[0003] Thirdly, the existing parameter optimization method is mainly based on fixed threshold or simple rules, which is difficult to adapt to complex and variable actual working environment, and does not fully consider the coupling influence relationship between key parameters such as laser bias current, extinction ratio and receiving sensitivity. Finally, when the system performance declines or fails, the fault positioning process is tedious and inefficient, and it is difficult to quickly and accurately identify and locate the root cause of the fault. The above technical bottlenecks seriously restrict the work efficiency of optical modules in the debugging and production and fault analysis links. SUMMARY

[0004] Therefore, the purpose of the present application is to provide an optical module operation and maintenance method, device and electronic equipment, which can enhance the response accuracy and decision transparency in the operation and maintenance process of optical modules.

[0005] In order to achieve the above purpose, the technical solutions adopted by the embodiments of the present application are as follows: In a first aspect, the present application provides an optical module operation and maintenance method, which comprises: In response to a user's operation and maintenance request for a target optical module, receiving target optical module information and operation and maintenance information; the operation and maintenance request includes a performance tuning request or a fault diagnosis request; Determining a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information; Taking the target optical module node as the starting point, performing a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, and generating at least one set of operation and maintenance scheme; the operation and maintenance scheme is used to echo to the user for viewing.

[0006] In an optional implementation, the operation and maintenance information includes target performance indicators and current environment data, and the operation and maintenance scheme is generated by performing a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, which comprises: If the operation and maintenance request is a performance tuning request, determining target performance parameters and target performance values according to the target performance indicators; search target adjustable parameters that affect the target performance parameter based on the target performance parameter in the optical module knowledge graph, and obtain constraint boundaries of each target adjustable parameter; search historical optimal cases with similarity exceeding a preset first threshold based on the current environment data and the target optical module information in the optical module knowledge graph, and determine an optimization starting value of each target adjustable parameter according to a search result of the historical optimal cases; take the target performance value as an optimization target, take the optimization starting value of each target adjustable parameter as a starting point, and perform multi-objective collaborative optimization search on each target adjustable parameter within the constraint boundaries of all target adjustable parameters to obtain at least one parameter optimization combination; the parameter optimization combination includes a recommended optimization value and an expected performance value of the target adjustable parameter; calculate an optimization confidence of each parameter optimization combination; one parameter optimization combination and the corresponding optimization confidence constitute a set of operation and maintenance schemes.

[0007] In an optional implementation, the target optical module information includes a plurality of attribute information of the target optical module; the search of the historical optimal cases with similarity exceeding the preset first threshold based on the current environment data and the target optical module information in the optical module knowledge graph includes: find a historical case that is completely same as the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data in the optical module knowledge graph; if the historical case that is completely same as the target optical module information and has the first similarity exceeding the preset first threshold with the current environment data is found, the historical optimal case is obtained; if the historical case that is completely same as the target optical module information and has the first similarity exceeding the preset first threshold with the current environment data is not found, candidate information is generated according to a priority of each attribute information; find a historical case that is completely same as the candidate information and has a first similarity exceeding a preset first threshold with the current environment data in the optical module knowledge graph according to the candidate information; if the historical case that is completely same as the candidate information and has the first similarity exceeding the preset first threshold with the current environment data is found, the historical optimal case is obtained.

[0008] In an optional implementation, the determination of the optimization starting value of each target adjustable parameter according to the search result of the historical optimal cases includes: if the historical optimal case exists, a value of a target adjustable parameter in the historical optimal case is determined as the optimization starting value of the target adjustable parameter; If the historical optimal case does not exist, a default value of the target adjustable parameter is determined as an optimization starting value of the target adjustable parameter.

[0009] In an optional implementation, the calculation of the optimization confidence of each parameter optimization combination comprises: The second similarity of each parameter optimization combination is calculated according to the target optical module information, the target performance index, the current environment data and the historical optimal case; The third similarity of each parameter optimization combination is calculated according to the expected performance value and the target performance value of each parameter optimization combination; The safety margin ratio of each parameter optimization combination is calculated according to the recommended optimization value and the constraint boundary of each target adjustable parameter in each parameter optimization combination; The optimization confidence of each parameter optimization combination is determined according to the second similarity, the third similarity and the safety margin ratio of each parameter optimization combination.

[0010] In an optional implementation, the operation and maintenance information comprises current fault information, and the graph traversal operation is performed in the optical module knowledge graph based on the operation and maintenance information to generate at least one operation and maintenance scheme, which comprises: If the operation and maintenance request is a fault diagnosis request, the current fault information is standardized to obtain a standard fault phenomenon; A fault root cause affecting the standard fault phenomenon is searched in the optical module knowledge graph to obtain at least one candidate root cause; A solution corresponding to each candidate root cause is searched in the optical module knowledge graph according to each candidate root cause; A historical case most similar to the fault root cause and the candidate root cause is searched in the optical module knowledge graph according to each candidate root cause to obtain a root cause similar case corresponding to each candidate root cause; A diagnosis confidence of each candidate root cause is calculated according to the standard fault phenomenon, each candidate root cause and the corresponding root cause similar case; each candidate root cause, the corresponding solution and the corresponding diagnosis confidence constitute an operation and maintenance scheme.

[0011] In an optional implementation, the calculation of the diagnosis confidence of each candidate root cause according to the standard fault phenomenon, each candidate root cause and the corresponding root cause similar case comprises: A fault type caused by the candidate root cause is searched in the optical module knowledge graph, and a prior probability of the candidate root cause is determined according to a number of times of occurrence of the fault type caused by the candidate root cause and a total number of occurrences of the fault type; According to the similarity analysis of the fault phenomena of the root cause similar cases corresponding to the standard fault phenomenon and each candidate root cause, a support degree of each candidate root cause is obtained; According to the prior probability of the candidate root cause, a preset case weight, and the support degree, a diagnosis confidence of each candidate root cause is determined.

[0012] In an optional implementation, the optical module knowledge graph is constructed by the following manner: Obtaining multi-modal engineering data corresponding to the optical module; the multi-modal engineering data includes research and development records, generated data, test data, user feedback, and maintenance reports; According to a pre-set rule, extracting entities, entity attributes, and association relationships between entities from the multi-modal engineering data; the entities include optical module attributes, performance parameters, adjustable parameters, fault phenomena, fault root causes, solutions, environmental factors, and historical cases; Storing the extracted entities, entity attributes of the entities, and association relationships between the entities into a graph database to obtain the optical module knowledge graph.

[0013] In a second aspect, the present application provides an optical module operation and maintenance device, the device comprising: A processing module, configured to receive target optical module information and operation and maintenance information in response to a user's operation and maintenance request for a target optical module; the operation and maintenance request includes a performance tuning request or a fault diagnosis request; An operation and maintenance module, configured to determine a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information; and perform a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, taking the target optical module node as a starting point, to generate at least one set of operation and maintenance solutions; the operation and maintenance solutions are used to be echoed to the user for viewing.

[0014] In a third aspect, the present application provides an electronic device comprising a processor and a memory, the memory storing a computer program executable by the processor, and the processor being capable of executing the computer program to implement the optical module operation and maintenance method of any of the preceding embodiments.

[0015] Compared with the prior art, the optical module operation and maintenance method, device and electronic equipment provided by the embodiments of the present application ensure that the actual operation and maintenance scene can be processed differently by responding to the performance optimization request or fault diagnosis request of the user to the target optical module, receiving the target optical module information and operation and maintenance information, determining the target optical module node in the pre-constructed optical module knowledge graph according to the target optical module information, and realizing the accurate mapping between the specific device instance and the structured knowledge system. Taking the target optical module node as the starting point, performing a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, generating at least one set of operation and maintenance scheme, and using the operation and maintenance scheme to echo to the user for viewing. Using the correlation reasoning ability of the knowledge graph, searching and expanding along the semantic relationship path from a single node, thereby breaking the data island, overcoming the experience dependence, realizing the personalized and explainable operation and maintenance suggestion generation for the specific device state and operation and maintenance target, enhancing the response accuracy and decision transparency in the optical module operation and maintenance process, and improving the efficiency and performance of the optical module optimization and diagnosis.

[0016] In order to make the above-mentioned objects, features and advantages of the present application more apparent and easy to understand, the following preferred embodiments are specifically described below, and the accompanying drawings are described in detail as follows. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0018] Figure 1 A flowchart of the optical module operation and maintenance method provided by the embodiments of the present application is shown.

[0019] Figure 2 Another flowchart of the optical module operation and maintenance method provided by the embodiments of the present application is shown.

[0020] Figure 3 A block diagram of the optical module operation and maintenance device provided by the embodiments of the present application is shown.

[0021] Figure 4 A block diagram of the electronic device provided by the embodiments of the present application is shown. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various different configurations.

[0023] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed present application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0024] It should be noted that the relational terms such as "first" and "second" and the like are used only to distinguish one entity or operation from another entity or operation, without necessarily requiring or implying any such actual relationship or order between such entities or operations. Moreover, the terms "comprising", "including" or any other variant thereof are intended to cover non-exclusive inclusion, so that a process, method, article or apparatus that includes a list of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus including the element.

[0025] The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0026] Please refer to Figure 1 , Figure 1 A flowchart of a method for operating and maintaining an optical module is shown. The method comprises the following steps: Step S10, in response to a user's request for operating and maintaining a target optical module, receiving target optical module information and operating and maintaining information; the request for operating and maintaining includes a request for performance tuning or a request for fault diagnosis.

[0027] In the embodiments of the present application, when a user initiates a request for operating and maintaining a specific optical module (i.e. a target optical module), the target optical module information and the operating and maintaining information are received through an interface. The target optical module information is used to uniquely identify a physical entity (i.e. an optical module), and the operating and maintaining information is used to represent the user's intention for operating and maintaining the target optical module.

[0028] Step S20, determining a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information.

[0029] In the embodiment of the present application, the optical module knowledge graph refers to a graph database that stores various entities in the field of optical modules and their mutual relationships, including but not limited to the attributes of different types of optical modules, and the performance parameters, adjustable parameters, fault phenomena, fault causes, solutions, environmental factors and historical cases associated with the optical module attributes.

[0030] The received target optical module information is used as a search key to perform accurate matching in the optical module knowledge graph, and the corresponding graph node is located to obtain the target optical module node. The target optical module node is used as the starting anchor point for subsequent graph traversal operations, which ensures that all reasoning processes are carried out around the target optical module, and enhances the relevance and pertinence of the decision results.

[0031] Step S30: Based on the operation and maintenance information, a graph traversal operation is performed in the optical module knowledge graph with the target optical module node as the starting point, and at least one set of operation and maintenance solutions is generated; the operation and maintenance solutions are used for echoing to the user for viewing.

[0032] In the embodiment of the present application, by introducing the path exploration mechanism in the graph structure, the operation and maintenance problem is converted into a semantic reasoning task in the knowledge space. When the operation and maintenance request is a performance tuning request, starting from the target optical module node, along the predefined correlation relationships such as "has parameter" and "mutual influence", a plurality of adjustable parameters associated with the performance of the target optical module are tracked, and one or more potential influence paths are constructed.

[0033] When the operation and maintenance request is a fault diagnosis request, the observed fault phenomenon is taken as the target endpoint, and various potential cause paths that may lead to the fault phenomenon are traced in reverse. The graph traversal process performs depth-first or breadth-first search according to the semantic relationships established in the graph, and finally aggregates several feasible technical paths, each of which corresponds to a possible solution strategy. The output set formed by these technical paths is a set of operation and maintenance solutions, which can be used for subsequent display to the user for reference.

[0034] In summary, the optical module operation and maintenance method provided in this embodiment of the invention receives target optical module information and operation and maintenance information in response to user requests for performance tuning or fault diagnosis of the target optical module, ensuring differentiated processing based on actual operation and maintenance scenarios. Based on the target optical module information, the method determines the target optical module node in a pre-built optical module knowledge graph, achieving precise mapping between specific device instances and a structured knowledge system. Starting from the target optical module node, a graph traversal operation is performed in the optical module knowledge graph based on the operation and maintenance information to generate at least one set of operation and maintenance solutions, which are then displayed to the user. Utilizing the associative reasoning capabilities of the knowledge graph, the method searches and expands along semantic relationship paths starting from a single node, thereby breaking down data silos and overcoming reliance on experience. This enables the generation of personalized and interpretable operation and maintenance suggestions for specific device states and operation and maintenance goals, enhancing the response accuracy and decision-making transparency in the optical module operation and maintenance process, and improving the efficiency and performance of optical module tuning and diagnosis.

[0035] Optionally, the operational information includes target performance metrics and current environment data. The following provides a possible implementation method for generating performance tuning requests. Figure 1 The sub-steps of step S30 may include: Step S300: If the operation and maintenance request is a performance tuning request, determine the target performance parameters and target performance values ​​based on the target performance indicators.

[0036] In this embodiment of the invention, when a performance tuning request initiated by a user is received, the system first receives operation and maintenance information, including target optical module information, target performance metrics, and current environmental data, as input conditions. The target performance metrics are used to define the desired technical effect, such as optimizing the bit error rate (BER). The current environmental data includes external variables that may affect module performance, such as temperature, humidity, and power supply voltage in the actual deployment environment.

[0037] If the user inputs a specific performance target (e.g., a bit error rate lower than...), If the target performance parameter is set to the bit error rate, then the target performance value is set to... If the user inputs an abstract performance target (e.g., optimizing BER), then the performance parameters in the optical module knowledge graph are searched, and the abstract target performance metric is converted into a performance parameter that conforms to the optical module knowledge graph. In other words, the target performance parameter is set to the bit error rate, and the target performance value is set to the default performance experience value.

[0038] It should be understood that the target performance parameter refers to a key performance characterization quantity that is explicitly defined and modeled in the optical module knowledge graph, such as a bit error rate, an extinction ratio (ER), and a receiving sensitivity. The bit error rate is a key performance indicator for measuring the transmission quality of the optical module. The extinction ratio is a parameter for measuring the modulation instruction and is one of the important targets for tuning. The receiving sensitivity is the minimum optical power that can be correctly detected by the receiver of the optical module, and is a key parameter for measuring the receiving performance.

[0039] In step S310, the target adjustable parameters that affect the target performance parameter are searched in the optical module knowledge graph based on the target performance parameter, and the constraint boundary of each target adjustable parameter is obtained.

[0040] In the embodiment of the present application, starting from the target performance parameter, all adjustable parameter nodes that have a “mutual influence” relationship with the target performance parameter are traversed in reverse in the optical module knowledge graph to form a localized adjustable parameter sub-set (i.e., the target adjustable parameter). For example, a laser bias current (directly affecting the working state and service life of the laser of the optical module) and a modulation current (directly related to the optical power and extinction ratio of the optical module). The laser includes but is not limited to a direct modulation laser and an electro-absorption modulation laser.

[0041] It should be noted that the target adjustable parameters that affect the target performance parameter are not searched in the optical module knowledge graph without limitation, but a limited depth-first search (DFS) or breadth-first search (BFS) algorithm is used to set the maximum traversal depth to avoid excessive consumption of computing resources and the risk of combinatorial explosion. The target adjustable parameters generated in this way not only limit the decision space of this optimization, but also convert the topological structure in the optical module knowledge graph into the association relationship between the target adjustable parameters in real time.

[0042] Further, in order to improve the scientificity and rationality of the optimization process, the constraint boundary of each target adjustable parameter is obtained from the optical module knowledge graph, including but not limited to the hardware safety range, the manufacturer's recommended working interval, and the forbidden area identified in the historical failure record, which constitutes a hard limit condition that cannot be exceeded in the optimization process.

[0043] In step S320, a historical optimal case with a similarity exceeding a preset first threshold is searched in the optical module knowledge graph based on the current environment data and the target optical module information, and the tuning starting value of each target adjustable parameter is determined according to the search result of the historical optimal case.

[0044] In the embodiment of the present application, the historical optimal cases similar to the current environment data and the target optical module information are searched in the optical module knowledge graph, and the tuning starting value of each target adjustable parameter is set according to the search result of the historical optimal cases.

[0045] In step S330, the tuning starting value of each target adjustable parameter is taken as a starting point, and a multi-objective collaborative optimization search is performed on each target adjustable parameter within the constraint boundary of all target adjustable parameters, taking the target performance value as an optimization target, to obtain at least one parameter tuning combination; the parameter tuning combination includes a recommended tuning value of the target adjustable parameter and an expected performance value.

[0046] In the embodiment of the present application, the tuning starting value of each target adjustable parameter is taken as a starting point, and a multi-objective collaborative optimization search is performed under the condition of meeting the constraint boundary of all target adjustable parameters, taking the target performance value as an optimization target, to obtain at least one parameter tuning combination. The multi-objective optimization algorithm includes but is not limited to genetic algorithm, particle swarm optimization, and gradient-based numerical optimization algorithm, aiming to find one or more parameter tuning combinations, so that the output expected performance value is as close as possible to or even better than the target performance value, while considering the coordination between multiple performance indicators to prevent the degradation of other key performance due to the aggressive adjustment of a single parameter.

[0047] It is worth noting that after generating each parameter tuning combination, conflict detection needs to be performed on the parameter tuning combination to check whether the combination violates the pre-defined “taboo rule” in the knowledge graph, and the parameter tuning combination conflicting with the taboo rule is deleted. The taboo rule is summarized by expert experience.

[0048] It should be noted that in the multi-objective collaborative optimization search process, in order to explore more abundant parameter tuning combinations, the single target performance value can be expanded to a target performance optimization range to expand the search space, that is, the target performance value is floated up and down according to a preset adjustment ratio.

[0049] For example, if the target performance value is , the preset adjustment ratio is 10%, the target performance value reduced by the preset adjustment ratio (10%) is taken as the lower limit of the target performance optimization range (i.e. ), and the target performance value increased by the preset adjustment ratio (10%) is taken as the upper limit of the target performance optimization range (i.e. ), forming an optimization task with the interval as the target. By introducing the performance tolerance range, more parameter configuration schemes with potential advantages can be found under the premise of meeting the performance requirements, improving the optimization flexibility and robustness.

[0050] Further, in order to improve the efficiency and rationality of the multi-objective collaborative optimization search, an influence weight calculation mechanism driven by domain knowledge is introduced to accurately reflect the relative importance of different adjustable parameters in the overall performance regulation. This mechanism provides prior guidance for the optimization algorithm by fusing expert experience and graph structure information, thereby reducing the invalid search space and accelerating the convergence speed.

[0051] In specific implementation, the influence weight can be obtained by an analytic hierarchy process (AHP): based on historical data and technical manuals, domain experts compare the importance of each adjustable parameter with each other, construct a judgment matrix, and obtain the weight distribution of each parameter by solving the maximum eigenvector, while performing consistency check to ensure the rationality of the evaluation logic. In addition, the weight can also be assigned based on the network structure characteristics of the graph itself, for example, according to the out-degree of a certain adjustable parameter to judge its influence, the higher the out-degree, the more other performance parameters it will affect, and it should be given higher priority in the optimization process.

[0052] In addition, the expected performance value of each parameter tuning combination is generated based on a performance theoretical curve. The performance theoretical curve is a mathematical expression describing the functional relationship between adjustable parameters and key performance indicators based on a physical model or experimental fitting. According to the recommended tuning value of each target adjustable parameter in the parameter tuning combination, the theoretical performance value that can be achieved when the value of the target adjustable parameter is the recommended tuning value is found in the performance theoretical curve.

[0053] It should be understood that the generation process of the parameter tuning combination can be optimized as follows: taking the target performance optimization range as the optimization target, taking the tuning starting value of each target adjustable parameter as the starting point, and performing multi-objective collaborative optimization search on each target adjustable parameter within the constraint boundary of all target adjustable parameters according to the influence weight to obtain at least one parameter tuning combination. The target performance optimization range is obtained by adjusting the target performance value up and down by a preset adjustment ratio; the expected performance value is generated according to the recommended tuning value of each target adjustable parameter in the parameter tuning combination and the performance theoretical curve.

[0054] Step S340, calculate the tuning confidence of each parameter tuning combination; a set of operation and maintenance schemes is formed by a parameter tuning combination and the corresponding tuning confidence.

[0055] In the embodiment of the present application, the matching degree between the parameter tuning combination and the historical optimal case, the degree of achievement between the expected performance value and the target performance value, and the margin of the recommended tuning value from the constraint boundary are comprehensively evaluated, and then the results of these dimensions are fused to generate a tuning confidence. The tuning confidence is used to represent the possibility of successfully achieving the expected performance improvement in actual application of the parameter tuning combination. Finally, each parameter tuning combination and its corresponding tuning confidence together form a set of independent operation and maintenance schemes for user reference and selection.

[0056] Optionally, the target optical module information includes a plurality of attribute information of the target optical module. As to how to obtain the historical optimal case, a possible implementation is provided as follows. The sub-step of step S320, "searching, based on the current environment data and the target optical module information, a historical optimal case with a similarity exceeding a preset first threshold in the optical module knowledge graph", can include: Step S320-1, searching, in the optical module knowledge graph, a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data.

[0057] Step S320-2, if a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is found, obtaining the historical optimal case.

[0058] In the embodiment of the present application, first, the target optical module information and the current environment data currently obtained are taken as query conditions to search, in the optical module knowledge graph, whether there is a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold between its corresponding environment data and the current environment data.

[0059] The first similarity is used to measure the closeness between the current running environment and the environment of the historical case, and can be determined by calculating the Euclidean distance between the vector of the current environment data and the vector of the environment data of the historical case. The smaller the Euclidean distance is, the higher the first similarity is. If there is a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold in the optical module knowledge graph, the historical case is directly determined as the historical optimal case.

[0060] Step S320-3, if a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is not found, generating candidate information according to the priority of each attribute information.

[0061] Step S320-4, searching, in the optical module knowledge graph, a historical case that is completely identical to the candidate information and has a first similarity exceeding a preset first threshold with the current environment data.

[0062] Step S320-5, if a historical case that is completely identical to the candidate information and has a first similarity exceeding a preset first threshold with the current environment data is found, a historical optimal case is obtained.

[0063] In the embodiment of the present application, if a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold is not found in the optical module knowledge graph, a degradation matching mechanism based on attribute priority is introduced. According to the importance of each attribute information in influencing the performance of the optical module, a candidate information set with a hierarchical structure is generated. The generation of the candidate information follows the attribute priority principle to ensure that the selected replacement case is as close as possible to the actual characteristics of the current target optical module in the key technical dimension.

[0064] For example, in the optical module attributes, features such as the number of channels, transmission rate levels (such as 800G, 600G), etc. usually have a high priority, so after the original matching fails, the requirements for some low-priority attributes can be relaxed in turn, and the historical case of a replacement optical module with similar high-priority attributes is searched for. For example, when no matching case of a 4-channel 800G module is found, the historical optimal case of an 8-channel 800G module can be tried to search for; if there is no 800G matching, it is further extended to the module case of the same channel structure of 600G.

[0065] It can be seen that, by dynamically constructing a hierarchical matching logic based on attribute priority, the embodiment of the present application realizes a smooth transition from strict matching to approximate matching, which not only ensures the ability to directly reuse high-quality historical experience under optimal conditions, but also enhances the robustness and adaptability in the face of new models or sparse data scenarios, thereby expanding the search coverage of effective historical cases and providing reliable support for the reasonable setting of the subsequent tuning starting value.

[0066] Optionally, as for how to obtain the tuning starting value, a possible implementation manner is provided as follows. The sub-step of "determining the tuning starting value of each target adjustable parameter according to the search result of the historical optimal case" in step S320 can include: Step S320-6, if there is a historical optimal case, the value of the target adjustable parameter in the historical optimal case is determined as the tuning starting value of the target adjustable parameter.

[0067] In the embodiment of the present application, after the search for the historical optimal case is completed, if there is a historical optimal case, a group of adjustable parameter configurations that are effective in the then operation and maintenance scene, i.e. the actual values of each target adjustable parameter in the historical optimal case, are extracted from the historical optimal case, and these values are directly used as the tuning starting value of the corresponding target adjustable parameter, which realizes the knowledge transfer of the existing successful experience, so that the current optimization algorithm can fine-tune and iterate on the basis of the existing high-performance configuration, rather than starting from zero to explore the entire parameter space.

[0068] Step S320-7, if there is no historical optimal case, determining a default value of the target adjustable parameter as an optimization starting value of the target adjustable parameter.

[0069] In the embodiment of the present application, when no historical optimal case is found, it indicates that the technical situation of the current target optical module lacks sufficient historical data support. In this case, the experiential input of the historical case is no longer relied on, and the default value of the target adjustable parameter is determined as the optimization starting value of the target adjustable parameter. The default value is usually derived from the factory configuration of the equipment, the recommended settings under standard test conditions, or the industry general specification, which has no personalized adaptation ability, but can ensure the basic functional availability and safety, and ensure that the optimization process will not be interrupted due to data loss.

[0070] It can be seen that the embodiment of the present application dynamically selects the source path of the optimization starting value according to the existence of the historical optimal case, makes full use of the empirical and effective parameter configuration when there is a similar running precedent, and still relies on the standardized configuration to maintain the system running continuity when there is no matching case, so as to ensure that the multi-target collaborative optimization search always has reasonable and feasible initial input.

[0071] Optionally, for how to calculate the optimization confidence of the parameter optimization combination, a possible implementation is provided as follows. The sub-step of step S340 can include: Step S340-1, calculating a second similarity of each parameter optimization combination according to the target optical module information, the target performance index, the current environment data, and the historical optimal case.

[0072] In the embodiment of the present application, the feature vectorization processing is performed according to the target optical module information, the target performance index, and the current environment data, to obtain a current feature vector. The feature vectorization processing is performed according to the optical module attribute in the historical optimal case, the performance parameter and the performance value corresponding to the target performance index, and the environmental factor, to obtain a historical feature vector. The second similarity is calculated according to the current feature vector and the historical feature vector.

[0073] For example, the cosine similarity method is used to calculate the angle cosine value between the two, to measure the similarity between the current operation and maintenance situation and the past successful cases, to obtain a historical case similarity score. The historical case similarity score is normalized to map into the interval [0, 1], as the second similarity (i.e. the historical similarity). The higher the second similarity is, the closer the current optimization task is to the existing high-performance practice, and the higher the optimization confidence is.

[0074] Step S340-2, calculating a third similarity of each parameter optimization combination according to the expected performance value and the target performance value of each parameter optimization combination.

[0075] The smaller the deviation between the expected performance value and the target performance value is, the higher the tuning confidence is in the embodiment of the present application. The calculation formula of the third similarity is:

[0076] wherein, is the third similarity of the parameter tuning combination; The closer to 1, the closer the predicted performance value is to the target requirement. is the expected performance value of the parameter tuning combination; is the target performance value.

[0077] Step S340-3, calculating the safety margin ratio of each parameter tuning combination according to the recommended tuning value of each target adjustable parameter in each parameter tuning combination and the constraint boundary.

[0078] In the embodiment of the present application, it is checked whether each recommended tuning value is within the preset safety range (i.e. the constraint boundary), and the distance between the recommended tuning value and the constraint boundary is further quantified. The farther the recommended tuning value corresponding to the constraint boundary is, the lower the operation risk is, and the higher the tuning confidence is. The calculation formula of the safety margin ratio of the parameter tuning combination is:

[0079] wherein, is the safety margin ratio of the parameter tuning combination; N is the total number of target adjustable parameters in the parameter tuning combination; is the recommended tuning value of the i-th target adjustable parameter; is the nearest boundary value in the safety boundary corresponding to the i-th target adjustable parameter; is the total safety range of the safety boundary corresponding to the i-th target adjustable parameter.

[0080] Step S340-4, determining the tuning confidence of each parameter tuning combination according to the second similarity, the third similarity and the safety margin ratio of each parameter tuning combination.

[0081] In the embodiment of the present application, for each parameter combination, the second similarity, the third similarity and the safety margin ratio of each parameter tuning combination are weighted and fused to generate the tuning confidence of each parameter tuning combination. The calculation formula of the tuning confidence is:

[0082] wherein, C is the tuning confidence of the parameter tuning combination; is the second similarity of the parameter tuning combination; , , is a weight coefficient, and The weight coefficient can be set according to an actual application scenario, and the application is not limited in this regard. For example, if historical cases are rich and reliable, a higher value can be given to .

[0083] Optionally, the operation and maintenance information includes current fault information. For how to generate the operation and maintenance scheme of the fault diagnosis request, a possible implementation is provided as follows. Figure 1 The sub-step of step S30 can include: Step S350, if the operation and maintenance request is a fault diagnosis request, the current fault information is standardized to obtain a standard fault phenomenon.

[0084] In the operation and maintenance process of the optical module, the fault phenomenon is often presented in the form of diversified alarm information. These information can come from different monitoring systems or log records, and their expression methods are not unified and have semantic ambiguity, which will lead to a decrease in accuracy if directly used for root cause analysis. To solve this problem, after receiving the operation and maintenance request containing the current fault information, the application first standardizes the current fault information, aiming to convert it into a standard fault phenomenon defined in the knowledge graph, so as to ensure that the subsequent reasoning process can be accurately performed in the structured semantic space.

[0085] Firstly, accurate matching is performed, that is, the key fields (such as the combination of parameter name and fault action) are extracted from the current fault information, and are compared with the fault phenomena defined in the optical module knowledge graph. If there is a completely consistent name, the completely consistent name is directly determined as the standard fault phenomenon.

[0086] If accurate matching cannot be achieved, fuzzy matching and semantic similarity calculation mechanism are started to cope with the matching difficulties caused by synonymous expressions, spelling differences or different description granularities. The key fields extracted from the current fault information and the description texts of each candidate fault phenomenon in the knowledge graph are respectively converted into word vector or sentence vector representation forms, and then the cosine similarity algorithm is used to calculate the semantic distance between the input alarm vector and each candidate vector.

[0087] Subsequently, the candidate fault phenomenon with the highest similarity score and the similarity score exceeding a preset similarity threshold (for example, 0.8) is selected as the standard fault phenomenon; if the similarity scores of all candidate fault phenomena do not exceed the preset similarity threshold, the fault phenomenon is determined as an "unknown phenomenon", and an artificial intervention process is triggered to ensure the reliability of the diagnosis starting point.

[0088] Step S360, searching for fault root causes affecting the standard fault phenomenon in the optical module knowledge graph to obtain at least one candidate root cause.

[0089] In the embodiment of the present application, the standard fault phenomenon is located in the optical module knowledge graph, and the graph traversal operation is performed starting from the standard fault phenomenon. All upstream nodes connected to the standard fault phenomenon node through the "may cause" or "association" relationship are searched, which are the fault root cause nodes that have causal or statistical association with the fault phenomenon. Thus, a candidate root cause covering multiple possibilities is formed (for example, connector is seriously contaminated). This process realizes the multi-path hypothesis generation from the observed phenomenon to the potential cause, breaking through the limitations of traditional single linear rule judgment.

[0090] In step S370, the solution corresponding to each candidate root cause is searched in the optical module knowledge graph according to the candidate root cause.

[0091] In the embodiment of the present application, for each candidate root cause, the solution corresponding to the candidate root cause is retrieved in the optical module knowledge graph through the "can be solved by" association relationship. For example, the candidate root cause is that the connector is seriously contaminated, and the solution is to clean the fiber connector.

[0092] In step S380, the historical case most similar to the fault root cause is searched in the optical module knowledge graph according to each candidate root cause, to obtain the root cause similar case corresponding to each candidate root cause.

[0093] To further improve the diagnosis accuracy, for each candidate root cause, the historical successful diagnosis case related thereto, i.e., the root cause similar case, is retrieved in the optical module knowledge graph. The root cause similar case records the complete disposal process and verification result in the past same or similar fault situation.

[0094] In step S390, the diagnosis confidence of each candidate root cause is calculated according to the standard fault phenomenon, each candidate root cause, and the corresponding root cause similar case. Each candidate root cause, the corresponding solution, and the corresponding diagnosis confidence constitute a set of operation and maintenance schemes.

[0095] In the embodiment of the present application, the diagnosis confidence of each candidate root cause is obtained by comparing the current fault situation with the fault phenomenon in the historical case to evaluate the matching degree therebetween. Each combination of the candidate root cause and the corresponding diagnosis confidence is regarded as a set of independent operation and maintenance schemes. Multiple such schemes together constitute a complete diagnosis suggestion list, which is sorted according to the diagnosis confidence and then echoed to the user for viewing, so as to realize the explainable, multi-level, and efficient diagnosis support for the complex fault scenario.

[0096] Optionally, for how to calculate the diagnosis confidence of each candidate root cause, a possible implementation is provided as follows. The sub-step of step S390 can include: Step S390-1, find the fault type caused by the candidate root cause in the optical module knowledge graph, and determine the prior probability of the candidate root cause according to the number of times of the fault type caused by the candidate root cause and the total number of times of the fault type.

[0097] In the embodiment of the application, the set of fault types associated with the candidate root cause is found in the optical module knowledge graph, and the prior probability of the candidate root cause is determined according to the ratio of the number of times of the fault phenomenon (i.e. the fault type) caused by the candidate root cause to the total number of times of the fault phenomenon. The prior probability reflects the universality or commonality of the candidate root cause in the historical operation data.

[0098] Step S390-2, similarity analysis is performed on the standard fault phenomenon and the fault phenomenon of each candidate root cause corresponding to the root cause similar case to obtain the support degree of each candidate root cause.

[0099] In the embodiment of the application, similarity analysis is performed on the standard fault phenomenon and the fault phenomenon recorded in each root cause similar case. This process converts the text description into a vector representation and uses a semantic matching algorithm such as cosine similarity to calculate the closeness between the two, thereby obtaining a numerical indicator reflecting the consistency of the current fault and the historical case, i.e. the support degree. The support degree is used to measure whether the current fault performance is highly consistent with the typical performance mode of the candidate root cause.

[0100] Step S390-3, determine the diagnostic confidence of each candidate root cause according to the prior probability of the candidate root cause, the preset case weight and the support degree.

[0101] In the embodiment of the application, the product of the preset case weight and the support degree is calculated, and the obtained product is added to the prior probability to obtain the diagnostic confidence of each candidate root cause. If there are multiple root cause similar cases for the candidate root cause, the diagnostic confidence of each diagnostic root cause is calculated according to the following formula.

[0102]

[0103] wherein, is the diagnostic confidence of the i th candidate root cause; is the prior probability of the i th candidate root cause; is the preset case weight of the j th root cause similar case; is the support degree of the j th root cause similar case; and Q is the number of root cause similar cases corresponding to the i th candidate root cause.

[0104] As can be seen, in this embodiment of the invention, each candidate root cause is given an independent diagnostic confidence level. The diagnostic confidence level not only inherits the static causal logic in the knowledge graph, but also integrates the dual verification results of dynamic operation data and historical handling experience, which effectively improves the intelligence level and engineering practicality of fault root cause localization, so that the diagnostic conclusions are both theoretically based and close to real operation and maintenance scenarios.

[0105] Optionally, regarding how to construct the optical module knowledge graph, the following is one possible implementation method. Please refer to... Figure 2 The optical module maintenance method also includes the following steps: Step S40: Obtain the multimodal engineering data corresponding to the optical module; the multimodal engineering data includes R&D records, generated data, test data, user feedback, and maintenance reports.

[0106] In this embodiment of the invention, by integrating multi-source heterogeneous engineering data and performing structured modeling, a technological transformation from raw information to computable knowledge is achieved, aiming to form a unified, scalable, and semantically clear knowledge foundation. Specifically, multimodal engineering data corresponding to optical modules is acquired as the data foundation for knowledge graph construction. This multimodal engineering data encompasses various sources such as R&D records, production data, test data, user feedback, and maintenance reports. Its data format includes both structured data like tables and unstructured text content such as technical documents and debugging logs, ensuring that the extracted knowledge covers key experiences and facts throughout the product's entire lifecycle.

[0107] Step S50: Extract entities, entity attributes, and relationships between entities from multimodal engineering data according to pre-set rules; entities include optical module attributes, performance parameters, adjustable parameters, fault phenomena, root causes of faults, solutions, environmental factors, and historical cases.

[0108] In this embodiment of the invention, entities, entity attributes, and relationships between entities are extracted from multimodal engineering data according to pre-defined rules. Specifically, for structured data, a data mapping method is used to directly identify the meaning of fields, extracting entities such as optical module model, transmission distance, and bit error rate, along with their entity attributes (i.e., attribute values).

[0109] For unstructured text, natural language processing technology is used, combined with named entity recognition and relation extraction algorithms, to identify entities such as extinction ratio and laser aging, and to extract semantic relationships (i.e., association relationships) such as the effect of bias current on extinction ratio and temperature drift leading to a decrease in receiver sensitivity, thereby making the professional knowledge implicit in the text explicit.

[0110] It should be understood that the extracted entities include but are not limited to optical module attributes, performance parameters, adjustable parameters, fault phenomena, fault root causes, solutions, environmental factors and historical cases. Each type of entity is defined with corresponding entity attributes: for example, "optical module attributes" have attributes such as manufacturer, data rate and transmission distance; "adjustable parameters" have attributes such as current value and adjustable range; "fault root causes" include attributes such as description and occurrence probability; and "solutions" record execution information such as operation steps.

[0111] Meanwhile, the association relationship is defined to depict the dynamic and causal connection between entities, including but not limited to "has parameter" representing the ownership relationship between the optical module and its parameters, "mutual influence" expressing the coupling effect between the adjustable parameters and the performance parameters, "causes" connecting the fault root cause and the fault phenomenon, "applies to" establishing the adaptation relationship between the solution and the fault type, and "best works in" describing the matching state of the module and the environmental condition.

[0112] In step S60, the extracted entities, entity attributes of the entities and the association relationship between the entities are stored into a graph database to obtain an optical module knowledge graph.

[0113] In the embodiment of the application, all the extracted entities, entity attributes and the association relationship between the entities are imported into a graph database to complete the persistent storage and organization of knowledge, and thus an optical module knowledge graph is constructed. The optical module knowledge graph represents entities by nodes and represents the association relationship by edges, and the entity attributes are additional feature fields of the nodes, supporting efficient graph traversal and query operations, and providing structured knowledge support for subsequent parameter collaborative optimization and fault root cause analysis.

[0114] It should be understood that the embodiment of the application realizes a fundamental change from relying on artificial experience to intelligent collaborative optimization of parameter tuning by constructing an optical module knowledge graph, and can perform accurate multi-parameter collaborative optimization on key adjustable parameters such as bias current and extinction ratio. Meanwhile, relying on the semantic reasoning capability of the optical module knowledge graph, accurate diagnosis from phenomenon alarm to root cause positioning of faults can be realized, and the operation and maintenance efficiency is significantly improved; the whole process converts the originally dispersed and implicit engineering experience into a calculable and interpretable structured knowledge model, and on the basis of effectively solving the knowledge inheritance problem, a closed-loop learning mechanism formed by continuous accumulation of historical cases has dynamic evolution and self-optimization capabilities, thereby comprehensively improving the performance control level and production debugging efficiency of the optical module.

[0115] Based on the same inventive concept, the basic principles and technical effects of the optical module operation and maintenance device provided by the embodiment of the application are the same as those of the above-mentioned embodiment. For brevity of description, the part not mentioned in this embodiment can refer to the corresponding content in the above-mentioned embodiment.

[0116] Please refer toFigure 3 , Figure 3 A block diagram of the optical module operation and maintenance device 400 is provided. The optical module operation and maintenance device 400 includes a processing module 410 and an operation and maintenance module 420.

[0117] The processing module 410 is configured to receive target optical module information and operation and maintenance information in response to a user's operation and maintenance request for a target optical module. The operation and maintenance request includes a performance tuning request or a fault diagnosis request. The operation and maintenance module 420 is configured to determine a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information, perform a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information, and generate at least one set of operation and maintenance schemes starting from the target optical module node. The operation and maintenance schemes are used to echo to the user for viewing.

[0118] In summary, the optical module operation and maintenance device provided by the embodiments of the present application can respond to a user's performance tuning request or fault diagnosis request for a target optical module, receive target optical module information and operation and maintenance information, and ensure that differentiated processing can be performed according to actual operation and maintenance scenarios. The target optical module node is determined in a pre-constructed optical module knowledge graph according to the target optical module information, thereby realizing accurate mapping between specific device instances and structured knowledge systems. The graph traversal operation is performed in the optical module knowledge graph based on the operation and maintenance information starting from the target optical module node, at least one set of operation and maintenance schemes is generated, and the operation and maintenance schemes are used to echo to the user for viewing. The correlation reasoning capability of the knowledge graph is used to search and expand along the semantic relationship path from a single node, thereby breaking data silos and overcoming experience dependence, realizing personalized and interpretable operation and maintenance suggestion generation for specific device states and operation and maintenance targets, enhancing response accuracy and decision transparency in the optical module operation and maintenance process, and improving the efficiency and performance of optical module tuning and diagnosis.

[0119] Optionally, the operation and maintenance information includes a target performance indicator and current environment data. The operation and maintenance module 420 is specifically configured to, if the operation and maintenance request is a performance tuning request, determine a target performance parameter and a target performance value according to the target performance indicator; search, based on the target performance parameter, a target adjustable parameter that affects the target performance parameter in the optical module knowledge graph, and obtain a constraint boundary of each target adjustable parameter; search, based on the current environment data and the target optical module information, a historical optimal case with a similarity exceeding a preset first threshold in the optical module knowledge graph, and determine a tuning starting value of each target adjustable parameter according to a search result of the historical optimal case; take the target performance value as an optimization target, take the tuning starting value of each target adjustable parameter as a starting point, and perform a multi-objective collaborative optimization search on each target adjustable parameter within the constraint boundary of all target adjustable parameters to obtain at least one parameter tuning combination; the parameter tuning combination includes a recommended tuning value and an expected performance value of the target adjustable parameter; calculate a tuning confidence of each parameter tuning combination; and one parameter tuning combination and the corresponding tuning confidence constitute a set of operation and maintenance schemes.

[0120] Optionally, the target optical module information includes a plurality of attribute information of the target optical module. The operation and maintenance module 420 is specifically configured to find, in the optical module knowledge graph, a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data; if a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is found, a historical optimal case is obtained; if a historical case that is completely identical to the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is not found, candidate information is generated according to a priority of each attribute information; find, in the optical module knowledge graph, a historical case that is completely identical to the candidate information and has a first similarity exceeding a preset first threshold with the current environment data; if a historical case that is completely identical to the candidate information and has a first similarity exceeding a preset first threshold with the current environment data is found, a historical optimal case is obtained.

[0121] Optionally, the operation and maintenance module 420 is specifically configured to, if the historical optimal case exists, determine a value of the target adjustable parameter in the historical optimal case as the tuning starting value of the target adjustable parameter; and if the historical optimal case does not exist, determine a default value of the target adjustable parameter as the tuning starting value of the target adjustable parameter.

[0122] Optionally, the operation and maintenance module 420 is specifically configured to calculate a second similarity of each parameter tuning combination according to the target optical module information, the target performance index, the current environment data and the historical optimal case; determine a theoretical performance value of each target adjustable parameter according to the performance theory curve and the recommended tuning value of each target adjustable parameter, and calculate a third similarity of each parameter tuning combination according to the expected performance value and the target performance value of each parameter tuning combination; calculate a safety margin ratio of each parameter tuning combination according to the recommended tuning value of each target adjustable parameter in each parameter tuning combination and the constraint boundary; and determine a tuning confidence of each parameter tuning combination according to the second similarity, the third similarity and the safety margin ratio of each parameter tuning combination.

[0123] Optionally, the operation and maintenance information includes current fault information. The operation and maintenance module 420 is specifically configured to, if the operation and maintenance request is a fault diagnosis request, perform standardization processing on the current fault information to obtain a standard fault phenomenon; search for fault root causes affecting the standard fault phenomenon in the optical module knowledge graph to obtain at least one candidate root cause; search for a solution corresponding to each candidate root cause in the optical module knowledge graph according to each candidate root cause; search for a historical case most similar to the fault root cause and the candidate root cause in the optical module knowledge graph according to each candidate root cause to obtain a root cause similar case corresponding to each candidate root cause; calculate a diagnosis confidence of each candidate root cause according to the standard fault phenomenon, each candidate root cause and the corresponding root cause similar case; and form a group of operation and maintenance solutions from each candidate root cause, the corresponding solution and the corresponding diagnosis confidence.

[0124] Optionally, the operation and maintenance module 420 is specifically configured to find a fault type caused by the candidate root cause in the optical module knowledge graph, and determine a prior probability of the candidate root cause according to the number of times of the fault type caused by the candidate root cause and the total number of the fault type; perform similarity analysis on the fault phenomenon of the standard fault phenomenon and the root cause similar case corresponding to each candidate root cause to obtain a support degree of each candidate root cause; and determine the diagnosis confidence of each candidate root cause according to the prior probability, the preset case weight and the support degree of the candidate root cause.

[0125] Optionally, the processing module 410 is further configured to acquire multi-modal engineering data corresponding to the optical module; the multi-modal engineering data includes research and development records, generated data, test data, user feedback and maintenance reports; extract entities, entity attributes and association relationships between entities from the multi-modal engineering data according to a pre-set rule; the entities include optical module attributes, performance parameters, adjustable parameters, fault phenomena, fault root causes, solutions, environmental factors and historical cases; and store the extracted entities, entity attributes of the entities and association relationships between the entities in a graph database to obtain the optical module knowledge graph.

[0126] Please refer to Figure 4Fig. 1 shows a block schematic diagram of an electronic device 500 according to an embodiment of the present application. The electronic device 500 can be, but is not limited to, a personal computer (PC), a Personal Digital Assistant (PDA), a notebook computer, a tablet computer, a server. The electronic device 500 comprises a memory 510, a processor 520 and a communication module 530. The memory 510, the processor 520 and the communication module 530 are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, the memory 510, the processor 520 and the communication module 530 can be electrically connected to each other through one or more communication buses or signal lines.

[0127] The memory 510 is configured to store programs or data. The memory 510 can be, but is not limited to, a random access memory, a read only memory, a programmable read only memory, an erasable programmable read only memory, an electrically erasable programmable read only memory, etc.

[0128] The processor 520 is configured to read / write the data or programs stored in the memory 510 and perform corresponding functions. For example, when the computer program stored in the memory 510 is executed by the processor 520, the optical module operation and maintenance method disclosed in the above embodiments can be implemented.

[0129] The communication module 530 is configured to establish a communication connection between the electronic device 500 and other communication terminals through a network and to receive / transmit data through the network.

[0130] It should be understood that, Figure 4 The structure shown in Fig. 1 is only a schematic diagram of the structure of the electronic device 500. The electronic device 500 can further comprise more or less components than those shown in Fig. 1, or have a different configuration from that shown in Fig. 1. Figure 4 The components shown in Fig. 1 can be implemented in hardware, software or a combination thereof. Figure 4 Figure 4 The embodiments of the present application further provide a computer readable storage medium having a computer program stored thereon, which, when executed by the processor 520, implements the optical module operation and maintenance method disclosed in the above embodiments.

[0131] The embodiments of the present application further provide a program product, which, when executed by the processor 520, implements the optical module operation and maintenance method disclosed in the above embodiments.

[0132] The embodiments of the present application further provide a program product, which, when executed by the processor 520, implements the optical module operation and maintenance method disclosed in the above embodiments.

[0133] ​In several embodiments provided by the present application, it should be understood that the disclosed apparatus and method can also be implemented in other manners. The described embodiments of the apparatus are merely exemplary, and the present application can be implemented in other manners. For example, the apparatus embodiments described above can be implemented in a manner as shown in the flowcharts and block diagrams of the accompanying drawings. In this regard, each block in the flowcharts and block diagrams can represent a module, a segment, or a portion of code which comprises one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions shown in the blocks can occur in a different order than that shown in the figures. For example, two blocks shown in succession can in fact be executed substantially concurrently or in the reverse order, depending on the functionality involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented by dedicated hardware-based systems which perform the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0134] In addition, each functional module in the various embodiments of the present application can be integrated together to form a separate part, or each module can exist independently, or two or more modules can be integrated to form a separate part.

[0135] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part that contributes to the prior art, or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0136] The above only represents the preferred embodiments of the present application and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. An optical module operation and maintenance method, characterized by, The method comprises: In response to a user's operation and maintenance request for a target optical module, receiving target optical module information and operation and maintenance information; the operation and maintenance request includes a performance tuning request or a fault diagnosis request; According to the target optical module information, determining a target optical module node in a pre-constructed optical module knowledge graph; Taking the target optical module node as a starting point, performing a graph traversal operation in the optical module knowledge graph based on the operation and maintenance information to generate at least one set of operation and maintenance schemes; the operation and maintenance scheme is used to echo to the user for viewing.

2. The optical module operation and maintenance method according to claim 1, characterized in that, The operation and maintenance information includes a target performance index and current environment data, and the operation and maintenance information is used to perform a graph traversal operation in the optical module knowledge graph to generate at least one set of operation and maintenance schemes, which includes: If the operation and maintenance request is a performance tuning request, determine a target performance parameter and a target performance value according to the target performance index; Based on the target performance parameter, search for target adjustable parameters affecting the target performance parameter in the optical module knowledge graph, and obtain the constraint boundary of each target adjustable parameter; Based on the current environment data and the target optical module information, search for a historical optimal case with a similarity exceeding a preset first threshold in the optical module knowledge graph, and determine a tuning starting value of each target adjustable parameter according to the search result of the historical optimal case; Taking the target performance value as an optimization target, and taking the tuning starting value of each target adjustable parameter as a starting point, performing a multi-objective collaborative optimization search on each target adjustable parameter within the constraint boundary of all target adjustable parameters to obtain at least one parameter tuning combination; the parameter tuning combination includes a recommended tuning value and an expected performance value of the target adjustable parameter; Calculate the tuning confidence of each parameter tuning combination; one parameter tuning combination and the corresponding tuning confidence constitute a set of operation and maintenance schemes.

3. The optical module operation and maintenance method according to claim 2, characterized in that, The target optical module information includes a plurality of attribute information of the target optical module; and the historical optimal case with a similarity exceeding a preset first threshold is searched in the optical module knowledge graph based on the current environment data and the target optical module information, which includes: Find a historical case in the optical module knowledge graph that is completely the same as the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data; If a historical case that is completely the same as the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is found, the historical optimal case is obtained; If a historical case that is completely the same as the target optical module information and has a first similarity exceeding a preset first threshold with the current environment data is not found, candidate information is generated according to the priority of each attribute information; Find a historical case in the optical module knowledge graph that is completely the same as the candidate information and has a first similarity exceeding a preset first threshold with the current environment data; If a historical case that is completely the same as the candidate information and has a first similarity exceeding a preset first threshold with the current environment data is found, the historical optimal case is obtained.

4. The optical module operation and maintenance method according to claim 2, characterized in that, The determination of the tuning starting value of each target adjustable parameter according to the search result of the historical optimal case includes: If the historical optimal case exists, a value of a target adjustable parameter in the historical optimal case is determined as an optimization starting value of the target adjustable parameter; If the historical optimal case does not exist, a default value of the target adjustable parameter is determined as the optimization starting value of the target adjustable parameter.

5. The optical module operation and maintenance method according to claim 2, characterized in that, The calculation of the optimization confidence of each parameter optimization combination comprises: calculating a second similarity of each parameter optimization combination according to the target optical module information, the target performance index, the current environment data and the historical optimal case; calculating a third similarity of each parameter optimization combination according to an expected performance value and a target performance value of each parameter optimization combination; calculating a safety margin ratio of each parameter optimization combination according to a recommended optimization value and a constraint boundary of each target adjustable parameter in each parameter optimization combination; determining an optimization confidence of each parameter optimization combination according to the second similarity, the third similarity and the safety margin ratio of each parameter optimization combination.

6. The optical module operation and maintenance method according to claim 1, characterized in that, The operation and maintenance information comprises current fault information, and the graph traversal operation is performed in the optical module knowledge graph based on the operation and maintenance information to generate at least one set of operation and maintenance schemes, which comprises: If the operation and maintenance request is a fault diagnosis request, the current fault information is standardized to obtain a standard fault phenomenon; searching for a fault root cause affecting the standard fault phenomenon in the optical module knowledge graph to obtain at least one candidate root cause; searching for a solution corresponding to each candidate root cause in the optical module knowledge graph according to each candidate root cause; searching for a historical case most similar to the candidate root cause in the optical module knowledge graph according to each candidate root cause to obtain a root cause similar case corresponding to each candidate root cause; calculating a diagnosis confidence of each candidate root cause according to the standard fault phenomenon, each candidate root cause and the corresponding root cause similar case; each candidate root cause, the corresponding solution and the corresponding diagnosis confidence constitute a set of operation and maintenance schemes.

7. The optical module operation and maintenance method according to claim 6, characterized in that, The calculation of the diagnosis confidence of each candidate root cause according to the standard fault phenomenon, each candidate root cause and the corresponding root cause similar case comprises: finding a fault type caused by the candidate root cause in the optical module knowledge graph, and determining a prior probability of the candidate root cause according to a number of times of occurrence of the fault type caused by the candidate root cause and a total number of occurrences of the fault type; performing similarity analysis on the fault phenomenon of the standard fault phenomenon and the root cause similar case corresponding to each candidate root cause to obtain a support degree of each candidate root cause; determining the diagnosis confidence of each candidate root cause according to the prior probability of the candidate root cause, a preset case weight and the support degree.

8. The optical module operation and maintenance method according to claim 1, characterized in that, The optical module knowledge graph is constructed by the following method: obtaining multi-modal engineering data corresponding to an optical module; the multi-modal engineering data comprises research and development records, generated data, test data, user feedback and maintenance reports; extracting entities, entity attributes and associations between entities from the multi-modal engineering data according to pre-set rules; the entities include optical module attributes, performance parameters, adjustable parameters, failure phenomena, failure root causes, solutions, environmental factors and historical cases; storing the extracted entities, entity attributes and associations between entities into a graph database to obtain the optical module knowledge graph.

9. An optical module operation and maintenance device, characterized by comprising: The device comprises: a processing module configured to receive target optical module information and operation and maintenance information in response to a user request for operation and maintenance of a target optical module; the operation and maintenance request includes a performance tuning request or a fault diagnosis request; an operation and maintenance module configured to determine a target optical module node in a pre-constructed optical module knowledge graph according to the target optical module information, perform a graph traversal operation in the optical module knowledge graph based on the target optical module node as a starting point and the operation and maintenance information, and generate at least one set of operation and maintenance solutions; the operation and maintenance solutions are used to be echoed to the user for viewing.

10. An electronic device, comprising: comprising a processor and a memory, the memory storing a computer program executable by the processor, and the processor being capable of executing the computer program to implement the optical module operation and maintenance method of any one of claims 1-8.

Images