Fault diagnosis method for end-side cloud cooperative equipment based on micro-service architecture

Abstract

The invention provides an end-side cloud cooperative equipment fault diagnosis method based on a micro-service architecture, which comprises the four steps of building an equipment fault detection platform, training and issuing a diagnosis model, and collecting data and diagnosing faults, and is convenient for workers to operate by building the equipment fault detection platform, thereby improving the fault diagnosis efficiency. According to the end-side cloud cooperative equipment fault diagnosis method based on the micro-service architecture, the cloud processing center only needs to be responsible for establishment and training of a fault diagnosis model by adopting an end-side cloud cooperative mode, so that the fault diagnosis efficiency is improved, and the fault diagnosis efficiency is improved. The computing task is issued to the edge processing center, that is, the equipment fault diagnosis task is segmented, fault diagnosis is performed on the equipment through an end-side cloud cooperation mode, then the data computing pressure of the cloud processing center is relieved, and the edge processing center is closer to the equipment relative to the cloud processing center, so that the delay of data response is effectively reduced, and the data computing efficiency is improved. Therefore, the fault diagnosis quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to an edge-cloud collaborative device fault diagnosis method based on a microservices architecture. Background Art

[0002] The microservices architecture is a new technology for deploying applications and services in the cloud. Microservices can run in "their own programs" and communicate through "lightweight devices and HTTP-based APIs". The key is that the service can run in its own program. Through this, we can distinguish the public service from the microservices architecture (distributing an API in the existing system). In the public service, many services can be restricted by internal independent processes. If any of these services needs to add a certain function, then the process scope must be reduced. In the microservices architecture, only the required function needs to be added to a specific service without affecting the overall process architecture;

[0003] From the perspective of data, fault diagnosis is also known as fault detection, also called deviation detection or outlier detection. From the perspective of data, fault detection first analyzes the relationships between relevant feature data, mines useful value information, learns the laws between data information, then detects a special observation value that is very different from many other observed values, and finally determines whether the special observation value is fault data;

[0004] With the development of the Internet of Things, various intelligent devices have emerged. As the number of intelligent devices increases, they will generate a large amount of business data. At the same time, as the location distribution of the devices becomes more extensive, for the cloud computing center, when it needs to perform fault diagnosis on a certain device, its burden is heavier. It not only requires high computing performance, but also needs to have better network bandwidth and lower network latency. It is prone to various problems when facing different situations, such as low data transmission efficiency, etc. Therefore, the present invention proposes an edge-cloud collaborative device fault diagnosis method based on a microservices architecture to solve the problems existing in the prior art. Summary of the Invention

[0005] Aiming at the above problems, the purpose of the present invention is to propose an edge-cloud collaborative device fault diagnosis method based on a microservices architecture. This edge-cloud collaborative device fault diagnosis method based on a microservices architecture has the advantage of good processing effect and solves the problems in the prior art.

[0006] To achieve the purpose of the present invention, the present invention is realized through the following technical solutions: An edge-cloud collaborative device fault diagnosis method based on a microservices architecture, including the following steps:

[0007] Step 1: Build a device fault detection platform

[0008] First, based on the microservices architecture, an equipment fault monitoring platform is built. The equipment fault monitoring platform includes a front-end layer, an edge layer, a cloud center layer, and a service layer. The front-end layer includes a mobile end and a Web end. The edge layer includes an edge processing center. The cloud center layer includes a cloud processing center. The service layer includes a data access end, a data collection end, and a data temporary storage end. At the same time, a wireless communication solution is adopted for data transmission between the various layers within the equipment fault monitoring platform;

[0009] Step 2: Diagnosis model training and distribution

[0010] The cloud processing center in Step 1 constructs a fault detection model and iteratively trains the constructed fault diagnosis model. After continuous iterative training, a trained fault detection model can be obtained, and then the fault detection model is deployed into the edge processing center;

[0011] Step 3: Data collection

[0012] The cloud processing center issues a fault detection requirement for the equipment to the edge processing center. After receiving the corresponding requirement instruction, the edge processing center collects data on the equipment specified in the requirement instruction through the data access end. The collected data is processed by the edge processing center to obtain data related to equipment faults;

[0013] Step 4: Fault diagnosis

[0014] The edge processing center inputs the data related to equipment faults obtained into the fault detection model deployed in the edge processing center. The edge processing center conducts fault detection and diagnosis on the equipment and uploads the diagnosis result to the cloud processing center, and the cloud processing center gives feedback.

[0015] The further improvement lies in: in Step 1, the edge processing center includes a data processing module, a data receiving module, a model diagnosis module, and a data storage module.

[0016] The further improvement lies in: in Step 1, the cloud processing center includes a model training module, a distributed database, and a cloud center processing module.

[0017] The further improvement lies in: in Step 1, the method for training the fault diagnosis model includes the following steps:

[0018] S1: Collect and process the data of the equipment through the edge processing center and upload the processed data to the cloud processing center;

[0019] S2: Divide the processed data into two parts, namely a training set and a test set;

[0020] S3: Use the data in the training set to train the fault diagnosis model, and then use the test set to test the trained fault diagnosis model.

[0021] A further improvement is that in step 2, since the types of devices are different, the data collected by them also vary. Therefore, the access terminal devices are classified, and corresponding fault diagnosis model training is performed for the device data of different categories.

[0022] A further improvement is that in step 3, a data processing module, a feature extraction module, and a normalization module are provided in the edge processing center. The data processing module preliminarily processes the collected device data, that is, deletes the redundant data in the data. Then, the feature extraction module extracts the feature data related to device faults, and then the normalization module normalizes the extracted data.

[0023] A further improvement is that in step 4, a device classification and recognition module is also provided in the edge processing center. The device classification and recognition module is used to identify the classification to which the device data belongs, and the corresponding fault diagnosis model is called according to this classification.

[0024] A further improvement is that in step 4, a result comparison module is also provided in the cloud processing center. A grading table corresponding to the output result of the fault diagnosis model is preset in the result comparison module, and the fault state of the current device is determined according to the result compared by the result comparison module.

[0025] The beneficial effects of the present invention are as follows: This end-edge-cloud collaborative device fault diagnosis method based on a microservices architecture divides the device fault diagnosis task by adopting the end-edge-cloud collaborative method, enabling the cloud processing center to only be responsible for the establishment and training of the fault diagnosis model, and sending its computing tasks to the edge processing center. By means of end-edge-cloud collaboration, the device is fault-diagnosed, thereby reducing the data computing pressure on the cloud processing center. Moreover, since the edge processing center is closer to the device than the cloud processing center, the data response delay is effectively reduced, improving the quality of fault diagnosis. Description of the Drawings

[0026] Figure 1 It is a step schematic diagram of Embodiment 1 of the present invention.

[0027] Figure 2 It is a schematic structural diagram of the device fault detection platform in Embodiment 1 of the present invention. Detailed Embodiment

[0028] To deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with embodiments. These embodiments are only used to explain the present invention and do not constitute a limitation to the protection scope of the present invention.

[0029] According to Figure 1 - Figure 2 As shown, this embodiment proposes a device fault diagnosis method based on a microservices architecture, including the following steps:

[0030] Step 1: Build a device fault detection platform

[0031] First, based on the microservices architecture, build a device fault monitoring platform. Building the device fault monitoring platform is mainly to facilitate staff to log in to the platform and operate on the device fault diagnosis. That is, by the way of manual login to the platform for operation, send the requirement for the fault diagnosis of a certain device. The device fault monitoring platform includes a front-end layer, an edge layer, a cloud center layer, and a service layer. The front-end layer includes a mobile end and a Web end. The mobile end is convenient for staff to log in through mobile devices, and the Web end is convenient for staff to access in the form of a website through computer devices or tablet devices. At the same time, when staff access through the mobile end and the Web end, they need to perform login verification. The edge layer includes an edge processing center, and the edge processing center includes a data processing module, a data receiving module, a model diagnosis module, and a data storage module. The cloud center layer includes a cloud processing center, and the cloud processing center includes a model training module, a distributed database, and a cloud center processing module. The service layer includes a data access end, a data collection end, and a data temporary storage end. The service layer is mainly to provide corresponding data access services for the edge processing center. Thus, the edge processing center can collect device data through the data collection end. Among them, the device data includes, but is not limited to, the device number, type, voltage, current, temperature, humidity, working status, timestamp, signal strength, etc. At the same time, a wireless communication solution is adopted for data transmission between each layer in the device fault monitoring platform. Among them, the data temporary storage end mainly provides a data temporary storage function, that is, it itself is a database. When the edge processing center collects data, it is stored by the data temporary storage end. Since it is a temporary storage end, the retention period of its data can be set according to requirements;

[0032] Step 2: Train and distribute the diagnosis model

[0033] The cloud processing center in Step 1 constructs a fault detection model and iteratively trains the constructed fault diagnosis model. After continuous iterative training, a trained fault detection model can be obtained. Then, this fault detection model is deployed to the edge processing center. Since the types of devices are different, the collected data also varies. Subsequently, the accessed terminal devices are classified, and corresponding fault diagnosis model training is carried out for the device data of different categories. That is, a targeted approach is adopted. For example, the devices are divided into two different types, and corresponding fault diagnosis models are constructed for these two different types of devices respectively. Thus, in terms of data processing and model construction, the pressure on the edge center and the cloud processing center can be reduced. Since the electromechanical device data has multidimensionality and diversity, its fault forms and the data related to faults are also different. Subsequently, constructing a fault diagnosis model suitable for all devices will increase the pressure on the cloud processing center. Therefore, a targeted approach is adopted to call the corresponding fault diagnosis model for the device data of the same category, that is, the devices of the same category are grouped together, and there are corresponding fault diagnosis models for different types of devices;

[0034] Since the types of devices are different, the collected data also varies. Subsequently, the accessed terminal devices are classified, and corresponding fault diagnosis model training is carried out for the device data of different categories;

[0035] Step 3: Data collection

[0036] The cloud processing center issues a device fault detection requirement to the edge processing center. After receiving the corresponding requirement instruction, the edge processing center collects data on the devices specified in the requirement instruction through the data access end. That is, the collected data is processed by the edge processing center to obtain data related to device faults. The edge processing center is equipped with a data processing module, a feature extraction module, and a standardization module. The data processing module preliminarily processes the collected device data, that is, deletes the redundant data in the data. Then, the feature extraction module extracts the feature data related to device faults, and afterwards, the standardization module standardizes the extracted data;

[0037] Step 4: Fault diagnosis

[0038] The edge processing center inputs the data related to equipment failures into the fault detection model deployed in the edge processing center, and the edge processing center conducts fault detection and diagnosis on the equipment, and uploads the diagnosis results to the cloud processing center, and the cloud processing center gives feedback. There is also an equipment classification and identification module in the edge processing center, which is used to identify the classification to which the equipment data belongs, and call the corresponding fault diagnosis model according to this classification. There is also a result comparison module in the cloud processing center. A grading table corresponding to the output result of the fault diagnosis model is preset in the result comparison module, and the fault state of the current equipment is determined according to the result compared by the result comparison module.

[0039] In step one, the method for training the fault diagnosis model includes the following steps:

[0040] S1: The edge processing center collects and processes the data of the equipment, and uploads the processed data to the cloud processing center;

[0041] S2: Divide the processed data into two parts, namely the training set and the test set;

[0042] S3: Use the data in the training set to train the fault diagnosis model, and then use the test set to test the trained fault diagnosis model;

[0043] Through continuous training and testing, an optimal result can finally be obtained, and this optimal result is the trained fault diagnosis model.

[0044] Embodiment 2

[0045] The main difference between this embodiment and Embodiment 1 is that multiple groups of edge processing centers are set in the edge layer. Multiple groups of edge processing centers access and manage the equipment in different regions in batches, and then multiple groups of edge processing centers are connected to the cloud processing center through a wireless communication network. Then, in different regions, corresponding edge processing centers can be set. When the staff issues a demand for equipment diagnosis, the cloud processing center sends a diagnosis instruction to the edge processing center in the corresponding region. For the method of determining which region the equipment to be diagnosed belongs to, after the equipment is accessed, the information data of the equipment and the corresponding edge processing center need to be recorded in the cloud processing center, and a corresponding regional table is established. When a new device is added and accessed, the corresponding regional table will add the relevant information of the new device.

[0046] The foregoing has shown and described the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and what is described in the above embodiments and the specification is only to illustrate the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An edge-cloud collaborative device fault diagnosis method based on a microservices architecture, characterized in that: It includes the following steps: Step 1: Build a device fault detection platform First, based on the microservice architecture, build a device fault monitoring platform. This device fault monitoring platform includes a front-end layer, an edge layer, a cloud center layer, and a service layer. The front-end layer includes a mobile end and a Web end. The edge layer includes an edge processing center. The cloud center layer includes a cloud processing center. The service layer includes a data access end, a data collection end, and a data temporary storage end. At the same time, a wireless communication solution is adopted for data transmission between each layer in the device fault monitoring platform; Step 2: Train and distribute the diagnostic model The cloud processing center in Step 1 constructs a fault detection model and iteratively trains the constructed fault diagnosis model. After continuous iterative training, a trained fault detection model can be obtained, and then this fault detection model is deployed into the edge processing center; Step 3: Data collection The cloud processing center issues a device fault detection requirement to the edge processing center. After the edge processing center receives the corresponding requirement instruction, it collects data on the device specified in the requirement instruction through the data access end. The collected data is processed by the edge processing center to obtain data related to device faults; Step 4: Fault diagnosis The edge processing center inputs the data related to device faults obtained into the fault detection model deployed in the edge processing center. The edge processing center performs fault detection and diagnosis on the device and uploads the diagnosis result to the cloud processing center, and the cloud processing center gives feedback.

2. A method for diagnosing equipment faults in end-edge-cloud collaboration based on a microservices architecture according to claim 1, characterized in that: In Step 1, the edge processing center includes a data processing module, a data receiving module, a model diagnosis module, and a data storage module.

3. A method for diagnosing equipment faults in edge-cloud collaboration based on a microservices architecture according to claim 1, characterized in that: In Step 1, the cloud processing center includes a model training module, a distributed database, and a cloud center processing module.

4. A method for diagnosing equipment faults in edge-cloud-end collaboration based on a microservice architecture according to claim 1, characterized in that: In Step 1, the method for training the fault diagnosis model includes the following steps: S1: The edge processing center collects and processes the device data and uploads the processed data to the cloud processing center; S2: Divide the processed data into two parts, namely a training set and a test set; S3: Use the data in the training set to train the fault diagnosis model, and then use the test set to test the trained fault diagnosis model.

5. A method for diagnosing equipment faults in end-edge-cloud collaboration based on a microservices architecture according to claim 1, characterized in that: In Step 2, due to different types of devices, the collected data also has differences. Then, the accessed terminal devices are classified, and corresponding fault diagnosis model training is carried out for the device data of different categories.

6. A method for diagnosing equipment faults in end-edge-cloud collaboration based on a microservices architecture according to claim 1, characterized in that: In Step 3, a data processing module, a feature extraction module, and a standardization module are provided in the edge processing center. The data processing module preliminarily processes the collected device data, that is, deletes the redundant data in the data. Then, the feature extraction module extracts the feature data related to device faults, and then the standardization module standardizes the extracted data.

7. A method for diagnosing equipment faults in edge-cloud collaboration based on a microservices architecture according to claim 1, characterized in that: In Step 4, a device classification and recognition module is also provided in the edge processing center. The device classification and recognition module is used to identify the classification to which the device data belongs, and call the corresponding fault diagnosis model according to this classification.

8. A method for diagnosing equipment faults in edge-cloud-end collaboration based on a microservice architecture according to claim 1, characterized in that: In the fourth step, a result comparison module is also provided in the cloud processing center. A grading table corresponding to the output result of the fault diagnosis model is preset in the result comparison module, and the fault state of the current device is determined according to the result compared by the result comparison module.

Images