A fault monitoring method and device for a respiratory warming tube

Abstract

The application discloses a kind of failure monitoring method and device of breathing warming pipeline, it is related to medical equipment technical field.The method is applied to the warming breathing pipeline matched with breathing machine, the method includes: obtaining the heating current corresponding to the current of breathing warming pipeline, heating current is the heating wire current corresponding to the current of each warming pipeline unit in breathing warming pipeline;Heating current is input into the trained failure monitoring model, and the distribution probability of the monitoring result corresponding to breathing warming pipeline is obtained, the distribution probability of monitoring result includes the probability of failure and the probability of normal work;According to the distribution probability of monitoring result and the probability processing rule of pre-set, determine the failure detection result corresponding to breathing warming pipeline.In the application, the failure monitoring model is obtained by deep learning of historical data, at this time, failure detection and diagnosis can be automatically carried out based on the failure monitoring model, which greatly improves the accuracy and efficiency of detection.

Description

Technical Field

[0001] This application relates to the field of medical devices, and in particular to a method and apparatus for fault monitoring of a respiratory heating pipeline. Background Technology

[0002] Heated breathing tubing is commonly used to connect ventilators and anesthesia machines to masks or endotracheal tubes. Its purpose is to prevent condensation by heating the gas within the tubing, thus reducing patient discomfort. In practical applications, fault identification in heated tubing systems is a complex and critical task. Current technologies typically involve manually extracting the electrical signals when a fault occurs in the heated tubing, and then manually extracting statistical results from these signals, such as RMS values. Alternatively, they may manually extract frequency domain features of the current from the electrical signals to obtain feature vectors, and then use data-driven models for classification or detection.

[0003] However, extensive practical applications have revealed that these manually extracted features have subjectivity and limitations. To a certain extent, due to differences between datasets, it is difficult to further improve the accuracy of identification. Therefore, there is an urgent need for a method that can quickly and accurately identify faults in heating pipeline systems. Summary of the Invention

[0004] The purpose of this application is to address at least one of the aforementioned technical deficiencies.

[0005] On one hand, embodiments of this application provide a fault monitoring method for a warming respiratory tubing. This method is applied to a warming respiratory tubing used with a ventilator, and includes:

[0006] Obtain the current heating current corresponding to the breathing heating pipeline. The heating current is the current heating current corresponding to the heating wire of each heating pipeline unit in the breathing heating pipeline.

[0007] The heating current is input into the trained fault monitoring model to obtain the probability distribution of the monitoring results corresponding to the breathing heating pipeline. The probability distribution of the monitoring results includes the probability of a fault occurring and the probability of normal operation.

[0008] Based on the probability distribution of the monitoring results and the preset probability processing rules, the fault detection results corresponding to the breathing heating pipeline are determined.

[0009] Optionally, the fault monitoring model includes convolutional layers, a normalized activation module, an attention mechanism module, a residual module, a fully connected layer, a Dropout layer, and a Softmax layer. The normalized activation module includes a first Batch Normalization and a first ReLU activation function.

[0010] Optionally, the convolutional layer includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. Each of the first, second, and third convolutional layers is connected to a normalized activation module. The input to the first convolutional layer is the heating current, the input to the second convolutional layer is the output of the normalized activation module connected to the first convolutional layer, the input to the third convolutional layer is the output of the normalized activation module connected to the second convolutional layer, and the output of the normalized activation module connected to the third convolutional layer serves as the input to the attention mechanism module.

[0011] Optionally, the residual module includes two residual blocks. Each residual block includes a second BatchNormalization, a second ReLU activation function, a weight layer, and a third Batch Normalization, which are connected in sequence. The input of the residual module is the output of the attention mechanism module, and the output of the residual module is the input of the fully connected layer.

[0012] Optionally, the attention mechanism module includes a channel attention module and a spatial attention module. The output of the attention mechanism module is obtained in the following way:

[0013] The output features of the normalized activation module connected after the third convolutional layer are obtained and then input into the spatial attention module to obtain the first feature;

[0014] The output feature and the first feature are multiplied element by element to obtain the product feature;

[0015] The product features are normalized, and the normalized product features are input into the channel attention module to obtain the second feature;

[0016] The output of the attention mechanism module is obtained by multiplying the second feature and the normalized product feature element by element.

[0017] Optionally, the fully connected layer includes a first fully connected layer, a second fully connected layer, and a third fully connected layer connected in sequence. The input of the first fully connected layer is the output of the residual module, the output of the third fully connected layer is the input of the Dropout layer, and the output of the Dropout layer is the input of the Softmax layer.

[0018] Optionally, the fault monitoring model is trained in the following ways:

[0019] Obtain the training dataset, which includes a training sample set and a test sample set. The training dataset includes each sample pair, and each sample pair includes sample data and corresponding sample labels. The sample data includes the heating current of the heating wire of each heating pipeline unit in the breathing heating pipeline, and the sample label represents the annotation result of the corresponding fault detection result of the sample pair.

[0020] The initial neural network model is trained based on the training sample set until the loss function corresponding to the initial neural network model converges, thus obtaining the initial fault detection model. The input of the initial neural network model is a sample pair, and the output is the fault detection result corresponding to the sample pair. The value of the loss function characterizes the difference between the fault detection result corresponding to each sample pair output by the model and the labeled result corresponding to each sample pair.

[0021] The initial fault monitoring model is tested and verified based on the test sample set. If the test and verification results of the initial fault monitoring model do not meet the set verification requirements, the step of training the initial neural network model based on the training sample set is repeated until the test and verification results of the initial fault monitoring model meet the set verification requirements.

[0022] Optionally, the validation sample set includes a validation sample set and a test sample set, and the sample data includes positive sample data and negative sample data. The training dataset includes:

[0023] Acquire at least one heating current of the breathing heating tubing during normal operation as at least one negative sample data;

[0024] Acquire at least one heating current in the breathing heating tubing when a fault occurs as at least one positive sample data;

[0025] At least one negative sample data and at least one positive sample data are randomly combined and divided according to a set quantity requirement to obtain the training sample set, the validation sample set and the test sample set respectively.

[0026] Optionally, validation requirements can be set, including recall and precision requirements for each sample pair.

[0027] On the other hand, embodiments of this application provide a fault monitoring device for a warming respiratory tubing. This device is applied to a warming respiratory tubing used with a ventilator, and includes:

[0028] The current acquisition module is used to acquire the current corresponding to the breathing heating pipeline. The heating current is the current corresponding to the heating wire of each heating pipeline unit in the breathing heating pipeline.

[0029] The probability acquisition module is used to input the heating current into the trained fault monitoring model to obtain the probability distribution of the monitoring results corresponding to the breathing heating pipeline. The probability distribution of the monitoring results includes the probability of a fault occurring and the probability of normal operation.

[0030] The result determination module is used to determine the fault detection result corresponding to the breathing heating pipeline based on the distribution probability of the monitoring results and the preset probability processing rules.

[0031] In another aspect, embodiments of this application provide an electronic device, including a processor and a memory:

[0032] The memory is configured to store machine-readable instructions that, when executed by the processor, cause the processor to perform any of the methods in a fault monitoring method for a breathing heating pipeline.

[0033] The beneficial effects of the technical solutions provided in this application include at least the following:

[0034] In this embodiment, a fault monitoring model is obtained through deep learning of historical data, enabling automated fault detection and diagnosis based on the model, significantly improving detection accuracy and efficiency. Furthermore, it exhibits good scalability and adaptability, applicable to heating pipeline systems of different types and scales. Through continuous learning and model updates, this method can adapt to system changes and new fault modes, ensuring sustained high efficiency in actual operation. This allows for better control of the heating pipeline temperature, improved heating efficiency, reduced ventilator heating power consumption, and better prevention of condensation.

[0035] In this embodiment, since the fault monitoring model includes a residual module, that is, a residual learning mechanism is introduced, the gradient vanishing and gradient exploding problems in deep neural network training can be effectively solved, allowing the network to be trained deeper. It also reduces the weight layers and ReLU in the classic residual layer structure, simplifies the classic residual structure, reduces the amount of computation, and makes the model converge faster.

[0036] In this embodiment, the fault monitoring model also includes an attention mechanism module, which adopts a processing method that combines spatial attention and channel attention. This is an effective feature enhancement strategy that can improve the performance and generalization ability of the model without significantly increasing computational costs.

[0037] In this embodiment, a Dropout layer is introduced between the fully connected layer and the Softmax layer. This layer can randomly select a certain proportion of neurons in each training batch and temporarily disable their outputs, thereby reducing the network's dependence on specific neurons, preventing overfitting, and helping to improve the model's generalization ability. This allows the fault monitoring model to perform better on unseen data. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1a A schematic diagram of a breathing heating pipeline provided in an embodiment of this application;

[0040] Figure 1b A schematic diagram of a fault monitoring process for a respiratory heating pipeline provided in an embodiment of this application;

[0041] Figure 2 This is a schematic diagram of the structure of a fault monitoring model provided in an embodiment of this application;

[0042] Figure 3 This is a schematic diagram of the structure of each residual block provided in the embodiments of this application;

[0043] Figure 4 A schematic diagram of the attention mechanism module provided in an embodiment of this application;

[0044] Figure 5 This is a schematic diagram of another fault monitoring model provided in an embodiment of this application;

[0045] Figure 6 A schematic diagram of the structure of the fault monitoring device for the breathing heating pipeline provided in the embodiments of this application;

[0046] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0047] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting the invention.

[0048] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this application means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is “connected” or “coupled” to another element, it can be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein can include wireless connections or wireless coupling. The term “and / or” as used herein includes all or any units and all combinations of one or more associated listed items.

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the key proprietary terms involved in the embodiments of this application will be explained below with reference to the accompanying drawings.

[0050] Heated breathing tubing: This is commonly used to connect ventilators and anesthesia machines to masks or endotracheal tubes. Its purpose is to prevent condensation by heating the gas inside the tubing, thus reducing patient discomfort. Specifically, for example... Figure 1a As shown, a breathing heating circuit typically consists of multiple heating circuit units equipped with heating wires and temperature sensors. The heating wire of each heating circuit unit is independently connected to a temperature control unit. Below the breathing heating circuit is a temperature control unit for the ventilator and anesthetic.

[0051] The Central Processing Unit (CPU) is one of the main components of a computer, primarily responsible for interpreting computer instructions and processing data within computer software. In a broader sense, the CPU refers to a series of logical machines capable of executing complex computer programs.

[0052] Feature engineering, also known as feature extraction or feature discovery, is the process of extracting features (characteristics, attributes, properties) from raw data using domain knowledge. Its motivation, rather than simply providing raw data to the machine learning process, is to use these additional features to improve the quality of the machine learning results.

[0053] 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.

[0054] Specifically, such as Figure 1b As shown, this method can be applied to a heated breathing circuit used with a ventilator, and the method may include:

[0055] Step S101: Obtain the current heating current corresponding to the breathing heating pipeline. The heating current is the current heating current corresponding to the heating wire of each heating pipeline unit in the breathing heating pipeline.

[0056] Optionally, when monitoring whether a malfunction has occurred in the breathing heating pipeline, the current heating current corresponding to the heating wire of each heating pipeline unit in the breathing heating pipeline can be obtained, and then the malfunction of the breathing heating pipeline can be further determined based on the obtained heating current. The method of obtaining the heating current can be selected according to actual needs, and this application embodiment does not limit it; for example, the heating current can be obtained based on an installed current sensor.

[0057] Step S102: Input the heating current into the trained fault monitoring model to obtain the probability distribution of the monitoring results corresponding to the breathing heating pipeline. The probability distribution of the monitoring results includes the probability of a fault occurring and the probability of normal operation.

[0058] Optionally, after obtaining the heating current corresponding to the breathing heating pipeline, the heating current can be input into the trained fault monitoring model. At this time, the fault monitoring model will output the distribution probability of the monitoring results corresponding to the breathing heating pipeline. The distribution probability includes the probability of the breathing heating pipeline malfunctioning and the probability of the breathing heating pipeline working normally.

[0059] Step S103: Determine the fault detection result corresponding to the breathing heating pipeline based on the distribution probability of the monitoring results and the preset probability processing rules.

[0060] Furthermore, the probability distribution of the obtained monitoring results can be compared with a preset probability processing rule to determine whether a malfunction has occurred in the breathing heating pipeline. The preset probability processing rule can be set according to the required accuracy, and this embodiment does not impose any limitations. For example, if the preset probability processing rule is set such that a malfunction is considered to have occurred when the probability of a malfunction is greater than the probability of normal operation, then if the probability distribution of the obtained monitoring results shows a greater probability of a malfunction than the probability of normal operation, then the breathing heating pipeline is considered to be malfunctioning.

[0061] In this embodiment, a fault monitoring model is obtained through deep learning of historical data, enabling automated fault detection and diagnosis, significantly improving detection accuracy and efficiency. Furthermore, it exhibits good scalability and adaptability, applicable to heating pipeline systems of different types and scales. Through continuous learning and model updates, this method can adapt to system changes and new fault modes, ensuring sustained high efficiency in actual operation. This allows for better control of the heating pipeline temperature, improved heating efficiency, reduced ventilator heating power consumption, and better prevention of condensation.

[0062] In optional embodiments of this application, the fault monitoring model includes a convolutional layer, a normalized activation module, an attention mechanism module, a residual module, a fully connected layer, a Dropout layer, and a Softmax layer. The normalized activation module includes a first BatchNormalization and a first ReLU activation function.

[0063] Optionally, the structure diagram of the fault monitoring model in the embodiments of this application can be as follows: Figure 2 As shown, the structure specifically includes convolutional layers, a normalized activation module, an attention mechanism module, a residual module, a fully connected layer, a dropout layer, and a softmax layer. The normalized activation module includes a first Batch Normalization and a first ReLU activation function. The input to the convolutional layer is a heating current (i.e., input), and it is connected to the normalized activation module (i.e., Batch Normalization & ReLU). The output of the normalized activation module is connected to the attention mechanism module, the output of the attention mechanism module is connected to the residual module (also called a residual network), the output of the residual module is connected to the fully connected layer, and the dropout layer is connected to both the fully connected layer and the softmax layer.

[0064] In an optional embodiment of this application, the convolutional layer includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. Each of the first, second, and third convolutional layers is connected to a normalized activation module. The input to the first convolutional layer is a heating current. The input to the second convolutional layer is the output of the normalized activation module connected to the first convolutional layer. The input to the third convolutional layer is the output of the normalized activation module connected to the second convolutional layer. The output of the normalized activation module connected to the third convolutional layer serves as the input to the attention mechanism module.

[0065] Optionally, the convolutional layer may specifically include a first convolutional layer, a second convolutional layer, and a third convolutional layer connected in sequence. Specifically, the first convolutional layer (Conv1) has 3 input channels, 64 output channels, a 3x3 kernel size, a stride of 1, and padding of 1; the second convolutional layer (Conv2) has 64 input channels, 128 output channels, a 3x3 kernel size, a stride of 1, and padding of 1; and the third convolutional layer (Conv3) has 128 input channels, 256 output channels, a 3x3 kernel size, a stride of 1, and padding of 1.

[0066] Optionally, the first, second, and third convolutional layers are each connected to a batch normalization and ReLU module. The input to the first convolutional layer is the heating current, the input to the second convolutional layer is the output of the batch normalization and ReLU module connected to the first convolutional layer, the input to the third convolutional layer is the output of the batch normalization and ReLU module connected to the second convolutional layer, and the output of the batch normalization and ReLU module connected to the third convolutional layer is the input to the attention mechanism module.

[0067] In an optional embodiment of this application, the residual module includes two residual blocks. Each residual block includes a second batch normalization, a second ReLU activation function, a weight layer, and a third batch normalization connected in sequence. The input of the residual module is the output of the attention mechanism module, and the output of the residual module is the input of the fully connected layer.

[0068] Optionally, the input to the residual module in the fault monitoring model is the output of the attention mechanism module, and the output is the input to the fully connected layer. The residual module comprises two residual blocks, and the structure of each residual block can be specifically as follows: Figure 3 As shown, each residual block consists of sequentially connected Batch Normalization (BN in the figure), ReLU activation function (ReLU in the figure), weight layer (weight in the figure), and third Batch Normalization (BN in the figure). The input of each residual block is X. l The output is X l+1 .

[0069] In the application embodiment, since the fault monitoring model includes a residual module, i.e., a residual learning mechanism is introduced, it can effectively solve the gradient vanishing and gradient exploding problems in deep neural network training, allowing the network to be trained to be deeper. It also reduces the weight layers and ReLU in the classic residual layer structure, simplifying the classic residual structure, reducing computational cost, and allowing the model to converge faster. Specifically, it can include the following benefits:

[0070] 1. Alleviating the vanishing gradient problem: By using short-circuit connections, gradients can be directly transferred from the output to the input, thereby alleviating the vanishing gradient problem and enabling the network to be trained to a deeper level.

[0071] 2. Improve training speed: Since gradients can be directly passed, the network converges more easily, thereby improving training speed.

[0072] 3. Improve the model's generalization ability: Residual blocks can learn the direct mapping relationship between inputs and outputs, enabling the model to generalize better to unseen data.

[0073] 4. Reduce dependence on initialization: Since residual blocks improve training stability, they can reduce the network's dependence on weight initialization.

[0074] 5. Support for deeper network structures: Residual blocks enable the construction of deeper network structures, thereby improving the model's feature extraction capabilities.

[0075] 6. Improve model robustness: Residual blocks can learn the direct mapping relationship between input and output, making the model more robust to small changes in input data.

[0076] 7. Reduce overfitting: Residual blocks can improve the generalization ability of the model, thereby reducing the risk of overfitting.

[0077] In an optional embodiment of this application, the attention mechanism module includes a channel attention module and a spatial attention module, and the output of the attention mechanism module is obtained in the following manner:

[0078] The output features of the normalized activation module connected after the third convolutional layer are obtained and then input into the spatial attention module to obtain the first feature;

[0079] The output feature and the first feature are multiplied element by element to obtain the product feature;

[0080] The product features are normalized, and the normalized product features are input into the channel attention module to obtain the second feature;

[0081] The second feature and the normalized product feature are multiplied element-wise to obtain the output of the attention mechanism module.

[0082] Optionally, the attention mechanism module may include a channel attention module and a spatial attention module. Channel attention focuses on the importance of different channels, while spatial attention focuses on the importance of different spatial locations. This effectively concentrates attention on information closely related to the heating wire's heating current fault, while allocating less attention resources to information less correlated with the heating wire's heating current. This allows for differentiated attention to different frequency bands of the current signal in different contexts, thereby improving the detection sensitivity of heating wire heating current faults.

[0083] Optional, such as Figure 4 As shown, when the output feature of the normalized activation module connected after the third convolutional layer is used as the input feature to the spatial attention module, the output of the spatial attention module becomes the first feature. Then, the input feature of the spatial attention module (i.e., the original feature) and the first feature are multiplied element-wise to obtain the product feature. The product feature is then further normalized, such as by performing Softmax normalization (i.e.,...). Figure 4 The normalized multiplicative features are input into the channel attention module to obtain the second feature. Then, the second feature and the normalized multiplicative features are multiplied element-wise to obtain the output feature of the attention mechanism module. The output feature of the attention mechanism module at this time is the enhanced new feature representation.

[0084] In this embodiment, the combined spatial attention and channel attention approach is an effective feature enhancement strategy that can improve model performance and generalization ability without significantly increasing computational cost. Specifically, it offers the following benefits:

[0085] 1. Enhanced feature representation: Through the attention mechanism, the model can focus more on important features and suppress unimportant parts, thereby improving the feature representation ability.

[0086] 2. Improve model generalization ability: Attention modules can help models learn more robust features, reduce dependence on specific features, and improve the model's generalization ability under different data distributions.

[0087] 3. Reduce computational burden: Through the attention mechanism, the model can make more efficient use of computational resources and focus on the features most important to the task, thereby reducing unnecessary computation to some extent.

[0088] 4. Improved gradient flow: Attention modules can provide clearer gradient signals, which helps to better propagate gradients during training, thereby accelerating model convergence.

[0089] 5. Improve model performance: Introducing attention mechanisms has been shown to significantly improve model performance during training of many models.

[0090] 6. High adaptability: The combination of spatial attention and channel attention enables the model to pay attention to both the spatial distribution of features and the importance of channels, which allows the model to better adapt to diverse task requirements.

[0091] 7. Easy to integrate: Attention modules can usually be easily integrated into existing convolutional neural networks without requiring large-scale modifications to the network structure.

[0092] 8. Interpretability: Attention mechanisms can provide some interpretability for the model's decisions. By visualizing attention weights, we can understand which regions or features the model focuses on when making decisions.

[0093] In an optional embodiment of this application, the fully connected layer includes a first fully connected layer, a second fully connected layer, and a third fully connected layer connected in sequence. The input of the first fully connected layer is the output of the residual module, the output of the third fully connected layer is the input of the Dropout layer, and the output of the Dropout layer is the input of the Softmax layer.

[0094] Optionally, the fully connected layers in the fault monitoring module include a first fully connected layer, a second fully connected layer, and a third fully connected layer connected sequentially. The first fully connected layer (FC1) has an input dimension of 256*8*8 and an output dimension of 512. The second fully connected layer (FC2) has an input dimension of 512 and an output dimension of 256. The third fully connected layer (FC3) has an input dimension of 256 and an output dimension of 2 (used for binary classification in the Softmax layer). The input of the first fully connected layer is the output of the residual module, the output of the third fully connected layer is the input of the Dropout layer, and the output of the Dropout layer is the input of the Softmax layer.

[0095] The fault monitoring module in this embodiment introduces a Dropout layer between the fully connected layer and the Softmax layer. Its main function is to randomly "drop" (i.e., set) a portion of the neurons in the network to zero during model training. Specifically, introducing a Dropout layer between the fully connected layer and the Softmax layer allows for the random selection of a certain proportion of neurons in each training batch, temporarily disabling their output. This reduces the network's dependence on specific neurons and prevents overfitting. This approach forces the network to learn more robust feature representations because it cannot rely on the presence of any single neuron to make predictions. In this way, Dropout helps improve the model's generalization ability, enabling it to perform better on new, unseen data.

[0096] In an optional embodiment of this application, the fault monitoring model is trained in the following manner:

[0097] Obtain the training dataset, which includes a training sample set and a test sample set. The training dataset includes each sample pair, and each sample pair includes sample data and corresponding sample labels. The sample data includes the heating current of the heating wire of each heating pipeline unit in the breathing heating pipeline, and the sample label represents the annotation result of the corresponding fault detection result of the sample pair.

[0098] The initial neural network model is trained based on the training sample set until the loss function corresponding to the initial neural network model converges, thus obtaining the initial fault detection model. The input of the initial neural network model is a sample pair, and the output is the fault detection result corresponding to the sample pair. The value of the loss function characterizes the difference between the fault detection result corresponding to each sample pair output by the model and the labeled result corresponding to each sample pair.

[0099] The initial fault monitoring model is tested and verified based on the test sample set. If the test and verification results of the initial fault monitoring model do not meet the set verification requirements, the step of training the initial neural network model based on the training sample set is repeated until the test and verification results of the initial fault monitoring model meet the set verification requirements.

[0100] Optionally, this application may obtain a training sample set in the training dataset, which includes each sample pair. Each sample pair includes sample data and a corresponding sample label. The sample data refers to the heating current of the heating wire of each heating pipeline unit in the breathing heating pipeline, and the corresponding sample label is the annotation result of the fault detection result corresponding to the input sample data.

[0101] Furthermore, each sample pair can be input into the initial neural network model. This model outputs the fault detection result for each sample pair. Then, based on the fault detection result and the corresponding annotation for each sample pair, the value of the loss function for this model can be determined. If the determined loss function value does not converge, the network parameters of the initial neural network model can be adjusted. Then, each sample pair is input into the adjusted initial neural network model again, and the loss function value is determined again based on the output fault detection result and the corresponding annotation. If convergence still occurs, the network parameters of the initial neural network model are adjusted again until the corresponding loss function value converges. The value of the loss function represents the difference between the fault detection result and the corresponding annotation for each sample pair output by the model. When the loss function converges, it indicates that the difference between the fault detection result and the corresponding annotation for each sample pair output by the model meets the requirement, that is, the fault detection result of the sample pair output by the model approximates the annotation result of the sample pair.

[0102] Furthermore, when the loss function converges, an initial fault monitoring model can be obtained. At this point, the initial fault monitoring model can be tested and verified based on the test sample set in the training dataset. If the test verification result corresponding to the initial fault monitoring model does not meet the set verification requirements, the step of training the initial neural network model based on the training sample set is repeated until the test verification result corresponding to the initial fault monitoring model meets the set verification requirements.

[0103] In optional embodiments of this application, the verification requirements include the recall and accuracy requirements for each sample pair.

[0104] Optionally, the validation requirements can include recall and precision requirements for each sample pair. That is, once the recall and precision requirements for each sample pair are met, the resulting model can be used as a fault monitoring model. Here, recall refers to the proportion of correctly classified samples out of the total number of samples, while recall refers to the proportion of false positives among samples that are actually positive.

[0105] In optional embodiments of this application, the test sample set includes a validation sample set and a test sample set, the sample data includes positive sample data and negative sample data, and the training dataset is obtained by:

[0106] Acquire at least one heating current of the breathing heating tubing during normal operation as at least one negative sample data;

[0107] Acquire at least one heating current in the breathing heating tubing when a fault occurs as at least one positive sample data;

[0108] At least one negative sample data and at least one positive sample data are randomly combined and divided according to a set quantity requirement to obtain the training sample set, the validation sample set and the test sample set respectively.

[0109] Optionally, the test sample set includes a validation sample set and a test sample set, and the sample data includes positive sample data and negative sample data. In practice, at least one heating current of the breathing heating pipeline during normal operation can be obtained as at least one negative sample data, and at least one heating current of the breathing heating pipeline when a fault occurs can be obtained as at least one positive sample data. Then, the obtained at least one negative sample data and at least one positive sample data are randomly combined and divided into a training sample set, a validation sample set, and a test sample set. For example, if a total of 1000 positive and negative sample data are obtained, the dataset can be divided into 800 samples for the training sample set, 100 samples for the validation sample set, and 100 samples for the test sample set. In this application, the training dataset is obtained by randomly combining all data, which can simulate complex situations in real-world applications and helps improve the model's generalization ability.

[0110] Optionally, since heating pipeline fault monitoring is a binary classification problem, the sample data of the faulty ones are positive, and the sample data of the normally functioning ones are negative. In a binary classification problem, the samples can be classified by comparing the predicted results and the true labels of the sample data. That is, they can be divided into True Positives (TP), False Positives (FP), True Negatives (TN), and False Negatives (FN). The number of samples corresponding to each can be represented by the English abbreviation of their names. At this time, the confusion matrix of the classification results can be further obtained as shown in the table below.

[0111]

[0112] Furthermore, after obtaining the confusion matrix corresponding to the sample data, the model's performance can be evaluated periodically using the validation sample set. First, performance metrics on the validation set are calculated, such as accuracy, precision, recall, and F1 score. Then, based on the performance metrics on the validation set, the model's hyperparameters (such as learning rate, regularization parameters, and number of network layers) are adjusted. If the performance metrics on the validation set do not improve within a certain number of rounds, training can be stopped early to prevent overfitting. Finally, after model training and validation are completed, the final model is evaluated using the test sample set. The performance metrics on the test sample set are typically considered as a performance prediction of the model in practical applications. When the prediction capability meets the requirements, the final fault monitoring model can be obtained.

[0113] In this embodiment of the application, after the fault monitoring model is trained, it will be validated and evaluated based on the test sample set before the final fault monitoring model is obtained. This can help monitor the performance of the model on unseen data, thereby preventing the model from overfitting the training data. Furthermore, by using the performance indicators on the validation set, the hyperparameters of the model can be effectively adjusted to find the optimal model configuration.

[0114] Optionally, in order to better understand the method provided in the embodiments of this application, the following is based on... Figure 5 The fault monitoring model shown was trained and tested. The test data was used as input features to the convolutional layer (Conv). The output of the convolutional layer was then input to the Batch Normalization & ReLU activation function. The output and input of the Batch Normalization & ReLU activation function were sequentially connected to the attention mechanism module (Attention), the residual module (ResidualBlock), the fully connected layer (FC), and the Softmax layer to finally obtain the distribution probability of the monitoring results (output feature).

[0115] Furthermore, a comparative experiment was conducted on the heating current based on traditional methods and the method in this application. The parameters set are shown in the table below:

[0116]

[0117] Thresholding method:

[0118] Hyperparameters of the model without introducing an optimization residual module

[0119]

[0120] Hyperparameters of the model when introducing the residual optimization module

[0121]

[0122] Hyperparameters of the model when attention mechanism and residual structure are introduced

[0123]

[0124] Furthermore, the accuracy and recall rates obtained by predicting heating current based on traditional methods and the method in this application are shown in the table below:

[0125]

[0126] As can be seen from the table above, the method in this application can automatically extract multiple features at different levels without the need for manual feature design, which is more in line with the actual situation. The accuracy is higher and the false recognition rate is lower than that of the threshold method. The accuracy of the model with the introduction of residual module and attention mechanism module is also further improved.

[0127] This application provides a fault monitoring device for a breathing heating pipeline, such as... Figure 6 As shown, the device may include: a current acquisition module 601, a probability acquisition module 602, and a result determination module 603, wherein,

[0128] The current acquisition module is used to acquire the current corresponding to the breathing heating pipeline. The heating current is the current corresponding to the heating wire of each heating pipeline unit in the breathing heating pipeline.

[0129] The probability acquisition module is used to input the heating current into the trained fault monitoring model to obtain the probability distribution of the monitoring results corresponding to the breathing heating pipeline. The probability distribution of the monitoring results includes the probability of a fault occurring and the probability of normal operation.

[0130] The result determination module is used to determine the fault detection result corresponding to the breathing heating pipeline based on the distribution probability of the monitoring results and the preset probability processing rules.

[0131] Optionally, the fault monitoring model includes convolutional layers, a normalized activation module, an attention mechanism module, a residual module, a fully connected layer, a Dropout layer, and a Softmax layer. The normalized activation module includes a first Batch Normalization and a first ReLU activation function.

[0132] Optionally, the convolutional layer includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. Each of the first, second, and third convolutional layers is connected to a normalized activation module. The input to the first convolutional layer is the heating current, the input to the second convolutional layer is the output of the normalized activation module connected to the first convolutional layer, the input to the third convolutional layer is the output of the normalized activation module connected to the second convolutional layer, and the output of the normalized activation module connected to the third convolutional layer serves as the input to the attention mechanism module.

[0133] Optionally, the residual module includes two residual blocks. Each residual block includes a second BatchNormalization, a second ReLU activation function, a weight layer, and a third Batch Normalization, which are connected in sequence. The input of the residual module is the output of the attention mechanism module, and the output of the residual module is the input of the fully connected layer.

[0134] Optionally, the attention mechanism module includes a channel attention module and a spatial attention module. The output of the attention mechanism module is obtained in the following way:

[0135] The output features of the normalized activation module connected after the third convolutional layer are obtained and then input into the spatial attention module to obtain the first feature;

[0136] The output feature and the first feature are multiplied element by element to obtain the product feature;

[0137] The product features are normalized, and the normalized product features are input into the channel attention module to obtain the second feature;

[0138] The second feature and the normalized product feature are multiplied element-wise to obtain the output of the attention mechanism module.

[0139] Optionally, the fully connected layer includes a first fully connected layer, a second fully connected layer, and a third fully connected layer connected in sequence. The input of the first fully connected layer is the output of the residual module, the output of the third fully connected layer is the input of the Dropout layer, and the output of the Dropout layer is the input of the Softmax layer.

[0140] Optionally, the device also includes a model training module, which trains the fault monitoring model through the following methods:

[0141] Obtain the training dataset, which includes a training sample set and a test sample set. The training dataset includes each sample pair, and each sample pair includes sample data and corresponding sample labels. The sample data includes the heating current of the heating wire of each heating pipeline unit in the breathing heating pipeline, and the sample label represents the annotation result of the corresponding fault detection result of the sample pair.

[0142] The initial neural network model is trained based on the training sample set until the loss function corresponding to the initial neural network model converges, thus obtaining the initial fault detection model. The input of the initial neural network model is a sample pair, and the output is the fault detection result corresponding to the sample pair. The value of the loss function characterizes the difference between the fault detection result corresponding to each sample pair output by the model and the labeled result corresponding to each sample pair.

[0143] The initial fault monitoring model is tested and verified based on the test sample set. If the test and verification results of the initial fault monitoring model do not meet the set verification requirements, the step of training the initial neural network model based on the training sample set is repeated until the test and verification results of the initial fault monitoring model meet the set verification requirements.

[0144] Optionally, the test sample set includes a validation sample set and a test sample set, and the sample data includes positive sample data and negative sample data. When the model training module obtains the training dataset, it is specifically used for:

[0145] Acquire at least one heating current of the breathing heating tubing during normal operation as at least one negative sample data;

[0146] Acquire at least one heating current in the breathing heating tubing when a fault occurs as at least one positive sample data;

[0147] At least one negative sample data and at least one positive sample data are randomly combined and divided according to a set quantity requirement to obtain the training sample set, the validation sample set and the test sample set respectively.

[0148] Optionally, validation requirements can be set, including recall and precision requirements for each sample pair.

[0149] The fault monitoring device for a breathing heating pipeline in this embodiment can execute the fault monitoring method for a breathing heating pipeline shown in the embodiment of this application. The implementation principle is similar and will not be described again here.

[0150] This application provides an electronic device, which includes a processor and a memory configured to store machine-readable instructions that, when executed by the processor, cause the processor to perform a fault monitoring method for a breathing heating pipeline.

[0151] This application provides an electronic device, such as... Figure 7 As shown, Figure 7 The illustrated electronic device includes a processor 2001 and a memory 2003. The processor 2001 and the memory 2003 are connected, for example, via a bus 2002. Optionally, the electronic device 2000 may further include a transceiver 2004. It should be noted that in practical applications, the transceiver 2004 is not limited to one type, and the structure of this electronic device 2000 does not constitute a limitation on the embodiments of this application.

[0152] Processor 2001 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 2001 may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0153] Bus 2002 may include a pathway for transmitting information between the aforementioned components. Bus 2002 may be a PCI bus or an EISA bus, etc. Bus 2002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0154] The memory 2003 may be ROM or other type of static storage device capable of storing static information and instructions, RAM or other type of dynamic storage device capable of storing information and instructions, or EEPROM, CD-ROM or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0155] The memory 2003 stores the application code that executes the scheme of this application, and its execution is controlled by the processor 2001. The processor 2001 executes the application code stored in the memory 2003 to implement... Figure 6 The illustrated embodiment provides the operation of a fault monitoring device for a breathing heating pipeline.

[0156] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0157] The above description is only a partial embodiment of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of fault monitoring of a respiratory warming tube, the method comprising: The method is applied to a heated breathing pipeline matched with a breathing machine, and the method comprises the following steps: obtaining a current heating current corresponding to the breathing heating pipeline, wherein the heating current is a current heating current corresponding to a heating wire of each heating pipeline unit in the breathing heating pipeline; inputting the heating current into a trained fault monitoring model to obtain a distribution probability of a monitoring result corresponding to the breathing heating pipeline, wherein the distribution probability of the monitoring result comprises a probability of failure and a probability of normal operation; determining a fault detection result corresponding to the breathing heating pipeline according to the distribution probability of the monitoring result and a preset probability processing rule; the fault monitoring model comprises a convolution layer, a normalization activation module, an attention mechanism module, a residual module, a full connection layer, a Dropout layer and a Softmax layer, and the normalization activation module comprises a first Batch Normalization and a first ReLU activation function; the convolution layer comprises a first layer convolution layer, a second layer convolution layer and a third layer convolution layer, and the first layer convolution layer, the second layer convolution layer and the third layer convolution layer are all connected with a normalization activation module, the input of the first layer convolution layer is the heating current, the input of the second layer convolution layer is the output of the normalization activation module connected after the first layer convolution layer, the input of the third layer convolution layer is the output of the normalization activation module connected after the second layer convolution layer, and the output of the normalization activation module connected after the third layer convolution layer is taken as the input of the attention mechanism module; the attention mechanism module comprises a channel attention module and a spatial attention module, and the output of the attention mechanism module is obtained in the following manner: obtaining an output feature of the normalization activation module connected after the third layer convolution layer, and inputting the output feature into the spatial attention module to obtain a first feature; multiplying the output feature and the first feature element by element to obtain a product feature; normalizing the product feature, and inputting the normalized product feature into the channel attention module to obtain a second feature; multiplying the second feature and the normalized product feature element by element to obtain the output of the attention mechanism module.

2. The method of claim 1, wherein, the residual module comprises two residual blocks, each of which comprises a second Batch Normalization, a second ReLU activation function, a weight layer and a third Batch Normalization connected in sequence, the input of the residual module is the output of the attention mechanism module, and the output of the residual module is the input of the full connection layer.

3. The method of claim 1, wherein, the full connection layer comprises a first layer full connection layer, a second layer full connection layer and a third layer full connection layer connected in sequence, the input of the first layer full connection layer is the output of the residual module, the output of the third layer full connection layer is the input of the Dropout layer, and the output of the Dropout layer is the input of the Softmax layer.

4. The method of claim 1, wherein, the fault monitoring model is trained in the following manner: obtaining a training data set, the training data set comprising a training sample set and a test sample set, the training data set comprising sample pairs, each sample pair comprising sample data and a corresponding sample label, wherein the sample data comprises a heating current of a heating wire of each heating pipe unit in a breathing warming pipe, and the sample label represents a labeled result of a corresponding fault detection result of the sample pair; training an initial neural network model based on the training sample set until a loss function corresponding to the initial neural network model converges, to obtain an initial fault monitoring model, wherein the input of the initial neural network model is a sample pair, the output of the initial neural network model is a fault detection result corresponding to the sample pair, and the value of the loss function represents a difference between the fault detection result output by the model and the labeled result corresponding to the sample pair; testing and verifying the initial fault monitoring model based on the test sample set, and if the test and verification result corresponding to the initial fault monitoring model does not meet a set verification requirement, repeating the step of training the initial neural network model based on the training sample set until the test and verification result corresponding to the initial fault monitoring model meets the set verification requirement.

5. The method of claim 4, wherein, The test sample set comprises a verification sample set and a test sample set, the sample data comprises positive sample data and negative sample data, and the obtaining of the training data set comprises: obtaining at least one heating current of the breathing warming pipe when the breathing warming pipe is working normally as at least one negative sample data; obtaining at least one heating current of the breathing warming pipe when the breathing warming pipe is working normally as at least one positive sample data; dividing the at least one negative sample data and the at least one positive sample data according to a set quantity requirement to obtain the training sample set, the verification sample set and the test sample set respectively.

6. The method of claim 4, wherein, The set verification requirement comprises a recall rate requirement and an accuracy rate requirement corresponding to each sample pair.

7. A fault monitoring apparatus of a respiratory warming line for implementing the fault monitoring method of the respiratory warming line according to any one of claims 1 to 6, characterized by, The fault monitoring device is applied to a warming breathing pipe matched with a breathing machine, and the fault monitoring device comprises: a current obtaining module configured to obtain a current corresponding heating current of the breathing warming pipe, the current corresponding heating current being a current corresponding heating current of a heating wire of each heating pipe unit in the breathing warming pipe; a probability obtaining module configured to input the current corresponding heating current into a trained fault monitoring model to obtain a distribution probability of a monitoring result corresponding to the breathing warming pipe, the distribution probability of the monitoring result comprising a probability of a fault and a probability of normal working; a result determining module configured to determine a fault detection result corresponding to the breathing warming pipe according to the distribution probability of the monitoring result and a preset probability processing rule.

Images