Remote fault diagnosis system for breathing machine

Abstract

The invention relates to the technical field of breathing machines, and discloses a breathing machine remote fault diagnosis system which comprises a back-end server used for receiving and analyzing a breathing machine log file and recording fault information; the front-end display platform is connected with the rear-end server; and the AR glasses equipment is used for recording the fault picture and video of the breathing machine and receiving the fault solution sent by the back-end server. According to the remote fault diagnosis system for the breathing machine, the log file of the breathing machine is actively collected through the box, the fault is analyzed and recorded from the log file, a corresponding fault display interface is provided for the front end, and after the front end clicks fault details, a corresponding solution is generated by the intelligent maintenance knowledge base and is displayed to the PC end in an image-text mode. In addition, on-site personnel can wear AR glasses to record related pictures and videos of breathing machine faults, the back end reads and analyzes the content of key frames in the pictures, and related fault solutions are displayed to the AR glasses mobile terminal.

Description

Technical Field

[0001] The invention relates to the technical field of ventilators, and in particular to a remote fault diagnosis system for ventilators. Background Art

[0002] As a key component of modern medical equipment, ventilators are widely used in the treatment and assisted breathing of various respiratory diseases. It mechanically provides patients with an indispensable oxygen supply and can assist or completely replace the patient's spontaneous breathing when necessary, playing a vital role in stabilizing the patient's vital signs and promoting recovery. However, with the rapid development of medical technology, the design of ventilators has become increasingly complex, integrating sophisticated sensors, advanced controllers, and efficient actuators to achieve a more precise and personalized treatment experience.

[0003] But at the same time, this high degree of technical integration also brings new challenges to the fault diagnosis and maintenance of ventilators. Traditional ventilator fault diagnosis solutions highlight significant efficiency bottlenecks:

[0004] This is mainly reflected in the fact that the search methods for fault diagnosis are relatively simple and limited. The current process often relies on engineers to describe the fault phenomenon in words and then search in maintenance manuals or related materials. This method is not only inefficient, but also easily affected by human description accuracy and understanding differences. It may even lead to the omission of key information due to improper selection of description words, and thus fail to accurately locate and find relevant solutions. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention proposes a remote fault diagnosis system for a ventilator to solve the above technical problems.

[0006] A remote fault diagnosis system for a ventilator, comprising:

[0007] The back-end server is used to receive and parse the ventilator log files, record fault information, and provide a fault display interface;

[0008] The front-end display platform is connected to the back-end server to display fault details and solutions generated by the intelligent maintenance knowledge base;

[0009] AR glasses device, used to record ventilator failure images and videos, and receive fault solutions sent by the back-end server

[0010] Furthermore, the back-end server also includes:

[0011] The log collection module is used to interact with the ventilator through the HL7 protocol and actively collect log files;

[0012] Log parsing module, used to parse the fault codes in the log files and convert them into text descriptions of the fault phenomena;

[0013] The database storage module is used to save the parsed fault information into the database.

[0014] Furthermore, the front-end display platform is built using the Vue framework.

[0015] Furthermore, the AR glasses device also includes:

[0016] Video recording module, used to record ventilator failure images and videos;

[0017] Real-time transmission module, used to transmit recorded video and image data to the back-end server in real time;

[0018] The receiving and displaying module is used to receive and display the fault solution sent by the back-end server.

[0019] Furthermore, it also includes an intelligent maintenance knowledge base, which is used to generate corresponding solutions according to the fault phenomenon and display them in the form of pictures and texts on the front-end display platform.

[0020] Furthermore, the intelligent maintenance knowledge base includes:

[0021] A storage module, used to convert documents into vectors and store them in a vector database;

[0022] The retrieval module is used to calculate the cosine similarity based on the input vector and find the most similar document in the library;

[0023] The generation module is used to generate solutions based on the most similar documents using the GPT-3 model.

[0024] Furthermore, the back-end server also includes an image recognition module for identifying images transmitted by the AR glasses device, determining fault phenomena, and generating corresponding solutions.

[0025] Furthermore, the image recognition module uses the YOLOv8 algorithm to perform target detection and identify faulty components or phenomena in the image.

[0026] Furthermore, the back-end server also includes an Nginx static server for providing HTTP and Web reverse proxy functions, solving AJAX cross-domain problems, and optimizing the access speed of the front-end display platform.

[0027] Furthermore, it also includes a security module for encrypting the data in the system to ensure the security and privacy of the data during transmission and storage.

[0028] The invention adopting the above technical solution has the following advantages:

[0029] The present invention actively collects the log files of the ventilator through the box, parses and records the faults from the log files, and provides a corresponding fault display interface for the front end. After the front end clicks on the fault details, the intelligent maintenance knowledge base will generate the corresponding solution and display it to the PC in the form of pictures and texts. In addition, the on-site personnel can wear AR glasses to record the relevant screen pictures and videos of the ventilator fault, and the back end can read and analyze the content of the key frames in the pictures, and display the relevant fault solutions to the AR glasses mobile terminal. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the specific implementation of the present invention, the following will briefly introduce the drawings required for use in the specific implementation. In all the drawings, each element or part is not necessarily drawn according to the actual scale.

[0031] Figure 1 A flow chart of a remote fault diagnosis system for a ventilator according to the present invention;

[0032] Figure 2 This is an architecture diagram of a remote fault diagnosis system for a ventilator according to the present invention;

[0033] Figure 3 This is an architecture diagram of the front end of a remote fault diagnosis system for a ventilator according to the present invention. DETAILED DESCRIPTION

[0034] The following embodiments of the technical solution of the present invention are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and are therefore only used as examples, and cannot be used to limit the protection scope of the present invention.

[0035] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should be the common meanings understood by technicians in the field to which the present invention belongs. The terms "first", "second", etc. in the specification and claims of the embodiments of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so as to implement the embodiments of the present disclosure described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. Unless otherwise specified, the term "multiple" means two or more. In the embodiments of the present disclosure, the character " / " indicates that the objects before and after are in an "or" relationship. For example, A / B means: A or B. The term "and / or" is a description of the association relationship of objects, indicating that there can be three relationships. For example, A and / or B means: A or B, or, A and B. The term "corresponding" can refer to an association relationship or a binding relationship, and A and B correspondingly refer to an association relationship or a binding relationship between A and B.

[0036] like Figure 1 to Figure 3 As shown, a remote fault diagnosis system for a ventilator of the present invention includes a back-end server for receiving and parsing ventilator log files, recording fault information, and providing a fault display interface; a front-end display platform connected to the back-end server for displaying fault details and solutions generated by an intelligent maintenance knowledge base; an AR glasses device for recording ventilator fault images and videos, and receiving fault solutions sent by the back-end server.

[0037] Specifically, the three main technologies used are backend, server, and frontend, and the backend mainly uses Java as the development language.

[0038] The box actively collects the log files of the ventilator, analyzes and records the faults from the log files, and provides the corresponding fault display interface for the front end. After the front end clicks on the fault details, the intelligent maintenance knowledge base will generate the corresponding solution and display it to the PC in the form of pictures and texts.

[0039] In addition, on-site personnel can wear AR glasses to record pictures and videos of relevant images of the ventilator failure. The back-end can read and analyze the content of the key frames in the pictures, and display the relevant fault solutions to the AR glasses mobile terminal to guide and help on-site personnel to solve the fault remotely.

[0040] In some embodiments, the backend server further includes:

[0041] The log collection module is used to interact with the ventilator through the HL7 protocol and actively collect log files;

[0042] Log parsing module, used to parse the fault codes in the log files and convert them into text descriptions of the fault phenomena;

[0043] The database storage module is used to save the parsed fault information into the database.

[0044] Specifically, collect ventilator logs:

[0045] The system actively collects relevant logs of the ventilator through the box. Including time, equipment model, fault code, usage information, etc. The box and the large equipment exchange data through the HL7 Health Information Exchange Standard (Health Level 7) protocol (hl7 will be briefly introduced below). The log files of the equipment will be stored in a path. The code reads the device log file path and then obtains the log file list. After successfully reading the file by sorting the date, it reads the text line by line in the form of a file stream. Abnormal log content will contain specific error codes. Different code values ​​represent different fault phenomena. Based on this as a starting point, you can get text descriptions related to the exception. After that, generate a SQL statement and save the content to the database.

[0046] Parsing log content:

[0047] After the backend receives the logs collected by the box, it compares them with the HL7 protocol document provided by the manufacturer, identifies the fault codes through regular expression matching, and converts the fault codes into textual descriptions of the fault phenomena by comparing them with the document.

[0048] A regular expression (regex or regexp for short) is a pattern used to match a series of character combinations in a string. It is usually used to search, match, and replace specific patterns in text.

[0049] In some embodiments, the AR glasses device further includes:

[0050] Video recording module, used to record ventilator failure images and videos;

[0051] Real-time transmission module, used to transmit recorded video and image data to the back-end server in real time;

[0052] The receiving and displaying module is used to receive and display the fault solution sent by the back-end server.

[0053] Specifically, AR glasses remotely:

[0054] On-site engineers can wear AR glasses to record and transmit the fault images and videos of the ventilator, and call the back-end interface to obtain the corresponding fault solutions.

[0055] In some embodiments, the back-end server also includes an image recognition module for identifying images transmitted by the AR glasses device, determining fault phenomena, and generating corresponding solutions.

[0056] In some embodiments, the image recognition module uses the YOLOv8 algorithm to perform target detection and identify faulty components or phenomena in the image.

[0057] Specifically, the image recognition model identifies the cause of the fault in the remote image:

[0058] Based on AR remote video, on-site engineers can capture the image at that time and send it back to the server. The server calls the AI ​​image recognition model to identify the returned image and determine the fault phenomenon. The overall process diagram is as follows:

[0059] The ventilator remote fault diagnosis AI model was trained using the YOLOv8 algorithm. The YOLO (You Only LookOnce) series is a widely used target detection model in the field of computer vision and has received widespread attention for its efficient real-time detection capabilities. YOLOv8 combines the advantages of previous generations of models and has been improved in many aspects, significantly improving its accuracy and speed.

[0060] YOLOv8 has been improved in the following aspects: Improved backbone network: A more powerful backbone network architecture, such as CSPDarknet53, is adopted, so that the model can capture more feature information while maintaining high efficiency. Efficient Feature Pyramid Network (FPN): The feature pyramid network is optimized to improve the detection ability of small targets. Optimized anchor box mechanism: Dynamically adjust the size and shape of the anchor box so that the model can predict the target position more accurately. Advanced data enhancement technology: Data enhancement technologies such as CutMix, MixUp, and Mosaic are introduced to enable the model to better generalize to different scenarios. Improved loss function: A more complex loss function, such as CIoU (Complete Intersection over Union), is adopted to improve the convergence speed and detection accuracy of the model. Mixed precision training: Supports mixed precision training, which accelerates the training process and reduces video memory usage while ensuring model accuracy.

[0061] In some embodiments, an intelligent maintenance knowledge base is also included, which is used to generate corresponding solutions according to the fault phenomenon and display them in the form of pictures and texts on the front-end display platform.

[0062] In some embodiments, the intelligent maintenance knowledge base includes:

[0063] A storage module, used to convert documents into vectors and store them in a vector database;

[0064] The retrieval module is used to calculate the cosine similarity based on the input vector and find the most similar document in the library;

[0065] The generation module is used to generate solutions based on the most similar documents using the GPT-3 model.

[0066] Specifically, maintenance solutions are provided based on the intelligent maintenance knowledge base:

[0067] The intelligent maintenance knowledge base consists of three modules: storage module, retrieval module and generation module.

[0068] The storage module mainly converts documents into vectors and then stores them in the milvus vector database. The algorithm used here to convert documents into vectors is Distributed Bag of Words (DBOW).

[0069] In DBOW mode, the goal of the model is to predict random words in a document based only on the document vector. The basic idea is to learn a vector representation that captures the semantics of the document so that the distribution of words in the document can be well predicted based on this vector alone.

[0070] Mathematically, suppose we have a vocabulary V of size |V| and a document D represented as a series of words w1,w2,...,wn. Our goal is to learn a document vector d that maximizes the probability of predicting the occurrence of a word wk randomly sampled from document D.

[0071] Assuming we have a function f(d) to represent the probability distribution of predicted words based on the document vector d, then for document D, our optimization goal can be expressed as:

[0072]

[0073] Among them, P(w k |d) The softmax function is usually used to calculate:

[0074]

[0075] where v i is the word w in the vocabulary i The corresponding vector representation.

[0076] During the training process, the document vector d is updated through optimization algorithms such as stochastic gradient descent (SGD) to minimize the loss function (usually the negative log-likelihood), thereby learning a vector that can better represent the semantics of the document.

[0077] The retrieval module obtains the document in the library that is most similar to the input vector by calculating the cosine distance between document vectors.

[0078] Assuming there are two document vectors A = [a1, a2, a3, ..., an] and B = [b1, b2, b3, ..., bn], their cosine distance can be calculated by the following formula:

[0079]

[0080] in,

[0081]

[0082]

[0083]

[0084] The final cos(θ) is the cosine similarity of the two document vectors, and its value is between 0 and 1. The closer the value is to 1, the closer the directions of the two vectors are, that is, the more similar the two documents are; the closer the value is to 0, the greater the difference in the directions of the two vectors is, that is, the greater the difference between the two documents; the value close to 0 means that the two vectors are almost orthogonal, that is, the two documents are not very related.

[0085] If you want to get the cosine distance, it is usually defined as 1-cos(θ), so that the distance value is between 0 and 1, and the larger the value, the greater the difference between the two documents.

[0086] After obtaining the most similar n (usually 3) documents in the database, we will use the GPT-3 model to generate answers from these n documents based on the question. This is the generation module.

[0087] In some embodiments, the back-end server also includes an Nginx static server for providing HTTP and Web reverse proxy functions, solving AJAX cross-domain issues, and optimizing the access speed of the front-end display platform.

[0088] In some embodiments, a security module is also included for encrypting data in the system to ensure the security and privacy of the data during transmission and storage.

[0089] server

[0090] The server uses Nginx as a static server, providing high-performance HTTP and Web reverse proxy functions, and solving AJAX cross-domain issues.

[0091] front end

[0092] In some embodiments, the front-end display platform is built using the Vue framework.

[0093] Specifically, the front-end uses the Vue framework to display text or pictures and other content such as fault solutions in a single-page platform application. The main principle is to obtain the fault code, troubleshooting process, solution and other content of the ventilator fault according to the search interface provided by the back-end. Scrolling the mouse down the page can fully display the content. Click the picture to view the large picture.

[0094] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A remote fault diagnosis system for a ventilator, characterized in that: include: The back-end server is used to receive and parse the ventilator log files, record fault information, and provide a fault display interface; The front-end display platform is connected to the back-end server to display fault details and solutions generated by the intelligent maintenance knowledge base; AR glasses device, used to record ventilator failure images and videos, and receive fault solutions sent by the back-end server.

2. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: The back-end server also includes: The log collection module is used to interact with the ventilator through the HL7 protocol and actively collect log files; Log parsing module, used to parse the fault codes in the log files and convert them into text descriptions of the fault phenomena; The database storage module is used to save the parsed fault information into the database.

3. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: The front-end display platform is built using the Vue framework.

4. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: The AR glasses device also includes: Video recording module, used to record ventilator failure images and videos; Real-time transmission module, used to transmit recorded video and image data to the back-end server in real time; The receiving and displaying module is used to receive and display the fault solution sent by the back-end server.

5. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: It also includes an intelligent maintenance knowledge base, which is used to generate corresponding solutions based on the fault symptoms and display them in the form of graphics and text on the front-end display platform.

6. The remote fault diagnosis system for a ventilator according to claim 5, characterized in that: The intelligent maintenance knowledge base includes: A storage module, used to convert documents into vectors and store them in a vector database; The retrieval module is used to calculate the cosine similarity based on the input vector and find the most similar document in the library; The generation module is used to generate solutions based on the most similar documents using the GPT-3 model.

7. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: The back-end server also includes an image recognition module for recognizing images transmitted by the AR glasses device, determining fault phenomena, and generating corresponding solutions.

8. The remote fault diagnosis system for a ventilator according to claim 7, characterized in that: The image recognition module uses the YOLOv8 algorithm to perform target detection and identify faulty components or phenomena in the image.

9. The remote fault diagnosis system for a ventilator according to claim 1, characterized in that: The back-end server also includes an Nginx static server, which is used to provide HTTP and Web reverse proxy functions, solve AJAX cross-domain problems, and optimize the access speed of the front-end display platform.

10. The remote fault diagnosis system for a ventilator according to any one of claims 1 to 9, characterized in that: It also includes a security module for encrypting data in the system to ensure the security and privacy of data during transmission and storage.

Images