Fault detection method and apparatus, communication device, storage medium, and computer program product

Abstract

The application relates to a fault detection method and device, communication equipment, a storage medium and a computer program product. The method comprises the following steps: performing initial fault detection on sample data of a production device by using a first model to obtain fault information of the production device; based on the fault information, initiating a cooperative detection task to a cloud end; the cooperative detection task is used for instructing the cloud end to perform fault detection on the production device; comparing a fault detection result fed back by the cloud end with information related to the fault detection result in a local knowledge base, and when the comparison result is consistent, updating a newly added fault detection result to the local knowledge base. The method can improve the accuracy of fault diagnosis results.

Description

Technical Field

[0001] This application relates to the field of industrial internet technology, and in particular to a fault detection method, device, communication equipment, storage medium, and computer program product. Background Technology

[0002] With the development of industrial technology, industrial production failure prediction technology has emerged. This technology uses equipment and service data, such as equipment usage data, operating condition data, main unit and accessory performance data, and accessory replacement data, to predict equipment failures, service, and accessory needs. The aim is to reduce the impact of failures on production by identifying potential failures in advance and taking corresponding measures.

[0003] In traditional technologies, the operating status data of faulty equipment is usually collected, and the operating status data of the faulty equipment is analyzed and processed through a local fault prediction model to obtain the fault detection result of the faulty equipment.

[0004] However, in traditional technologies, only the operating status data of faulty equipment is collected, and the faulty equipment is detected through a fault detection model. The data content is limited and the accuracy of the fault detection results is low. Summary of the Invention

[0005] This application provides a fault detection method, apparatus, communication device, storage medium, and computer program product.

[0006] Firstly, a fault detection method is provided, the method comprising:

[0007] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain the fault information of the production equipment.

[0008] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0009] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0010] In one embodiment, the method further includes:

[0011] Raw data from the production equipment is collected by sensors, and the raw data is processed to obtain sample data.

[0012] In one embodiment, the initial fault detection using a first model on sample data of the production equipment to obtain fault information of the production equipment includes:

[0013] The first model is used to perform initial fault detection on sample data of production equipment to identify abnormal signals and / or potential fault point information in the sample data.

[0014] Based on the abnormal signals and / or the potential fault point information, the candidate fault area of ​​the production equipment is determined.

[0015] In one embodiment, initiating a collaborative detection task to the cloud based on the fault information includes:

[0016] Detect whether the fault information is a new detection task;

[0017] If it is a new detection task, a collaborative detection task request is sent to the cloud, and the collaborative detection request contains the full sample data of the fault candidate region;

[0018] If it is not the new detection task, a collaborative detection task request is sent to the cloud, and the collaborative detection request includes the consistency detection result.

[0019] In one embodiment, if it is a new detection task, a collaborative detection task request is sent to the cloud, including:

[0020] If it is a new detection task, a data transmission strategy shall be formulated based on the urgency and bandwidth resources corresponding to the candidate fault regions;

[0021] Based on the data transmission strategy, all sample data in the collaborative detection task request is transmitted.

[0022] In one embodiment, transmitting the full amount of sample data in the collaborative detection task request based on the data transmission strategy includes:

[0023] If the urgency of the new detection task exceeds the urgency threshold, the full amount of sample data is transmitted in real time.

[0024] If the urgency of the new detection task is lower than the urgency threshold and the bandwidth resources are limited, the full amount of sample data will be transmitted in batches multiple times.

[0025] In one embodiment, comparing the fault detection result fed back from the cloud with information related to the fault detection result in the local knowledge base includes:

[0026] Retrieve data related to the fault detection results from the local knowledge base;

[0027] The fault detection results are compared with the information related to the fault detection results to analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

[0028] In one embodiment, the method further includes:

[0029] If the comparison results are inconsistent, the collaborative detection task will be re-initiated.

[0030] In one embodiment, the sensor collects raw data from the production equipment including: vibration signals, temperature, humidity, gas concentration, pressure, liquid flow rate, rotational speed, current, voltage, and video.

[0031] In one embodiment, processing the raw data includes: data cleaning, integration, and transformation.

[0032] Secondly, this application provides a fault detection method, the method comprising:

[0033] In response to a collaborative detection task initiated by the terminal, various detection methods are used to detect global information related to the collaborative detection task from the global knowledge base;

[0034] The second model is used to combine the global information related to the collaborative detection task with the sample data of the fault candidate region to generate a fault detection result, and the fault detection result is sent to the terminal.

[0035] In one embodiment, the step of combining global information related to the collaborative detection task with sample data of the fault candidate region using a second model to generate a fault detection result, and sending the fault detection result to the terminal, includes:

[0036] The sample data of the fault candidate region is integrated, filtered, and rearranged with the global information related to the collaborative detection task to obtain the processed fault detection task problem.

[0037] The fault detection task problem is input into the second model to generate fault detection results;

[0038] The fault detection results and the global information are sent to the terminal together.

[0039] Thirdly, this application provides a fault detection device, the device comprising:

[0040] The detection module is used to perform initial fault detection on the sample data of the production equipment using the first model to obtain the fault information of the production equipment.

[0041] The task initiation module is used to initiate a collaborative detection task to the cloud based on the fault information; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0042] The update module is used to compare the fault detection results fed back from the cloud with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0043] Fourthly, this application provides a fault detection device, the device comprising:

[0044] The detection module is used to respond to the collaborative detection task initiated by the terminal and use multiple detection methods to detect global information related to the collaborative detection task from the global knowledge base;

[0045] The generation module is used to combine global information related to the collaborative detection task with sample data of the fault candidate region using the second model to generate fault detection results, and send the fault detection results to the terminal.

[0046] Fifthly, this application provides a communication device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0047] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain the fault information of the production equipment.

[0048] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0049] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0050] Sixthly, this application provides a communication device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0051] In response to a collaborative detection task initiated by the terminal, various detection methods are used to detect global information related to the collaborative detection task from the global knowledge base;

[0052] The second model is used to combine the global information related to the collaborative detection task with the sample data of the fault candidate region to generate a fault detection result, and the fault detection result is sent to the terminal.

[0053] In a seventh aspect, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0054] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain the fault information of the production equipment.

[0055] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0056] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0057] Eighthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the fault detection method provided in the embodiments of this application. This method may include:

[0058] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain the fault information of the production equipment.

[0059] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0060] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0061] The aforementioned fault detection method, apparatus, communication equipment, storage medium, and computer program product employ a first model to perform initial fault detection on sample data of the production equipment, obtaining fault information of the production equipment. Based on the fault information, a collaborative detection task is initiated to the cloud. The collaborative detection task instructs the cloud to perform fault detection on the production equipment. The fault detection results fed back by the cloud are compared with information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base. Using this method, collaborative detection of production equipment is performed between the terminal and the cloud. Information is queried not only based on the local knowledge base but also on the global knowledge base in the cloud, achieving a comprehensive diagnosis of production equipment faults and improving the accuracy of fault diagnosis results. Attached Figure Description

[0062] Figure 1This is a diagram illustrating the application environment of a fault detection method in one embodiment;

[0063] Figure 2 This is a flowchart illustrating a fault detection method in one embodiment;

[0064] Figure 3 This is a flowchart illustrating the steps for obtaining sample data in one embodiment;

[0065] Figure 4 This is a flowchart illustrating the steps for determining candidate fault regions of production equipment in one embodiment.

[0066] Figure 5 This is a flowchart illustrating the steps of detecting whether a fault detection task is a new detection task in one embodiment.

[0067] Figure 6 This is a flowchart illustrating the steps for determining a data transmission strategy and sample data for a new detection task in one embodiment.

[0068] Figure 7 This is a flowchart illustrating the steps of determining a data transmission strategy based on the urgency of a detection task in one embodiment.

[0069] Figure 8 This is a flowchart illustrating the consistency verification steps for fault detection results based on a local knowledge base in one embodiment.

[0070] Figure 9 This is a flowchart illustrating the steps of re-initiating the collaborative detection task in one embodiment;

[0071] Figure 10 This is a flowchart illustrating the cloud-based fault detection method in one embodiment.

[0072] Figure 11 This is a flowchart illustrating the fault detection result feedback step in one embodiment;

[0073] Figure 12 This is a schematic diagram illustrating a specific example of a fault detection method in one embodiment;

[0074] Figure 13 This is a flowchart of the steps of a fault detection method in one embodiment;

[0075] Figure 14 This is a structural block diagram of a fault detection device on the terminal side in one embodiment;

[0076] Figure 15 This is a structural block diagram of a fault detection device on the cloud side in one embodiment;

[0077] Figure 16This is an internal structural diagram of a communication device in one embodiment. Detailed Implementation

[0078] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0079] Figure 1 This is a schematic diagram illustrating an application scenario for fault detection provided in an embodiment of this application. For example... Figure 1 As shown, this industrial scenario includes industrial equipment, terminal 100 (which can also be called a fault detection device), and cloud 200. The terminal 100 and cloud 200 transmit data via a network.

[0080] Terminal 100 can be a wireless terminal, which can be a device providing voice and / or other service data connectivity to a user, a handheld device with wireless connectivity, or other processing devices connected to a wireless modem. The wireless terminal can communicate with one or more core networks via a Radio Access Network (RAN). The wireless terminal can be a mobile terminal, such as a mobile phone (or "cellular" phone) or a computer with a mobile terminal, for example, a portable, pocket-sized, handheld, computer-embedded, or vehicle-mounted mobile device, exchanging voice and / or data with the RAN. The wireless terminal can also be referred to as a system, subscriber unit, subscriber station, mobile station, mobile, remote station, remote terminal, access terminal, user terminal, user agent, or user device or user equipment, without limitation herein.

[0081] Industrial production failure prediction is a predictive maintenance technology based on data analysis and intelligent algorithms. It uses equipment usage data, operating condition data, main unit and accessory performance data, and accessory replacement data to predict equipment failures, service, and accessory needs. The aim is to reduce the impact of failures on production by identifying potential failures in advance and taking corresponding measures.

[0082] Traditional techniques typically involve collecting operational status data from faulty equipment and then analyzing this data using a local fault prediction model (large model) to obtain the fault detection result. Therefore, traditional techniques primarily focus on how to leverage large models to improve the efficiency of fault detection and repair in industrial scenarios. However, traditional techniques, which only collect operational status data from faulty equipment and use large models for fault detection, suffer from limitations in data content and accuracy.

[0083] Based on the aforementioned traditional technologies, this application provides a fault detection method that performs collaborative detection of production equipment via a terminal and the cloud. This method not only queries information based on a local knowledge base but also on a global knowledge base in the cloud, thereby achieving a comprehensive diagnosis of production equipment faults and improving the accuracy of fault diagnosis results.

[0084] It should be noted that the beneficial effects or technical problems solved by the embodiments of this application are not limited to this one, but may also be other implicit or related problems. For details, please refer to the description of the embodiments below.

[0085] Before introducing specific embodiments of the present invention, the technical terms involved in the present invention will be explained:

[0086] Large-scale pre-trained models are deep learning models used in fields such as Natural Language Processing (NLP) and Computer Vision (CV). They are trained on massive amounts of data using unsupervised pre-training to achieve powerful representational and generalization capabilities. The core advantage of large-scale models lies in their ability to capture complex patterns and semantic relationships within the data, resulting in excellent performance across various downstream tasks, even on tasks with relatively small datasets.

[0087] Small models, in contrast to large models, typically refer to machine learning or deep learning models with fewer parameters and simpler structures. These models are designed to run on limited computing resources or to achieve low-latency, high-energy-efficiency inference on edge devices and mobile devices. The advantages of small models are their faster training and inference speeds, lower memory consumption, and suitability for real-time applications and resource-constrained scenarios.

[0088] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0089] In one embodiment, such as Figure 2 As shown, a fault detection method is provided, which improves the accuracy of fault detection in industrial production by combining large and small model-end (terminal) cloud coordinated reasoning and knowledge base-based retrieval. Specifically, this method is applied to... Figure 1 Taking terminal 100 as an example, the explanation includes the following steps:

[0090] Step 202: Use the first model to perform initial fault detection on the sample data of the production equipment to obtain fault information of the production equipment.

[0091] In implementation, a first model is pre-integrated locally on the terminal. This first model is a small model. The terminal acquires sample data from various production equipment in an industrial scenario. Then, it uses the first model and the sample data to perform initial fault detection, obtaining fault information for the production equipment. This fault information includes suspicious candidate regions where abnormal signals or potential fault points exist.

[0092] Step 204: Based on the fault information, initiate a collaborative detection task to the cloud.

[0093] Among them, the collaborative detection task is used to instruct the cloud to perform fault detection on production equipment.

[0094] In implementation, after the terminal provides initial fault detection, it initiates a collaborative detection task to the cloud based on the fault information characterizing the fault detection. Specifically, the terminal sends a collaborative detection task request to the cloud, which carries fault information, including abnormal signals or potential fault point data, as well as the location information of suspicious candidate areas of the abnormal signals or potential fault points. In this way, the cloud can perform fault detection on various production equipment in the current industrial scenario based on the received fault information. A large model and a global knowledge base are pre-deployed in the cloud. The cloud can analyze and process the fault information based on the large model and the global knowledge base, providing global fault detection results, which can then be fed back to the terminal.

[0095] Step 206: Compare the fault detection results fed back from the cloud with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, update the newly added fault detection results to the local knowledge base.

[0096] In implementation, the terminal needs to have a local knowledge base. This local knowledge base stores data related to fault detection results from various production equipment in the industrial scenario. For example, it might store instruction manuals, technical manuals, historical faults, fault causes, maintenance records, real-time operating status of the equipment, and current production tasks for each piece of equipment on the industrial production line. When the terminal receives fault detection results from the cloud, it compares these results with the information in the local knowledge base to further verify the fault detection results. If the fault detection results match the information in the local knowledge base, the determined fault detection result is accurate. Therefore, this result is output as the final result of the fault detection. Simultaneously, the terminal updates the local knowledge base with any newly added fault detection results to record the fault detection results for future querying and analysis.

[0097] In the aforementioned fault detection method, a first model is used to perform initial fault detection on sample data of the production equipment to obtain fault information. Based on the fault information, a collaborative detection task is initiated to the cloud. This collaborative detection task instructs the cloud to perform fault detection on the production equipment. The fault detection results fed back from the cloud are compared with information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base. Using this method, collaborative detection of production equipment is performed between the terminal and the cloud. This not only relies on the local knowledge base for information retrieval but also on the global knowledge base in the cloud, achieving a comprehensive diagnosis of production equipment faults and improving the accuracy of fault diagnosis results.

[0098] In one embodiment, multiple sensors are pre-deployed in an industrial setting. These sensors collect data on various production equipment and environmental information within the industrial environment to enable real-time monitoring of the operational status of the industrial setting. Figure 3 As shown, the method also includes:

[0099] Step 301: Collect raw data from the production equipment using sensors and process the raw data to obtain sample data.

[0100] In implementation, raw data from various production equipment is collected via sensors in an industrial setting. This raw data includes signals from each production device, operational status data, and environmental parameters specific to the industrial environment, such as vibration signals, temperature, humidity, gas concentration, pressure, liquid flow rate, rotational speed, current, voltage, and video data. This embodiment does not limit the type or quantity of raw data collected. A data acquisition and preprocessing module is pre-integrated in the terminal. This module collects various data from the production equipment on the production line and performs data cleaning, compression, integration, and feature extraction. Thus, after the sensors collect the raw data from each production device, the terminal's data acquisition and preprocessing module retrieves and processes this data to obtain sample data. Based on this sample data, fault detection is performed on all production equipment on the industrial production line in the entire industrial setting, allowing for early detection of potential faults and the implementation of corresponding measures to mitigate their impact on production.

[0101] Optionally, the processing of the raw data includes, but is not limited to, data cleaning, integration, transformation (standardization and normalization), etc. The processing method of the raw data in this application embodiment is not limited.

[0102] In this embodiment, the raw data of the production equipment is acquired and processed to obtain sample data for fault detection. Through preliminary processing of the raw data, the format of the raw data is unified, the data quality of the raw data is improved, and thus the accuracy of fault detection of the production equipment is improved.

[0103] In one embodiment, such as Figure 4 As shown, the specific processing steps of step 202 include:

[0104] Step 401: Use the first model to perform initial fault detection on the sample data of the production equipment, and identify abnormal signals and / or potential fault point information in the sample data.

[0105] During implementation, the local knowledge base stores information related to production line equipment, including but not limited to equipment manuals, technical manuals, historical fault records, and maintenance records. The terminal uses a first model to perform initial fault detection on sample data from the production equipment, conducting simple sample data classification and testing. Sample data with differences exceeding a threshold between it and baseline data is filtered out, identifying abnormal information and / or potential fault points within the sample data. For example, if the main control motor temperature is too high (e.g., exceeding the baseline by 20%), the main control motor temperature data is used as an abnormal signal, and the area of ​​the equipment containing that main control motor is marked as an abnormal area, i.e., a potential fault point.

[0106] Step 402: Based on abnormal signals and / or potential fault point information, determine the candidate fault areas of the production equipment.

[0107] In practice, based on abnormal signals and / or potential fault point information, the terminal determines the areas where abnormal signals and / or potential fault points occur in the distribution information of industrial scenarios, and uses these areas as candidate fault areas for production equipment, pending further fault verification.

[0108] In this embodiment, the sample data of the production equipment is initially detected by the local first model to determine the candidate fault areas of the production equipment, and then sent to the cloud for large model inference calculation. This reduces the amount of data transmission and the number of large model inferences in the fault detection task, thereby reducing data calculation and communication costs.

[0109] In one embodiment, before initiating a collaborative detection task to the cloud, a task type determination is required. This involves determining whether the current task is a new detection task. If it is not, it indicates an intermediate processing step in a fault detection task. Therefore, the data sent from the terminal to the cloud needs to include the results of local consistency checks. If the current detection task is a new one, it needs to be processed, such as... Figure 5 As shown, the specific processing steps of step 204 include:

[0110] Step 501: Check if the fault information is a new detection task.

[0111] During implementation, the terminal detects whether the fault information is a new detection task. Specifically, each fault detection task corresponds to a task identifier. The terminal can perform detection task based on the task identifier contained in the fault information, determine whether the task identifier has appeared (been recorded), and detect whether the fault detection task is a new detection task.

[0112] Step 502: If it is a new detection task, send a collaborative detection task request to the cloud.

[0113] The collaborative detection request includes full sample data of the fault candidate regions.

[0114] In practice, if the terminal is about to perform a new detection task, it sends a collaborative detection task request to the cloud. Since it is a new detection task, the collaborative detection task request needs to include all sample data of the area following the fault.

[0115] Step 503: If it is not a new detection task, send a collaborative detection task request to the cloud.

[0116] The collaborative detection request includes consistency detection results.

[0117] In practice, if the current fault detection task is not a new detection task, that is, if fault detection has already been performed, the terminal sends a collaborative detection task request to the cloud. This collaborative detection request contains the consistency detection results of the fault detection results based on the terminal's local knowledge base.

[0118] In this embodiment, the task type of the collaborative task request is checked to determine whether the collaborative task is a new detection task. This enables the marking and differentiation of coordinated detection tasks, and different information is contained in the collaborative detection requests sent at different stages, thereby reducing the amount of data transmission in the fault detection task.

[0119] In one embodiment, such as Figure 6 As shown, the specific processing steps of step 502 include:

[0120] Step 602: If it is a new detection task, formulate a data transmission strategy based on the urgency and bandwidth resources corresponding to the candidate fault area.

[0121] In implementation, the terminal pre-defines a data transmission strategy corresponding to the fault detection task. This strategy includes urgency levels and bandwidth resource allocation rules for different production equipment areas on the industrial production line. Thus, when a new fault detection task is detected, the terminal formulates a data transmission strategy based on the urgency level and bandwidth resources corresponding to the candidate fault area.

[0122] Step 604: Based on the data transmission strategy, transmit all sample data from the collaborative detection task request.

[0123] In implementation, the terminal transmits all sample data in the collaborative detection task request based on the established data transmission strategy. Specifically, different fault candidate areas correspond to different levels of urgency, and each urgency level corresponds to a different data transmission time limit requirement. Therefore, after determining the different urgency levels corresponding to the fault candidate areas, the terminal transmits all sample data in the collaborative detection task request based on the data transmission time limit requirement corresponding to that urgency level.

[0124] In this embodiment, different urgency levels are set for different fault candidate areas in the fault detection task. The data transmission time limit of the collaborative detection task is determined by the urgency level, thereby ensuring the timeliness of the fault detection task, so as to discover potential faults in advance and reduce the impact of faults.

[0125] In one embodiment, such as Figure 7 As shown, the specific processing steps of step 604 include:

[0126] Step 701: If the urgency of a new detection task exceeds the urgency threshold, transmit all sample data in real time.

[0127] In implementation, different levels of urgency in the terminal correspond to urgency thresholds, which are used to classify urgency levels. Assuming there is one urgency threshold, corresponding to two levels, if the urgency of a new detection task exceeds this threshold, it indicates that the current fault detection task is urgent and needs to be completed as soon as possible. Therefore, the terminal transmits all sample data in real time. Taking the fault detection of a precision component manufacturing equipment in an industrial scenario as an example, the production line collects vibration signals, temperature, pressure, rotational speed, current, voltage, and other data from the main control motor of the CNC lathe, the press, and the laser welding machine. By detecting each sample data, the fault candidate area is determined to be the main control motor area, which is the core equipment in lathe production. Therefore, if the urgency level corresponding to this main control motor area exceeds the preset urgency threshold, the relevant sample data of this main control motor area needs to be transmitted to the cloud in real time to complete the fault detection as quickly as possible.

[0128] Step 702: If the urgency of the new detection task is lower than the urgency threshold and bandwidth resources are limited, transmit the full amount of sample data in batches multiple times.

[0129] In practice, if the urgency of a new detection task is lower than the urgency threshold, indicating that the current fault detection task is not particularly urgent, and the bandwidth resources of the current network are limited, then in order to ensure the smooth execution of other tasks in the current network, the terminal transmits the full amount of sample data in batches multiple times. This not only ensures the accurate transmission of sample data, but also reduces bandwidth usage and the amount of data transmitted each time.

[0130] In this embodiment, by dividing the urgency of fault candidate regions and differentiating the different data transmission timeliness corresponding to different fault candidate regions, not only is the timeliness of the transmission of sample data for urgent fault detection tasks guaranteed, but the impact of the entire fault detection task on the bandwidth usage between the terminal and the cloud is also reduced.

[0131] In one embodiment, such as Figure 8 As shown, the specific processing steps of step 206 include:

[0132] Step 802: Retrieve data information related to the fault detection results from the local knowledge base.

[0133] In practice, after receiving the fault detection results from the cloud, the terminal retrieves relevant data from its local knowledge base. For example, regarding the issue of overheating of the main control machine tool, after receiving the fault detection results from the cloud, the terminal organizes information related to the overheating of the main control machine tool and the refrigeration equipment from its local knowledge base. This includes the machine tool's instruction manual, technical manual, historical faults, fault causes, and maintenance records, as well as the machine tool's real-time operating status, including the operating status and refrigerant quantity of the auxiliary refrigeration equipment, maintenance records, and so on.

[0134] Step 804: Compare the fault detection results with the information related to the fault detection results, and analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

[0135] In implementation, the terminal compares the fault detection results with related information. For example, the fault detection result "the cause of the high temperature is the low amount of R134a refrigerant" is compared with the information "the refrigerant needs to be added once every 3 months, and it has not been added for more than 3 months" in the relevant information retrieved from the local knowledge base. The semantic conclusions represented by the two pieces of information are confirmed to be consistent, that is, the conclusion that the comparison results are consistent is given. In other words, it can be considered that the cause of the high temperature of the main control machine tool is the low amount of refrigerant in the refrigeration equipment.

[0136] In this embodiment, after preliminary detection by the terminal's small model and global detection by the cloud's large model, the fault detection results are obtained. Then, the fault detection results are further verified by combining the terminal's local knowledge base. The verification of multiple fault detection results improves the accuracy of the fault detection results.

[0137] In one embodiment, such as Figure 9 As shown, the method also includes:

[0138] Step 901: If the comparison results are inconsistent, re-initiate the collaborative detection task.

[0139] In practice, when the comparison results are inconsistent, the terminal needs to re-initiate the collaborative detection task. For example, if the fault detection result states that "the cause of the excessively high temperature is the low amount of R134a refrigerant," while the relevant information for the main control machine tool stored in the local knowledge base includes "new refrigerant was added within the last 10 days," these two pieces of information represent contradictory semantic conclusions. It can be considered that the current fault detection result may be problematic. Therefore, the terminal can re-initiate the collaborative detection task, with the terminal and the cloud re-performing fault detection and related information queries.

[0140] In this embodiment, the fault detection results are further verified by combining the terminal's local knowledge base, and the verification of multiple fault detection results improves the accuracy of the fault detection results.

[0141] In one embodiment, the sensor collects raw data from the production equipment, including vibration signals, temperature, humidity, gas concentration, pressure, liquid flow rate, rotational speed, current, voltage, and video.

[0142] In one embodiment, processing raw data in an industrial setting includes: data cleaning, integration, and transformation. This application does not limit the content of the sample data or the specific data processing methods used in the data preprocessing stage.

[0143] In one embodiment, such as Figure 10 As shown, a fault detection method is provided that improves the accuracy of fault detection in industrial production through coordinated reasoning between large and small model terminals (terminals) and knowledge base-based retrieval. Specifically, the method is illustrated using an example of its application in the cloud. The method includes:

[0144] Step 1001: In response to the collaborative detection task initiated by the terminal, use multiple detection methods to detect global information related to the collaborative detection task from the global knowledge base.

[0145] In implementation, the cloud responds to the collaborative detection task initiated by the terminal, and uses a retrieval tool to detect global information related to the collaborative detection task (collaborative fault detection task) initiated by the terminal from the global knowledge base using various retrieval methods. These various retrieval methods include, but are not limited to, vector retrieval, keyword retrieval, and knowledge image retrieval. This application embodiment does not limit the retrieval methods used by the cloud. The global knowledge base includes, but is not limited to, instruction manuals, technical manuals, historical fault and maintenance records of the corresponding equipment, the latest research reports in the industry, etc. This application does not limit the information contained in the global knowledge base on the cloud.

[0146] Step 1002: Use the second model to combine the global information related to the collaborative detection task with the sample data of the fault candidate region to generate the fault detection result, and send the fault detection result to the terminal.

[0147] In implementation, a second model is pre-integrated in the cloud. This second model is a larger model than the first model. The cloud uses this larger model to search for global information in the global database and combines the global information with sample data of the fault candidate area. The larger model is used to generate more accurate and specific fault classification and prediction results (i.e., fault detection results). The fault detection results are then sent to the terminal for further verification of the fault detection results.

[0148] In this embodiment, the production equipment is detected collaboratively by the terminal and the cloud. Information is queried not only based on the local knowledge base, but also based on the global knowledge base in the cloud, so as to achieve a comprehensive diagnosis of production equipment faults and improve the accuracy of fault diagnosis results.

[0149] In one embodiment, after initial fault detection on the terminal side, the terminal side feeds back fault information to the cloud. This fault information only includes fault candidate areas where faults may exist and the corresponding fault data. Then, after performing relevant global information retrieval on the cloud, such as... Figure 11 As shown, the specific processing steps of step 1002 include:

[0150] Step 1101: Integrate, filter, and rearrange the sample data of the fault candidate region with the global information related to the collaborative detection task to obtain the processed fault detection task problem.

[0151] In implementation, the generator in the cloud integrates, filters, and rearranges the sample data of the fault candidate region with the global information related to the collaborative detection task to obtain the processed fault detection task problem. Specifically, the cloud supplements the preliminary fault diagnosis problem corresponding to the sample data of the fault candidate region (i.e., the sample data of the problem area) based on the global information to obtain the processed fault detection task problem.

[0152] Step 1102: Input the fault detection task problem into the second model to generate fault detection results.

[0153] In implementation, the cloud inputs the fault detection task problem into a second model, which analyzes and processes the problem to generate fault detection results. These results can be a detailed fault classification and prediction report, including possible fault types, corresponding handling suggestions, and preventative measures.

[0154] Step 1103: Send the fault detection results and global information to the terminal together.

[0155] In implementation, the cloud sends the fault detection results and global information to the terminal together to increase the interpretability of the fault detection results and facilitate users' understanding and comparison of the fault detection results.

[0156] In this embodiment, after the fault detection results are generated by the large model in the cloud, the fault detection results and related global information are fed back to the terminal so that the terminal can further understand and verify the fault detection results and global information, thereby improving the detection accuracy of the fault detection results.

[0157] In one embodiment, such as Figure 12 As shown, a specific example flow of a fault detection method is provided. This specific example flow involves cooperation and collaboration between the terminal and the cloud. The fault detection method includes:

[0158] Step 1201: The terminal-side sensor collects raw data from each production equipment on the industrial production line and performs data preprocessing on the raw data to obtain sample data of the production equipment.

[0159] The sample data includes vibration signals, temperature, pressure, current, voltage, and other data from punch presses and arc welding machines on the production line. The data has been cleaned, compressed, denoised, and feature extracted.

[0160] Step 1202: The terminal performs initial fault detection on the sample data of each production equipment to obtain fault information of the production equipment.

[0161] Specifically, a local industrial mini-model (the first model) is used to classify and detect data, identifying samples whose differences from the benchmark exceed a threshold. For example, if the detection result shows that the mechanical vibration amplitude of the punch press is too large (exceeding the normal value by 20%), the area where the punch press is located is marked as a candidate area.

[0162] Step 1203: Based on the fault information, the terminal initiates a collaborative detection task to the cloud.

[0163] Specifically:

[0164] (1) It is determined that the collaborative detection task is a new detection task. The sample data sent is the sample data of the candidate area where the punch press is located, including the vibration amplitude, voltage, current and temperature data of the punch press and the voltage, current and temperature data of the control equipment.

[0165] (2) Formulate data transmission strategy: Since the candidate area is the punching area, which is the core equipment of production, the emergency level exceeds the threshold, and the sample data corresponding to the fault detection task needs to be sent in real time.

[0166] Step 1204: The cloud server receives the fault information and responds to the collaborative detection task initiated by the terminal by using multiple detection methods to detect global information related to the collaborative detection task from the global knowledge base.

[0167] The cloud server uses a keyword search method to retrieve the latest information and data related to excessive vibration amplitude of the punch press from the global knowledge base. The global knowledge base stores the equipment manual and technical manual for this model of punch press, the historical fault and maintenance records of the punch press, and the latest research reports on this punch press product.

[0168] Step 1205: The cloud server performs accurate fault analysis and prediction.

[0169] Specifically:

[0170] (1) Rewriting the problem based on the retrieval: The problem sent by the production line equipment, "The vibration amplitude of the punch press is too large (exceeding the normal value by 20%), needs to be further analyzed in conjunction with the control motherboard. The temperature of the control motherboard is normal, the current is normal, and the voltage is between 5 and 10W", is integrated and rearranged with the data retrieved from the global information to obtain the rewritten problem, "The vibration amplitude of the punch press is too large (exceeding the normal value by 20%), the temperature of the punch press control motherboard is normal, the voltage and current of the control line are normal, the usage time of the punch press robotic arm is 5 years and 11 months, and the service life of this model of robotic arm is 5 years."

[0171] (2) Large model generation: Input the rewritten problem into the large model to generate detailed fault classification and prediction results for the punch press: the punch press has a fault of excessive vibration amplitude, the cause of the fault is that the robotic arm has been used beyond its time limit and the components are aging, and the countermeasure is to replace the robotic arm.

[0172] (3) Feedback of detection results: Send the detection results generated by the large model and the information source of global information retrieval (version number of punch press technical manual) to the terminal together.

[0173] Step 1206: The terminal compares the fault detection result with the information related to the fault detection result and analyzes the degree of consistency between the fault detection result and the information related to the fault detection result.

[0174] Specifically:

[0175] (1) Conduct local retrieval of relevant information: organize information related to excessive punching amplitude and control motherboard in the local knowledge base, including the instruction manual, technical manual, historical faults, fault causes, maintenance records, real-time operating status of the punching machine and output voltage and current of the control motherboard, etc. The most relevant information obtained is: a new robotic arm was updated 12 months ago, and the robotic arm has been in use for 5 years.

[0176] (2) Consistency assessment: The results of the fault detection were compared with the information that “the fault was caused by the robotic arm being used beyond its service life and the components being aged” and “a new robotic arm was replaced 12 months ago and the robotic arm has been used for 5 years”. The results showed that the conclusions were inconsistent.

[0177] Step 1207: The terminal re-initiates the collaborative detection task.

[0178] Specifically:

[0179] (1) First determine that the collaborative detection task (fault detection task) is not a new detection task. The sent sample data contains fault detection results that are inconsistent with the local test. The fault detection results contain the information retrieved locally that "a new robotic arm was replaced 12 months ago and the robotic arm has been used for 5 years".

[0180] (2) Formulate data transmission strategy: Since the candidate area is the punching area, which is the core equipment of the production equipment, the emergency level exceeds the threshold, and the detection task needs to be sent in real time.

[0181] Step 1208: The cloud server responds to the collaborative detection task initiated by the terminal by detecting global information related to the collaborative task from the global knowledge base.

[0182] Yes, after receiving the collaborative detection task sent by the terminal, the cloud server uses a keyword retrieval method to retrieve information and data related to "excessive vibration amplitude of punch press" from the global knowledge base.

[0183] Step 1209: The cloud server performs accurate fault analysis and prediction.

[0184] After obtaining the results of the global information retrieval, 1) the problem was rewritten based on the retrieval: the problem sent by the terminal, "The vibration amplitude of the punch press is too large (exceeding the normal value by 20%), the temperature of the punch press control motherboard is normal, the control line voltage is between 5-10V, the current is normal, the robotic arm is normal, and further analysis is required," was integrated and rearranged with the data obtained from the global information retrieval to obtain the rewritten problem, "The vibration amplitude of the punch press is too large (exceeding the normal value by 20%), the temperature of the punch press control motherboard is normal, the control line voltage is between 5 and 10V, the current is normal, the robotic arm is normal, the control line needs to maintain a constant voltage within the range of 5 to 10V, and further analysis is required."

[0185] 2) Large model generation: Input the rewritten problem into the large model to generate detailed fault classification and prediction results for the punch press: The punch press has a fault of excessive vibration amplitude. The cause of the fault is the unstable voltage of the control circuit. The countermeasure is to stabilize the voltage of the control circuit at a fixed value within the range of 5 to 10V.

[0186] 3) Fault detection result feedback: The fault detection results generated by the large model and the information source of the global information retrieval (the version number of the punch press technical manual) are sent to the terminal together.

[0187] Step 1210: The terminal performs a local consistency check: After receiving the fault detection results from the cloud, the terminal compares them with the information related to the fault detection results in the local knowledge base to perform a consistency check.

[0188] Specifically: (1) Conduct local retrieval of relevant information: The terminal organizes information related to excessive punching amplitude and control motherboard in the local knowledge base, including punching manual, technical manual, historical faults, fault causes, maintenance records, real-time operating status of the punching and output voltage and current of the control motherboard, etc. The most relevant information obtained is: 6 months ago, the control motherboard of the punching had an overvoltage instability fault record. After replacing the voltage regulator capacitor on the motherboard, the voltage remained constant at 5V.

[0189] (2) Consistency assessment: The fault detection result “The cause of the fault is the unstable voltage of the control circuit. The countermeasure is to stabilize the voltage of the control circuit at a fixed value within the range of 5-10V” was compared with the retrieved information “Six months ago, the control motherboard of the punch press had a fault record of unstable voltage. After replacing the voltage regulator capacitor on the motherboard, the voltage remained constant at 5V”. The conclusion was found to be consistent.

[0190] Step 1211: The terminal updates the obtained fault detection results to the local knowledge base.

[0191] In one embodiment, such as Figure 13 The diagram shows a flowchart of a fault detection method. Figure 13 The fault detection method described above has been described in the various embodiments and will not be repeated here. Only the overall process will be described. Specifically:

[0192] Step 1301: Collect raw data from each production device using sensors and preprocess it to obtain sample data.

[0193] Step 1302: The terminal performs preliminary fault detection on the sample data to obtain fault information of the production equipment.

[0194] Step 1303: Based on the fault information, the terminal initiates a collaborative detection task to the cloud.

[0195] Step 1304: The cloud responds to the collaborative detection task initiated by the terminal and uses multiple detection methods to detect global information related to the collaborative detection task from the global knowledge base.

[0196] Step 1305: The cloud performs accurate fault analysis and prediction based on global information and sample data of fault candidate areas to obtain fault detection results.

[0197] Step 1306: The terminal receives the fault detection results from the cloud and performs a consistency check in the local knowledge base.

[0198] Step 1307: Compare the fault detection results with the relevant information in the local knowledge base. If they are inconsistent, proceed to step 1304. If they are consistent, proceed to step 1308.

[0199] Step 1308: Update the newly added fault detection results to the local knowledge base.

[0200] It should be understood that, although Figures 2 to 13 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 2 to 13 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0201] In one embodiment, such as Figure 14 As shown, a fault detection device 1400 is provided, including: a detection module 1401, a task initiation module 1402, and an update module 1403, wherein:

[0202] The detection module 1401 is used to perform initial fault detection on the sample data of the production equipment using the first model to obtain fault information of the production equipment.

[0203] The task initiation module 1402 is used to initiate a collaborative detection task to the cloud based on fault information; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0204] The update module 1403 is used to compare the fault detection results fed back from the cloud with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0205] In one embodiment, the device 1400 further includes:

[0206] The processing module is used to collect raw data from the production equipment via sensors and process the raw data to obtain sample data.

[0207] In one embodiment, the detection module 1401 is specifically used to perform initial fault detection on sample data of the production equipment using a first model, and to identify abnormal signals and / or potential fault point information in the sample data.

[0208] Based on abnormal signals and / or potential fault point information, identify the fault candidate area of ​​the production equipment.

[0209] In one embodiment, the detection module 1401 is specifically used to detect whether the fault information is a new detection task;

[0210] If it is a new detection task, a collaborative detection task request is sent to the cloud. The collaborative detection request contains the full sample data of the fault candidate region.

[0211] If it is not a new detection task, a collaborative detection task request is sent to the cloud, which includes the consistency detection results.

[0212] In one embodiment, the detection module 1401 is specifically used to formulate a data transmission strategy based on the urgency and bandwidth resources corresponding to the fault candidate area if it is a new detection task.

[0213] Based on the data transmission strategy, all sample data in the collaborative detection task request is transmitted.

[0214] In one embodiment, the detection module 1401 is specifically used to transmit the full amount of sample data in real time if the urgency of a new detection task exceeds an urgency threshold.

[0215] If the urgency of a new detection task is below the urgency threshold and bandwidth resources are limited, transmit the full amount of sample data in batches multiple times.

[0216] In one embodiment, the update module 1403 is specifically used to retrieve data information related to the fault detection result from the local knowledge base;

[0217] The fault detection results are compared with the information related to the fault detection results to analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

[0218] In one embodiment, the device 1400 further includes:

[0219] The task initiation module is used to re-initiate the collaborative detection task when the comparison results are inconsistent.

[0220] In one embodiment, the sensor collects raw data from the production equipment, including vibration signals, temperature, humidity, gas concentration, pressure, liquid flow rate, rotational speed, current, voltage, and video.

[0221] In one embodiment, processing the raw data includes: data cleaning, integration, and transformation.

[0222] In one embodiment, such as Figure 15 As shown, a fault detection device 1500 is provided, including: a detection module 1501 and a generation module 1502, wherein:

[0223] The detection module 1501 is used to respond to the collaborative detection task initiated by the terminal and use various detection methods to detect global information related to the collaborative detection task from the global knowledge base;

[0224] The generation module 1502 is used to combine global information related to the collaborative detection task with sample data of the fault candidate region using the second model to generate fault detection results and send the fault detection results to the terminal.

[0225] In one embodiment, the generation module 1502 is specifically used to integrate, filter, and rearrange the sample data of the fault candidate region with global information related to the collaborative detection task to obtain the processed fault detection task problem.

[0226] Input the fault detection task problem into the second model to generate fault detection results;

[0227] The fault detection results and global information are sent to the terminal together.

[0228] For specific limitations regarding the fault detection device, please refer to the limitations on the fault detection method above, which will not be repeated here. Each module in the aforementioned fault detection device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of the computer device in hardware form or independently of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0229] In one embodiment, a communication device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0230] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain fault information of the production equipment.

[0231] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0232] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0233] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0234] Raw data from production equipment is collected by sensors and processed to obtain sample data.

[0235] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0236] The first model is used to perform initial fault detection on sample data of production equipment to identify abnormal signals and / or potential fault point information in the sample data.

[0237] Based on abnormal signals and / or potential fault point information, identify the fault candidate area of ​​the production equipment.

[0238] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0239] Check if the fault information is a new detection task;

[0240] If it is a new detection task, a collaborative detection task request is sent to the cloud. The collaborative detection request contains the full sample data of the fault candidate region.

[0241] If it is not a new detection task, a collaborative detection task request is sent to the cloud, which includes the consistency detection results.

[0242] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0243] For new detection tasks, data transmission strategies should be formulated based on the urgency of the candidate fault areas and bandwidth resources.

[0244] Based on the data transmission strategy, all sample data in the collaborative detection task request is transmitted.

[0245] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0246] If the urgency of a new detection task exceeds the urgency threshold, the full amount of sample data will be transmitted in real time.

[0247] If the urgency of a new detection task is below the urgency threshold and bandwidth resources are limited, transmit the full amount of sample data in batches multiple times.

[0248] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0249] Retrieve data related to the fault detection results from the local knowledge base;

[0250] The fault detection results are compared with the information related to the fault detection results to analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

[0251] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0252] If the comparison results are inconsistent, the collaborative detection task will be restarted.

[0253] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0254] The first model is used to perform initial fault detection on the sample data of the production equipment to obtain fault information of the production equipment.

[0255] Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment.

[0256] The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

[0257] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0258] Raw data from production equipment is collected by sensors and processed to obtain sample data.

[0259] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0260] The first model is used to perform initial fault detection on sample data of production equipment to identify abnormal signals and / or potential fault point information in the sample data.

[0261] Based on abnormal signals and / or potential fault point information, identify the fault candidate area of ​​the production equipment.

[0262] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0263] Check if the fault information is a new detection task;

[0264] If it is a new detection task, a collaborative detection task request is sent to the cloud. The collaborative detection request contains the full sample data of the fault candidate region.

[0265] If it is not a new detection task, a collaborative detection task request is sent to the cloud, which includes the consistency detection results.

[0266] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0267] For new detection tasks, data transmission strategies should be formulated based on the urgency of the candidate fault areas and bandwidth resources.

[0268] Based on the data transmission strategy, all sample data in the collaborative detection task request is transmitted.

[0269] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0270] If the urgency of a new detection task exceeds the urgency threshold, the full amount of sample data will be transmitted in real time.

[0271] If the urgency of a new detection task is below the urgency threshold and bandwidth resources are limited, transmit the full amount of sample data in batches multiple times.

[0272] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0273] Retrieve data related to the fault detection results from the local knowledge base;

[0274] The fault detection results are compared with the information related to the fault detection results to analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

[0275] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0276] If the comparison results are inconsistent, the collaborative detection task will be restarted.

[0277] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0278] In one exemplary embodiment, a communication device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 16As shown, the communication device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements an image recognition method. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the communication device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the communication device, or external keyboards, touchpads, or mice, etc.

[0279] Those skilled in the art will understand that Figure 16 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the communication device to which the present application is applied. Specific communication devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0280] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0281] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0282] The above embodiments merely illustrate several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A fault detection method, characterized in that, The method includes: The first model is used to perform initial fault detection on the sample data of the production equipment to obtain the fault information of the production equipment. Based on the fault information, a collaborative detection task is initiated to the cloud; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment. The fault detection results fed back from the cloud are compared with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

2. The method according to claim 1, characterized in that, The method further includes: Raw data from the production equipment is collected by sensors and processed to obtain sample data.

3. The method according to claim 1, characterized in that, The initial fault detection of the production equipment using the first model to obtain fault information of the production equipment includes: The first model is used to perform initial fault detection on sample data of production equipment to identify abnormal signals and / or potential fault point information in the sample data. Based on the abnormal signals and / or the potential fault point information, the candidate fault area of ​​the production equipment is determined.

4. The method according to claim 3, characterized in that, The step of initiating a collaborative detection task to the cloud based on the fault information includes: Detect whether the fault information is a new detection task; If it is a new detection task, a collaborative detection task request is sent to the cloud, and the collaborative detection request contains the full sample data of the fault candidate region; If it is not the new detection task, a collaborative detection task request is sent to the cloud, and the collaborative detection request includes the consistency detection result.

5. The method according to claim 4, characterized in that, If it is a new detection task, a collaborative detection task request is sent to the cloud, including: If it is a new detection task, a data transmission strategy shall be formulated based on the urgency and bandwidth resources corresponding to the candidate fault regions; Based on the data transmission strategy, all sample data in the collaborative detection task request is transmitted.

6. The method according to claim 5, characterized in that, The transmission of all sample data in the collaborative detection task request based on the data transmission strategy includes: If the urgency of the new detection task exceeds the urgency threshold, the full amount of sample data is transmitted in real time. If the urgency of the new detection task is lower than the urgency threshold and the bandwidth resources are limited, the full amount of sample data will be transmitted in batches multiple times.

7. The method according to claim 1, characterized in that, The step of comparing the fault detection results fed back from the cloud with information related to the fault detection results in the local knowledge base includes: Retrieve data related to the fault detection results from the local knowledge base; The fault detection results are compared with the information related to the fault detection results to analyze the degree of consistency between the fault detection results and the information related to the fault detection results.

8. The method according to claim 1 or 7, characterized in that, The method further includes: If the comparison results are inconsistent, the collaborative detection task will be re-initiated.

9. The method according to claim 2, characterized in that, The sensors collect raw data from the production equipment, including vibration signals, temperature, humidity, gas concentration, pressure, liquid flow rate, rotational speed, current, voltage, and video.

10. The method according to claim 2, characterized in that, Processing the raw data includes: data cleaning, integration, and transformation.

11. A fault detection method, characterized in that, The method includes: In response to a collaborative detection task initiated by the terminal, various detection methods are used to detect global information related to the collaborative detection task from the global knowledge base; The second model is used to combine the global information related to the collaborative detection task with the sample data of the fault candidate region to generate a fault detection result, and the fault detection result is sent to the terminal.

12. The method according to claim 11, characterized in that, The step of combining global information related to the collaborative detection task with sample data of the fault candidate region using the second model to generate a fault detection result, and sending the fault detection result to the terminal, includes: The sample data of the fault candidate region is integrated, filtered, and rearranged with the global information related to the collaborative detection task to obtain the processed fault detection task problem. The fault detection task problem is input into the second model to generate fault detection results; The fault detection results and the global information are sent to the terminal together.

13. A fault detection device, characterized in that, The device includes: The detection module is used to perform initial fault detection on the sample data of the production equipment using the first model to obtain the fault information of the production equipment. The task initiation module is used to initiate a collaborative detection task to the cloud based on the fault information; the collaborative detection task is used to instruct the cloud to perform fault detection on the production equipment. The update module is used to compare the fault detection results fed back from the cloud with the information related to the fault detection results in the local knowledge base. When the comparison results are consistent, the newly added fault detection results are updated to the local knowledge base.

14. A communication device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 12.

15. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

16. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 12.

Images