Voice reporting management system and management method

Abstract

The invention discloses a voice report management system and management method, and belongs to the technical field of voice processing, and the system comprises a voice request module, an edge node module and a decision making module. The voice request module is used for acquiring voice report data of each device in the workshop and sending the voice report data to the edge node module; the edge node module is used for identifying and analyzing the voice report data to obtain voice text information; and the decision making module is used for determining an initial fault level, a maintenance suggestion and a historical maintenance record according to the voice text information based on a preset knowledge graph, determining a target fault level according to the historical maintenance record, the real-time equipment parameters and the initial fault level, and determining a maintenance decision scheme according to the target fault level and the maintenance suggestion. According to the invention, abnormity reporting is carried out through the voice data, and the voice data is analyzed and processed, so that rapid and simple reporting operation of workshop abnormal equipment is realized, and meanwhile, the reported abnormal state is responded in time.

Description

Technical Field

[0001] The present invention relates to the technical field of speech processing, and in particular, to a speech reporting management system and a management method. Background Art

[0002] With the increasingly fierce market competition, manufacturing enterprises are facing more and more uncertain factors. The production tasks of modern manufacturing enterprises often have characteristics such as a variety of products, small batch sizes, and high quality requirements. This complexity makes the production process more vulnerable to various factors, resulting in production anomalies. Factors such as production equipment, processes, and personnel within the enterprise may also change, and these changes may also trigger production anomalies. Through effective anomaly management, enterprises can timely discover and solve problems in the production process, thereby avoiding or reducing production downtime and delays, and improving production efficiency. Anomaly management helps to reduce waste in the production process, such as raw material waste and labor waste. By correcting anomalies in a timely manner, enterprises can reduce production costs and improve profitability. Anomaly management focuses on quality issues in the production process. By timely discovering and handling quality anomalies, enterprises can ensure the stability and reliability of product quality and enhance customer satisfaction. In short, doing a good job in anomaly management during the production process is an important matter for each manufacturing enterprise.

[0003] In the prior art, the first step in anomaly management is the timely reporting of anomalies. In the scenario of a production workshop, the main job of ordinary workers is to perform production execution. When an anomaly occurs and is discovered by a worker, the conventional method is for the worker to put down the work at hand and manually input a series of information such as the anomaly type, problem description, occurrence location, etc. through the system on the computer next to the work station, and finally submit it. The whole process is rather troublesome and is also somewhat difficult for ordinary workers with a low educational level or those who are not used to operating computers. Therefore, in this scenario, anomaly information usually cannot be reported well and in a timely manner, and intelligent guidance for handling cannot be provided.

[0004] Therefore, there is an urgent need for a speech reporting management system and a management method to solve the technical problem in the prior art that anomalies cannot be reported simply and in a timely manner. Summary of the Invention

[0005] In view of this, it is necessary to provide a speech reporting management system and a management method that can report anomalies through voice data, analyze and process the voice data, and timely process the abnormal equipment.

[0006] To solve the above technical problems, on the one hand, the present invention provides a speech reporting management system, including a voice request module, an edge node module, and a decision-making module; The voice request module is used to obtain the voice reporting data of each device in the workshop and send the voice reporting data to the edge node module; The edge node module is used to identify and analyze the voice reporting data to obtain voice text information; The decision-making module is used to determine the initial fault level, maintenance suggestions and historical maintenance records based on the preset knowledge graph according to the voice text information, and determine the target fault level according to the historical maintenance records, real-time device parameters and the initial fault level, and determine the maintenance decision plan according to the target fault level and maintenance suggestions.

[0007] In a possible implementation manner, the voice request module includes a voice acquisition sub-module and a reporting trigger sub-module; The voice acquisition sub-module is used to obtain voice reporting data through voice receiving devices installed on each device; The reporting trigger sub-module is used to control the voice reporting data to the edge node module.

[0008] In a possible implementation manner, the voice reporting data includes device name voice data and abnormal state voice data, and the edge node module includes a noise reduction and filtering sub-module, a voice recognition sub-module and a semantic analysis sub-module; The noise reduction and filtering sub-module is used to perform noise reduction and filtering operations on the voice reporting data to obtain preprocessed voice data; The voice recognition sub-module is used to perform voice recognition on the preprocessed voice data based on automatic speech recognition to obtain voice text data; The semantic analysis sub-module is used to identify the device name information and device abnormal state information in the voice text data based on natural language processing technology, and the device name information and device abnormal state information constitute voice text information.

[0009] In a possible implementation manner, the decision-making module includes an initial fault level determination sub-module, a historical maintenance record query sub-module, a fault level judgment sub-module and a decision-making sub-module; The initial fault level determination sub-module is used to obtain the equipment ledger information in the preset MES database according to the voice text information, and determine the initial fault level and maintenance suggestions according to the ledger information and the preset knowledge graph database; The historical maintenance record query sub-module is used to query the historical maintenance records of the equipment based on the knowledge graph database according to the voice text information; The fault level judgment sub-module is used to determine the target fault level based on the preset rule engine according to the historical maintenance records and the obtained real-time device parameters; The decision-making sub-module is used to determine the final maintenance decision-making plan based on the rule engine, according to the target fault level, historical maintenance records, and voice text information.

[0010] In a possible implementation manner, the initial fault level determination sub-module includes an account information query unit and an initial level prediction unit; The account information query unit is used to obtain the account information of the corresponding device in the preset MES system according to the device name information; The initial level prediction unit is used to determine the initial fault level and maintenance suggestions of the device based on the knowledge graph according to the device account information.

[0011] In a possible implementation manner, the fault level judgment sub-module includes a fault feature weight allocation unit and a level risk prediction unit; The fault feature weight allocation unit is used to extract the fault features in different dimensions from the historical maintenance records and current device parameters, and determine the weights of the fault features in each dimension based on the preset risk level rule engine; The level risk prediction unit is used to determine the target fault level based on the dynamic Bayesian algorithm according to the weights and the fault features in each dimension.

[0012] In a possible implementation manner, the fault feature weight allocation unit includes a feature extraction sub-unit and a weight allocation sub-unit; The feature extraction sub-unit is used to perform data fusion and processing on the historical maintenance records and current device parameters to obtain a multi-source data set, and extract the fault features in different dimensions from the multi-source data set; The weight allocation sub-unit is used to determine the weight allocation in each dimension based on the preset risk level rule engine according to the historical maintenance records and knowledge graph data of the fault features in different dimensions.

[0013] In a possible implementation manner, the system further includes a decision execution module; The decision execution module is used to generate a maintenance work order according to the maintenance decision-making plan, execute the maintenance work order to repair the device and / or trigger an alarm.

[0014] In a possible implementation manner, the system further includes a feedback confirmation module; The feedback confirmation module is used to perform multi-channel confirmation according to the decision-making plan and the response result of the decision-making plan, and at the same time write the device name information, device abnormal state information, fault type, fault level, and the corresponding decision-making plan into the preset knowledge graph.

[0015] Second aspect, the present invention also provides a voice reporting management method, which is applicable to the voice reporting management system described in any one of the above. The method includes: Obtain the voice reporting data of each device in the workshop; Identify and analyze the voice reporting data to obtain voice text information; Based on a preset knowledge graph, determine the initial fault level, maintenance suggestions, and historical maintenance records according to the voice text information, and determine the target fault level according to the historical maintenance records, real-time device parameters, and initial fault level. Determine the maintenance decision-making plan according to the target fault level and maintenance suggestions.

[0016] The beneficial effects of the present invention are as follows: The voice reporting management system provided by the present invention captures the abnormal voice reporting data of each device by installing a voice request module on each device, and at the same time triggers an abnormal report, and sends the voice reporting data to the edge node module. The voice request module only needs to obtain the simple natural voice of the production worker to describe the device name and abnormal status, and can report in the first time when an abnormality occurs, and the operation is simple, without the need to report on a computer; the edge node module is set on several nodes at the edge of the workshop, receives the voice reporting data, and screens, identifies, and analyzes the voice reporting data, and can quickly identify and analyze the abnormal status voice data received at close range, reducing the data processing process of the remote terminal; the decision-making module set at the terminal obtains the initial fault level and maintenance suggestions according to the voice text information including the device name information and device abnormal status information obtained by voice recognition and analysis, and at the same time queries the historical maintenance records; determines the final target fault level according to the obtained real-time device parameters, historical maintenance records, and initial fault level, and finally determines the maintenance decision-making plan according to the target level, and can quickly match the corresponding fault and maintenance suggestions to determine the maintenance decision. The present invention captures the voice reporting data through the voice request module on each device. When an abnormality occurs, only the voice reporting data needs to be provided, and then the voice reporting data is quickly identified and analyzed at the corresponding edge node in the workshop. The analyzed data is combined with a preset knowledge graph to obtain historical maintenance records and determine the corresponding maintenance plan, realizing the rapid and simple reporting of the devices in the abnormal state of the workshop, and being able to respond in time to the reported abnormal state. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0018] Figure 1Schematic structural diagram of an embodiment of the voice reporting management system provided by the present invention; Figure 2 Schematic structural diagram of an embodiment of the initial fault level determination sub-module provided by the present invention; Figure 3 Schematic structural diagram of an embodiment of the fault level judgment sub-module provided by the present invention; Figure 4 Schematic flow diagram of an embodiment of the voice reporting management method provided by the present invention. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0020] In the description of the embodiments of the present invention, unless otherwise specified, the meaning of "a plurality" is two or more.

[0021] In the embodiments of the present invention, the descriptions such as "first" and "second" are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Therefore, the technical features defined with "first" and "second" may explicitly or implicitly include at least one such feature.

[0022] Referring to "embodiments" in this article means that the specific features, structures or characteristics described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0023] Before the description of the embodiments, the relevant terms are defined: The MES system (Manufacturing Execution System) is a key execution layer system that connects the enterprise planning layer (such as the ERP system) and the industrial control layer (such as device control systems such as PLC and DCS). It optimizes the entire production process from order placement to product completion by collecting, processing and feeding back data in the production process in real time, helping enterprises achieve efficient, transparent and intelligent production management.

[0024] The present invention provides a voice reporting management system and a management method, which will be described separately below.

[0025] Figure 1 A schematic structural diagram of an embodiment of the voice reporting management system provided by the present invention is shown as Figure 1 shown. The voice reporting management system includes a voice request module 100, an edge node module 200, and a decision-making module 300; The voice request module 100 is used to obtain the voice reporting data of each device in the workshop and send the voice reporting data to the edge node module 200; It should be noted that through the voice receiving device and the reporting button set on each production device, the voice receiving device includes but is not limited to an industrial microphone, and the reporting button includes but is not limited to a pressure overlimit trigger switch. When an abnormality occurs in the production device, the production worker presses the pressure overlimit trigger switch and describes the name and abnormal state of the abnormal device to perform abnormal voice reporting.

[0026] The edge node module 200 is used to identify and parse the voice reporting data to obtain voice text information; It should be noted that in this embodiment, multiple edge nodes are set in the workshop. Each node is responsible for multiple production devices adjacent to the edge node. An edge node module 200 is installed on each edge node to process the voice reporting data sent by the voice request module 100 on the corresponding production device, perform operations such as filtering and noise reduction, voice recognition, and semantic parsing on the voice reporting data, and finally convert the voice reporting data into text information including device name information and device abnormal state information, and send the voice text information to the local terminal server for subsequent processing. For example, when the voice reporting data is "CNC-2035 spindle overheat", the edge node module 200 parses the voice reporting data into device name information = CNC-2035 and abnormal state information = spindle overheat.

[0027] The decision-making module 300 is used to determine the initial fault level, maintenance suggestions, and historical maintenance records based on a preset knowledge graph according to the voice text information, determine the target fault level according to the historical maintenance records, real-time device parameters, and the initial fault level, and determine the maintenance decision plan according to the target fault level and the maintenance suggestions.

[0028] It should be noted that the decision-making module 300 is processed through a terminal server, which is an intelligent terminal including the MES system. Based on the technology of the MES system, the decision-making module 300 is set to receive the voice text information sent by the edge node module 200, link with the database in the MES system according to the voice text information, obtain the knowledge graph database in the MES system for data matching, obtain the initial fault level and maintenance suggestions, query the historical maintenance records of the equipment, fuse these multi-source data, determine the final fault level, and determine the maintenance decision plan according to the fault level and maintenance suggestions.

[0029] It should be further noted that the voice reporting management system in this embodiment is developed on the basis of the MES system to solve the technical problem of untimely abnormal reporting in the MES system.

[0030] Through the voice reporting management system developed on the basis of the MES system in this embodiment, through experimental data testing, it has the following advantages when dealing with abnormal reporting. First, the response timeliness is faster than the traditional manual recording method. The average time-consuming of the traditional manual recording is 4.2 minutes, and the voice reporting management system of this embodiment can achieve timely response from the device to the server and from the server to the device within 55 seconds. Second, the integrity of information reporting is improved. There will be missing keywords in the traditional manual recording, and the missing rate reaches 37%. The structured field integrity rate of this embodiment reaches 95%. Third, the operation is simpler. The traditional operation page needs to be operated on a computer, and this embodiment only requires pure voice interaction. Fourth, the reuse of knowledge. The abnormal handling experience in the traditional method lacks digitization, and the management system of this embodiment can perform dynamic knowledge graph evolution. Therefore, this embodiment can achieve fast and accurate reporting and timely response to maintenance when the device has an abnormality.

[0031] In this embodiment, the voice request module 100 on each device captures the voice reporting data. When an abnormality occurs, only the voice reporting data needs to be provided, and then the voice reporting data is quickly identified and parsed on the corresponding edge node in the workshop. The parsed data is combined with the preset knowledge graph to obtain the historical maintenance records, and the corresponding maintenance plan is determined, realizing fast and simple reporting of the devices in the abnormal state in the workshop and being able to respond in a timely manner to the reported abnormal state.

[0032] In some embodiments of the present invention, the voice request module 100 includes a voice acquisition sub-module 110 and a reporting trigger sub-module 120; The voice acquisition sub-module 110 is used to acquire the voice reporting data through the voice receiving device installed on each device; It should be noted that the voice receiving device installed on each production device is used to receive the voice of equipment anomalies. When providing voice reporting data, the production worker only needs to provide the equipment name and the equipment anomaly status.

[0033] The reporting trigger sub-module 120 is used to control the upload of voice reporting data to the edge node module 200.

[0034] It should be noted that the equipment anomaly voice reporting is realized through the pressing pressure overrun trigger switch installed on each production device.

[0035] In this embodiment, through the voice receiving device and the anomaly reporting button, the process of anomaly reporting is simplified. The production worker only needs to provide the equipment name and the anomaly status, thus realizing the simplicity and timeliness of the anomaly reporting operation.

[0036] In some embodiments of the present invention, the voice reporting data includes equipment name voice data and anomaly status voice data. The edge node module 200 includes a noise reduction and filtering sub-module 210, a voice recognition sub-module 220, and a semantic analysis sub-module 230; The noise reduction and filtering sub-module 210 is used to perform noise reduction and filtering operations on the voice reporting data to obtain preprocessed voice data; Specifically, based on the field programmable gate array (FPGA), through parallel processing and pipeline design, combined with a noise reduction algorithm, the voice reporting data is subjected to noise reduction and filtering to improve the quality of the voice signal and provide high-quality data basis for subsequent voice recognition and analysis.

[0037] The voice recognition sub-module 220 is used to perform voice recognition on the preprocessed voice data based on automatic speech recognition to obtain voice text data; Specifically, the automatic speech recognition (ASR) technology is adopted to convert the voice reporting data after noise reduction and filtering into text data. This process is completed at the edge node in the workshop, without uploading the voice reporting data to the server, thereby improving the data security and the timeliness of data processing.

[0038] The semantic analysis sub-module 230 is used to identify the equipment name information and the equipment anomaly status information in the voice text data based on natural language processing technology. The equipment name information and the equipment anomaly status information constitute the voice text information.

[0039] Specifically, the natural language processing (NLP) technology is adopted to further process and understand the text data recognized by ASR, and the text data is parsed into equipment name information and equipment anomaly status information.

[0040] In this embodiment, each voice processing device installed on the edge nodes of the workshop filters and denoises the received device anomaly voice reporting data, and identifies and analyzes the device name information and device anomaly status information therein, improving the timeliness and security of voice processing.

[0041] In some embodiments of the present invention, the decision-making module 300 includes an initial fault level determination sub-module 310, a historical maintenance record query sub-module 320, a fault level judgment sub-module 330, and a decision-making sub-module 340; The initial fault level determination sub-module 310 is used to obtain the equipment ledger information in the preset MES database according to the voice text information, and determine the initial fault level and maintenance suggestions according to the ledger information and the preset knowledge graph database; It should be noted that when the decision-making module 300 receives the voice text information from the edge node module 200, it automatically triggers a structured retrieval request, requests to be linked with the MES database, and queries the equipment ledger information.

[0042] Furthermore, it should be noted that various internal data and external data are collected in advance. The internal data includes the business system data within the enterprise, such as production data and supply chain data in the ERP system, customer complaint records and sales data in the CRM system, etc. These data contain a large amount of entity and relationship information related to anomaly management; the external data includes external data sources, such as industry reports, news, social media, etc. These data can help discover the general anomalies and trends within the industry, such as the market reaction caused by the product quality problems of competitors, etc.; integrate the data from different data sources, clean the data, remove duplicate, incorrect, and incomplete data records, and standardize the data; identify the entities related to anomaly management from the processed data, such as equipment, products, customers, and suppliers, etc.; use natural language processing technology for entity recognition, identify the relationships between entities, such as the relationship between equipment and faults, the relationship between customers and complaints, etc.; use rule-based methods, etc. for relationship extraction. For example, use a relationship extraction model to identify the association relationships between equipment and the causal relationships between equipment and faults from production data, etc.; adopt a graph database, knowledge graph construction tool, etc. to perform knowledge fusion on the identified entities and relationships, use the Neo4j graph database to store the entities and relationships as nodes and edges, construct a knowledge graph for anomaly management, and at the same time, update the knowledge graph with the anomaly reporting data and the decision-making maintenance plan to achieve the dynamic and automatic evolution of the knowledge graph.

[0043] The historical maintenance record query sub-module 320 is used to query the historical maintenance records of the equipment based on the knowledge graph database according to the voice text information; It should be noted that the knowledge graph also includes a historical maintenance record sub-database to query the historical maintenance records of the device in the last three times.

[0044] The fault level judgment sub-module 330 is used to determine the target fault level based on a preset rule engine according to the historical maintenance records and the obtained real-time device parameters. It should be noted that based on a Java-based business rule management system, a Drools rule engine is constructed to associate information among various data, between fault levels, and for maintenance decisions, etc. The historical maintenance records and the real-time obtained device parameters are associated according to the rule engine, and the weight distribution coefficient is calculated based on the association result, and the final fault level of the device is determined according to the weight distribution coefficient.

[0045] The decision-making sub-module 340 is used to determine the final maintenance decision plan based on the rule engine according to the target fault level, historical maintenance records, and voice text information.

[0046] Specifically, the rule engine is matched according to the obtained final fault level to determine the final maintenance strategy.

[0047] In some embodiments of the present invention, as Figure 2 shown, Figure 2 is a schematic structural diagram of an embodiment of the initial fault level determination sub-module provided by the present invention. The initial fault level determination sub-module 310 includes an account information query unit 311 and an initial level prediction unit 312. The account information query unit 311 is used to obtain the account information of the corresponding device in the preset MES system according to the device name information. The initial level prediction unit 312 is used to determine the initial fault level and maintenance suggestions of the device based on the preset knowledge graph according to the device account information.

[0048] Specifically, according to the account information of the device, the Neo4j knowledge graph database is queried. During the query process, the device name and abnormal state are input, and combined with the account information, the initial probability of each fault level and the maintenance suggestions corresponding to the fault level can be obtained. For example, when the device name information is CNC-2035, the device signal in its account information is queried as MAZAK-630, and the standard parameter values of this type of device. These information are matched with the Neo4j knowledge graph database to obtain the initial fault level and maintenance suggestions of the device.

[0049] In some embodiments of the present invention, Figure 3 is a schematic structural diagram of an embodiment of the fault level judgment sub-module provided by the present invention. The fault level judgment sub-module 330 includes a fault feature weight distribution unit 331 and a level risk prediction unit 332. The fault feature weight allocation unit 331 is used to extract fault features in different dimensions from historical maintenance records and current device parameters, and determine the weights of the fault features in each dimension based on a preset rule engine; In some embodiments of the present invention, the fault feature weight allocation unit 331 includes a feature extraction subunit and a weight allocation subunit; The feature extraction subunit is used to perform data fusion and processing on historical maintenance records and current device parameters to obtain a multi-source data set, and extract fault features in different dimensions from the multi-source data set; The weight allocation subunit is used to determine the weight allocation of each dimension based on a preset risk level rule engine according to the historical maintenance records and knowledge graph data of the fault features in different dimensions.

[0050] It should be noted that through the fusion analysis of historical maintenance records and current device parameters, the current value, knowledge graph benchmark value, and historical reference value of the device in different dimensions are determined to construct a feature matrix in different dimensions, and weight coefficients are assigned to the features in each dimension according to the feature matrix. In this embodiment, taking "CNC-2035 spindle overheating" as an example, the spindle temperature and historical maintenance records of the current device are analyzed, and combined with the rule engine, a feature matrix is obtained, and weight coefficients are assigned to each feature dimension, as shown in Table 1. Table 1 is the feature matrix in an embodiment provided by the present invention: Table 1 is the feature matrix in an embodiment provided by the present invention

[0051] The level risk prediction unit 332 is used to determine the target fault level based on the dynamic Bayesian algorithm according to the weights and fault features in each dimension.

[0052] It should be noted that by using the dynamic Bayesian algorithm, according to the current feature values and weight coefficients in each dimension, the probability of this abnormal state at each fault level is obtained, and the fault level corresponding to the maximum probability is the target fault level of the device.

[0053] In this embodiment, through the fusion analysis of multi-source data, the dynamic weight allocation of features, and the combination of the dynamic Bayesian algorithm, the target fault level of the device can be accurately determined, providing accurate data basis for the maintenance decision of the abnormal state.

[0054] In some embodiments of the present invention, the voice reporting management system further includes a decision execution module 400; The decision execution module 400 is used to generate a maintenance work order according to the maintenance decision plan, execute the maintenance work order to repair the device and / or trigger an alarm.

[0055] It should be noted that the maintenance decision-making plan includes control plans for the equipment, such as shutting down the machine or triggering an alarm, corresponding maintenance personnel information, maintenance suggestions, etc. These information are classified and sent, the maintenance suggestions and abnormal status are pushed to the corresponding maintenance personnel, and if shutdown is required, the equipment is controlled to shut down.

[0056] In some embodiments of the present invention, the voice reporting management system further includes a feedback confirmation module 500; The feedback confirmation module 500 is used to perform multi-channel confirmation according to the decision-making plan and the response result of the decision-making plan, and at the same time write the equipment name information, equipment abnormal status information, fault type, fault level, and the corresponding decision-making plan into a preset knowledge graph.

[0057] It should be noted that when the equipment executes the control information in the maintenance decision-making plan, confirmation feedback is performed through hearing and vision. The control information is broadcast in a directional sound field, such as "The machine has been shut down, please check the main shaft bearing", and light warning is performed through the optical signal installed on the workbench. At the same time, the information of the maintenance decision-making plan is sorted and classified and written into the knowledge graph to provide a data basis for the next abnormal management.

[0058] To better implement the voice reporting management system in the embodiments of the present invention, correspondingly, on the basis of the voice reporting management system, as Figure 4 shown, the embodiments of the present invention further provide a voice reporting management method, including: S401. Obtain the voice reporting data of each device in the workshop; S402. Identify and analyze the voice reporting data to obtain voice text information; S403. Based on a preset knowledge graph, determine the initial fault level, maintenance suggestions, and historical maintenance records according to the voice text information, determine the target fault level according to the historical maintenance records, real-time device parameters, and the initial fault level, and determine the maintenance decision-making plan according to the target fault level and the maintenance suggestions.

[0059] It should be noted that: the voice reporting management method provided in the above embodiments can implement the technical solutions described in the embodiments of the above voice reporting management system. The principles or specific implementation details implemented by the above steps can refer to the corresponding content in the embodiments of the above voice reporting system, and will not be elaborated here one by one.

[0060] Those skilled in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware (such as a processor, a controller, etc.) through a computer program. The computer program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.

[0061] The above has introduced in detail the voice reporting management method, device, equipment and storage device provided by the present invention. Specific examples are used in this article to elaborate on the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation manner and application scope. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A voice reporting management system, characterized in that: It includes voice request module, edge node module and decision making module; The voice request module is used to obtain the voice report data of each device in the workshop and send the voice report data to the edge node module; The edge node module is used to identify and analyze the voice reporting data to obtain voice text information; The decision-making module is used to determine the initial fault level, maintenance suggestions and historical maintenance records based on the voice text information based on a preset knowledge graph, and determine the target fault level based on the historical maintenance records, real-time equipment parameters and the initial fault level, and determine the maintenance decision plan based on the target fault level and maintenance suggestions.

2. The voice reporting management system according to claim 1, characterized in that: The voice request module includes a voice acquisition submodule and a reporting trigger submodule; The voice acquisition submodule is used to acquire voice reporting data through a voice receiving device installed on each device; The reporting triggering submodule is used to control the voice reporting data to the edge node module.

3. The voice reporting management system according to claim 2, characterized in that: The voice reporting data includes device name voice data and abnormal state voice data, and the edge node module includes a noise reduction filter submodule, a voice recognition submodule and a semantic analysis submodule; The noise reduction and filtering submodule is used to perform noise reduction and filtering operations on the voice reporting data to obtain pre-processed voice data; The speech recognition submodule is used to perform speech recognition on the pre-processed speech data based on automatic speech recognition to obtain speech text data; The semantic analysis submodule is used to identify the device name information and the device abnormal state information in the voice text data based on natural language processing technology, and the device name information and the device abnormal state information constitute the voice text information.

4. The voice reporting management system according to claim 3, characterized in that: The decision-making module includes an initial fault level determination submodule, a historical maintenance record query submodule, a fault level determination submodule and a decision submodule; The initial fault level determination submodule is used to obtain the equipment ledger information in the preset MES database according to the voice text information, and determine the initial fault level and maintenance suggestions according to the ledger information and the preset knowledge graph database; The historical maintenance record query submodule is used to query the historical maintenance record of the equipment according to the voice text information based on the knowledge graph database; The fault level judgment submodule is used to determine the target fault level based on the preset rule engine, according to the historical maintenance records and the acquired real-time equipment parameters; The decision submodule is used to determine a final maintenance decision plan based on the rule engine, according to the target fault level, historical maintenance records and voice text information.

5. The voice reporting management system according to claim 4, characterized in that: The initial fault level determination submodule includes an account information query unit and an initial level prediction unit; The ledger information query unit is used to obtain the ledger information of the corresponding equipment in the preset MES system according to the equipment name information; The initial level prediction unit is used to determine the initial fault level and maintenance suggestions of the equipment based on the knowledge graph and according to the equipment ledger information.

6. The voice reporting management system according to claim 4, characterized in that: The fault level judgment submodule includes a fault feature weight allocation unit and a level risk prejudgment unit; The fault feature weight allocation unit is used to extract fault features of different dimensions from the historical maintenance records and current equipment parameters, and determine the weights of the fault features in each dimension based on a preset risk level rule engine; The level risk prediction unit is used to determine the target fault level based on the dynamic Bayesian algorithm according to the weights and fault characteristics in each dimension.

7. The voice reporting management system according to claim 6, characterized in that: The fault feature weight allocation unit includes a feature extraction subunit and a weight allocation subunit; The feature extraction subunit is used to fuse and process the historical maintenance records and current equipment parameters to obtain a multi-source data set, and extract fault features under different dimensions in the multi-source data set; The weight allocation subunit is used to determine the weight allocation of each dimension based on the preset risk level rule engine, according to the historical maintenance records and knowledge graph data of fault characteristics in different dimensions.

8. The voice reporting management system according to claim 1, characterized in that: The system also includes a decision execution module; The decision execution module is used to generate a maintenance work order according to the maintenance decision plan, and execute the maintenance work order to repair the equipment and / or trigger an alarm.

9. The voice reporting management system according to claim 3, characterized in that: The system also includes a feedback confirmation module; The feedback confirmation module is used to perform multi-channel confirmation based on the decision plan and the response result of the decision plan, and at the same time write the device name information, device abnormal status information, fault type, fault level and corresponding decision plan into a preset knowledge graph.

10. A voice reporting management method, characterized in that: Applicable to the voice reporting management system according to any one of claims 1 to 9, the method comprising: Obtain voice reporting data from each device in the workshop; Identify and analyze the voice reporting data to obtain voice text information; Based on the preset knowledge graph, the initial fault level, maintenance suggestions and historical maintenance records are determined according to the voice text information, and the target fault level is determined according to the historical maintenance records, real-time equipment parameters and the initial fault level, and the maintenance decision plan is determined according to the target fault level and maintenance suggestions.

Images