Ship power device fault type identification method and device

Abstract

The invention provides a ship power device fault type identification method and device, and relates to the technical field of deep learning, and the method comprises the steps: obtaining the current fault data of a ship power device, extracting the deep features of the current fault data through employing a convolution layer in a convolution neural network fault diagnosis model combined with an attention mechanism, and obtaining the deep features of the current fault data; and key features in the deep features are extracted through an attention layer, and the fault type of the ship power device is identified based on the extracted key features. According to the method, the convolutional neural network fault diagnosis model combined with the attention mechanism is established, and based on the current fault data of the ship power device, the fault type of the ship power device is accurately and quickly identified in the early stage of an accident, so that the occurrence probability of major accidents is reduced, and the fault diagnosis efficiency is improved. And the operation safety of the ship power device is improved.

Description

Technical Field

[0001] The present invention relates to the field of deep learning technology, and in particular to a method and device for identifying the fault type of a ship power plant. Background Art

[0002] The ship's power plant is a very complex system involving hundreds of types of equipment. Therefore, the probability of failure of the ship's power plant during operation is relatively high. Once an accident occurs, it will have catastrophic consequences.

[0003] Traditional methods for identifying accident types of ship power plants are based on design rules, that is, based on expert system knowledge, fault diagnosis is performed by establishing complex and accurate mathematical and physical models. The advantage of this type of method is that the accident characteristics are relatively clear, and the accident can be clearly judged within the scope of the rules. However, in the actual operation of power plants, many accidents occur when they have never occurred before and the parameters do not change much, which brings great difficulties to the identification of fault types. It is often difficult to accurately and quickly identify the fault type, which leads to the occurrence of major accidents. Summary of the invention

[0004] In view of this, the present invention provides a method and device for identifying the fault type of a ship power plant, so as to solve the above-mentioned problems existing in the prior art.

[0005] The present invention provides a method for identifying the fault type of a ship power plant, comprising the following steps.

[0006] Obtain current fault data of ship power plant; Inputting current fault data of the ship power plant into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtaining a fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0007] According to a method for identifying the fault type of a ship power plant provided by the present invention, the convolutional neural network fault diagnosis model combined with the attention mechanism is obtained based on the following steps: Obtain historical fault data of ship power plants; Based on the attention mechanism, an attention layer is added to the CNN model to build a neural network model; The constructed neural network model is trained based on the historical fault data of the ship power plant to obtain a convolutional neural network fault diagnosis model combined with an attention mechanism.

[0008] According to a method for identifying a fault type of a ship power plant provided by the present invention, the deep features of the current fault data include at least one of time information and spatial information of the fault data.

[0009] According to a method for identifying the fault type of a ship power plant provided by the present invention, the key feature of the current fault data refers to the deep feature of the current fault data that has been subjected to weighted processing.

[0010] According to a method for identifying the fault type of a ship power plant provided by the present invention, the loss function of the convolutional neural network fault diagnosis model combined with the attention mechanism includes any one of a binary cross entropy loss function and a multi-classification cross entropy loss function.

[0011] According to a method for identifying the fault type of a ship power plant provided by the present invention, the method also includes: converting the fault type of the ship power plant output by the fully connected layer into a probability distribution through an activation function; the activation function includes any one of a Sigmoid function and a Softmax function.

[0012] The present invention also provides a device for identifying the fault type of a ship power plant, comprising the following modules: An acquisition module, used to acquire current fault data of a ship power plant; An identification module, used to input current fault data of the ship power unit into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtain a fault type of the ship power unit output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for identifying the fault type of a ship power plant as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for identifying the fault type of a ship power plant as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for identifying the fault type of a ship power plant as described above is implemented.

[0016] The method and device for identifying the fault type of a ship power plant provided by the present invention obtain the current fault data of the ship power plant, and use a convolutional neural network fault diagnosis model combined with an attention mechanism to extract features of the current fault data, thereby achieving accurate and rapid identification of the fault type of the ship power plant, thereby reducing the probability of major accidents and improving the safety of the operation of the ship power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a flow chart of the method for identifying the fault type of a ship power plant provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the convolutional neural network process provided by the present invention.

[0020] Figure 3 It is a convolution schematic diagram provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the convolutional neural network fault diagnosis model combined with the attention mechanism provided by the present invention.

[0022] Figure 5 It is a structural schematic diagram of a device for identifying a fault type of a ship power plant provided by the present invention.

[0023] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0024] Since the ship power plant is a very complex system, involving hundreds of types and quantities of equipment, the probability of failure of the power plant during operation is relatively high. Once an accident occurs, it will have catastrophic consequences. Therefore, it is necessary to propose a more reliable fault type identification method.

[0025] The present invention proposes a method for identifying the fault type of a ship power plant based on a convolutional neural network. The method takes the ship power plant as the research object and identifies the fault type of the power plant based on a convolutional neural network. Based on data-driven technology (data-driven means using historical data to evaluate the data of online detection to obtain the health status of the power plant system), a large amount of data input by the sensor is learned through a convolutional neural network, so that the type and severity of the accident can be judged in the early stage of the accident.

[0026] Compared with traditional diagnostic methods based on expert systems and mathematical and physical models, data-driven diagnosis can avoid the use of overly complex and limited expert knowledge, and avoid the establishment of complex and accurate mathematical and physical models. In addition, through the learning and training of all fault levels by the convolutional neural network model, the fault level of the actual power unit can be identified relatively easily. For the large amount of sensor sequence data built into the power unit, it is combined into a matrix and input into the convolutional neural network model to output the fault type of the power unit.

[0027] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] Combine the following Figures 1 to 6 The present invention describes a method and device for identifying the fault type of a ship power plant.

[0029] Figure 1 FIG. 1 is a flow chart of a method for identifying a fault type of a ship power plant provided by the present invention. Figure 1 As shown, the method includes step 101 and step 102.

[0030] Step 101: Acquire current fault data of a ship power plant.

[0031] According to the embodiment provided by the present invention, a large number of sensors built into the ship power unit are used to collect sequence data generated by the operation of the ship power unit when a fault occurs to identify the fault type of the ship power unit. The sequence data includes operating speed, wind speed, water speed, and water level depth of the ship. When one or more of the sequence data are abnormal, it means that one or more faults have occurred in the ship power unit.

[0032] Step 102: input the current fault data of the ship power unit into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtain the fault type of the ship power unit output by the convolutional neural network fault diagnosis model combined with the attention mechanism.

[0033] Furthermore, the convolutional neural network fault diagnosis model combined with the attention mechanism is obtained based on the following steps: Obtain historical fault data of ship power plants; Based on the attention mechanism, an attention layer is added to the CNN model to build a neural network model; The constructed neural network model is trained based on the historical fault data of the ship power plant to obtain a convolutional neural network fault diagnosis model combined with an attention mechanism.

[0034] Optionally, the corresponding fault data of each failure of the ship power plant in a certain historical period (such as the previous three years, the previous five years or the previous ten years, etc.) are collected and combined into a matrix for training the neural network model.

[0035] According to an embodiment provided by the present invention, a convolutional neural network fault diagnosis model combined with an attention mechanism is constructed by adding an attention layer after the convolution layer in the convolutional neural network model.

[0036] Specifically, Figure 2 It is a schematic diagram of the convolutional neural network process provided by the present invention, such as Figure 2 As shown in the figure. Convolutional Neural Network (CNN) is a deep learning model that is mainly used to process and analyze tasks with grid-structured data. The components of a convolutional neural network mainly include the basic input layer and output layer, as well as the core operation part of the convolution layer, pooling layer (also known as sampling layer) and fully connected layer.

[0037] In two-dimensional convolution, the main function of convolution is to extract features by sliding the convolution kernel (also called filter) on the input data. The essence of the convolution operation here is circular product and cumulative summation. The mathematical formula of two-dimensional convolution is as follows:

[0038] in, express OK The output matrix of columns, Indicates the height , the width is The two-dimensional convolution kernel, express OK, The input matrix of columns, and They represent the height and width of the two-dimensional convolution kernel respectively.

[0039] Figure 3 It is a convolution schematic diagram provided by the present invention, such as Figure 3 In the convolution operation, the convolution kernel starts to slide from the upper left corner of the input matrix, and each time it slides one pixel, the pixel value at the corresponding position is the sum of the products of the convolution kernel and the pixel value at the corresponding position of the input matrix.

[0040] After the convolution layer, adding an attention layer based on the attention mechanism allows the model to further process the features extracted by the convolution layer and distinguish the important and secondary information, thereby improving the accuracy of subsequent classification and recognition.

[0041] Next, the attention mechanism used in the present invention is described. When observing an object, the human brain tends to focus on certain parts of the object. These focused parts are often the key to obtaining information from the object. This information has a strong guiding role in the cognition of similar objects, and the attention mechanism is a special mechanism that imitates this cognitive process.

[0042] The attention mechanism can be simply understood as a weighted summator or a key feature extractor. It mainly performs weighted summation operations. Its main calculation formula is as follows:

[0043] in, Represents input features The probability distribution of Represents input features With the preset query vector The probability ratio function of the similarity between them; Represents input features With the preset query vector The probability ratio function of the similarity between them; According to the probability distribution From arrive All input features of The value after weighted summation, where , and it is a positive integer.

[0044] In the attention mechanism, a query vector is introduced, which represents the current content of attention or the information of interest. The query vector can be a fixed parameter or dynamically generated according to the input. The similarity between the query vector and the input feature is used to measure the degree of association between each element in the input and the query. Common similarity calculation methods include dot product, cosine similarity, and additive similarity. By normalizing the similarity, the attention weight of each input element is obtained. A higher weight indicates that the element is more important in the current task and requires more attention. The attention weight is multiplied by the input feature to obtain the weighted feature, which allows the model to pay more attention to elements with higher weights in subsequent calculations, thereby achieving centralized processing of specific information.

[0045] Aggregating the weighted representations, such as summing or taking a weighted average of all weighted representations, can obtain a comprehensive representation in which important information will receive more attention and less important information will receive less attention.

[0046] At this time, the fault data of the ship power plant in a certain historical period is divided into training set and test set in proportion, and input into the convolutional neural network model with the attention layer for supervised training, so that the model can learn and train all fault types, thereby obtaining a convolutional neural network fault diagnosis model combined with the attention mechanism that can identify the fault types of the ship power plant.

[0047] According to an embodiment provided by the present invention, the inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further comprises: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0048] Furthermore, the deep features of the current fault data include at least one of time information and space information of the fault data.

[0049] Furthermore, the key features of the current fault data refer to the deep features of the current fault data that have been weighted.

[0050] Figure 4 is a schematic diagram of the structure of the convolutional neural network fault diagnosis model combined with the attention mechanism provided by the present invention, such as Figure 4 shown.

[0051] When the convolutional neural network fault diagnosis model combined with the attention mechanism is applied to fault type identification, the current fault data of the ship power unit is first input into the convolutional neural network fault diagnosis model combined with the attention mechanism, and the convolution layer in the convolutional neural network module simultaneously extracts the deep features in the current fault data of the ship power unit, namely, time information and spatial information, and uses them as the input of the attention layer; the attention mechanism simulates the ability of the human brain to process a large amount of information, and automatically learns the input deep feature information to improve the weight value of important feature information in the deep feature, and then the attention layer maps the combined data to the attention module for automatic weight allocation, increases the attention weight value of key information in the fault data, and obtains key features; finally, the key features are mapped to the fully connected layer, and fault classification is performed in the fully connected layer.

[0052] Furthermore, the fault type of the ship power unit output by the fully connected layer is converted into a probability distribution through an activation function; the activation function includes any one of a Sigmoid function and a Softmax function.

[0053] Preferably, after the fully connected layer performs fault classification, the Softmax function is applied to convert the fault type of the ship power unit output by the fully connected layer into a probability distribution, and the predicted probability of each fault category output by the convolutional neural network fault diagnosis model combined with the attention mechanism is obtained, and the model performance can also be evaluated.

[0054] The formal description of the Softmax function is as follows:

[0055] Represents the input vector Middle The probability value of each element is used to convert the input vector into a probability distribution. Represents the input vector The The index of the elements, represents the length of the input vector, Represents the input vector The jThe Softmax function converts each element of the input vector into a probability value so that the sum of the probabilities of all elements is 1.

[0056] Furthermore, the loss function of the convolutional neural network fault diagnosis model combined with the attention mechanism includes any one of a binary cross entropy loss function and a multi-classification cross entropy loss function.

[0057] Preferably, according to the embodiment provided by the present invention, the loss function of the convolutional neural network fault diagnosis model combined with the attention mechanism can select a multi-classification cross entropy loss function.

[0058] The method for identifying the fault type of a ship power plant provided by the present invention obtains the current fault data of the ship power plant, uses a convolutional neural network fault diagnosis model combined with an attention mechanism to extract features of the current fault data, and based on the extracted key features, accurately and quickly identifies the type and degree of the ship power plant fault in the early stage of an accident, thereby reducing the probability of major accidents and improving the safety of the operation of the ship power plant.

[0059] The following is a description of the ship power plant fault type identification device provided by the present invention. The ship power plant fault type identification device described below and the ship power plant fault type identification method described above can be referenced to each other.

[0060] Based on any of the above embodiments, Figure 5 is a schematic diagram of the structure of a device for identifying a fault type of a ship power plant provided by an embodiment of the present invention. Figure 5 As shown, an embodiment of the present invention provides a device for identifying the fault type of a ship power plant, including an acquisition module 501 and an identification module 502, wherein: The acquisition module 501 is used to acquire the current fault data of the ship power plant; the identification module 502 is used to input the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtain the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism; wherein, the inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0061] The ship power plant fault type identification device provided by the present invention obtains the current fault data of the ship power plant, uses the convolution layer in the convolutional neural network fault diagnosis model combined with the attention mechanism to extract the deep features of the current fault data, and extracts the key features of the current fault data through the attention layer, and based on the extracted key features, accurately and quickly identifies the type and degree of the ship power plant fault in the early stage of an accident, thereby reducing the probability of major accidents and improving the safety of the operation of the ship power plant.

[0062] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6 As shown, the electronic device may include: a processor 610, a communications interface 620, a memory 630 and a communication bus 640, wherein the processor 610, the communications interface 620 and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute the method for identifying the fault type of a ship power plant, and the method includes: Obtain current fault data of ship power plant; Inputting current fault data of the ship power plant into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtaining a fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0063] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0064] On the other hand, the present invention further provides a computer program product, the computer program product comprising a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the ship power plant fault type identification method provided by the above methods, the method comprising: Obtain current fault data of ship power plant; Inputting current fault data of the ship power plant into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtaining a fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0065] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for identifying a fault type of a ship power plant provided by the above methods is implemented, and the method comprises: Obtain current fault data of ship power plant; Inputting current fault data of the ship power plant into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtaining a fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

[0066] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0067] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0068] 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 make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for identifying the fault type of a ship power plant, characterized in that: include: Obtain current fault data of ship power plant; Inputting current fault data of the ship power plant into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtaining a fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

2. The method for identifying the fault type of a ship power plant according to claim 1, characterized in that: The convolutional neural network fault diagnosis model combined with the attention mechanism is obtained based on the following steps: Obtain historical fault data of ship power plants; Based on the attention mechanism, an attention layer is added to the CNN model to build a neural network model; The constructed neural network model is trained based on the historical fault data of the ship power plant to obtain a convolutional neural network fault diagnosis model combined with an attention mechanism.

3. The method for identifying the fault type of a ship power plant according to claim 1, characterized in that: The deep features of the current fault data include at least one of time information and space information of the fault data.

4. The method for identifying the fault type of a ship power plant according to claim 1, characterized in that: The key features of the current fault data refer to the deep features of the current fault data that have been weighted.

5. The method for identifying the fault type of a ship power plant according to claim 1, characterized in that: The loss function of the convolutional neural network fault diagnosis model combined with the attention mechanism includes any one of a binary cross entropy loss function and a multi-classification cross entropy loss function.

6. The method for identifying the fault type of a ship power plant according to claim 1, characterized in that: Also includes: The fault type of the ship power unit output by the fully connected layer is converted into a probability distribution through an activation function; the activation function includes any one of a Sigmoid function and a Softmax function.

7. A device for identifying the fault type of a ship power plant, characterized in that: include: An acquisition module, used to acquire current fault data of a ship power plant; An identification module, used to input current fault data of the ship power unit into a convolutional neural network fault diagnosis model combined with an attention mechanism, and obtain a fault type of the ship power unit output by the convolutional neural network fault diagnosis model combined with an attention mechanism; Wherein, inputting the current fault data of the ship power plant into the convolutional neural network fault diagnosis model combined with the attention mechanism, and obtaining the fault type of the ship power plant output by the convolutional neural network fault diagnosis model combined with the attention mechanism, further includes: Based on the current fault data of the ship power plant, extracting deep features of the current fault data of the ship power plant through a convolutional layer; Based on the deep features of the current fault data of the ship power plant, extracting key features of the current fault data of the ship power plant through an attention layer; Based on the key features of the current fault data of the ship power plant, the fault type of the ship power plant output by the fully connected layer is obtained.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for identifying the fault type of a ship power plant as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for identifying the fault type of a ship power plant as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for identifying the fault type of a ship power plant as described in any one of claims 1 to 6 is implemented.

Images