Ship power equipment early warning method, device, equipment, medium and product
By combining deep convolution to generate adversarial networks and Gaussian hybrid models, the degradation sensitive characteristics of ship power equipment are extracted and early warning models are constructed, which solves the problem of poor early warning effect in the existing technology and achieves fast and accurate fault warning.
Patent Information
- Application Number
- CN202411864233.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-18
- Publication Date
- 2025-04-11
AI Technical Summary
In the early warning of ship power equipment failure, it is difficult for the prior art to quickly and accurately extract the degradation sensitive characteristics of each device and each component, resulting in poor early warning effect.
The method of combining deep convolution generation adversarial network (DCGAN) and Gaussian hybrid model (GMM) is adopted to extract the degradation sensitive characteristics of ship power equipment in historical healthy states by training deep convolution generation adversarial networks, and a warning model is used to build an early warning model to achieve accurate early warning of the current state.
This method can quickly and accurately extract degradation sensitive features, improve the accuracy and efficiency of ship power equipment fault warning, and solve the problem of large research workload.
Smart Images
Figure CN119807840A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of ship power equipment fault warning, and in particular to a ship power equipment fault warning method, device, equipment, medium and product. Background Art
[0002] Ship power equipment is the direct power source of the ship and the core part of maintaining the basic functions of the ship. The health status of the ship power equipment can directly affect the navigation status of the ship. Accurate early warning of the health status of the ship power equipment is an important basis for formulating ship routes, determining navigation cycles and other navigation plans, and is also a key analysis input for formulating inspection and maintenance plans.
[0003] With the continuous development of computer technology and machine learning methods, fault warning technology based on statistical models has been gradually applied to the field of ship power equipment fault warning, among which the accurate extraction of degradation sensitive features is the key to the construction of warning models. Due to the large number of potential failure modes of ship equipment and the complex and diverse structures, the workload of conducting research on degradation mechanisms and sensitive feature extraction for each device and component is large, and it is difficult to ensure the effectiveness of the selected sensitive features when only the health status data of the power equipment is available, which ultimately affects the warning effect. Summary of the invention
[0004] The purpose of this application is to provide a method, device, equipment, medium and product for early warning of ship power equipment, which can quickly and accurately warn of the fault status of ship power equipment.
[0005] To achieve the above objectives, this application provides the following solutions:
[0006] In the first aspect, the present application provides an early warning method for ship power equipment, comprising: collecting a vibration time domain signal sample set of the ship power equipment to be tested operating in a historical healthy state; using part of the samples in the vibration time domain signal sample set to train a deep convolutional generative adversarial network, and using the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested; inputting the remaining samples in the vibration time domain signal sample set into the feature extraction model to obtain a degradation sensitive feature set of the ship power equipment to be tested in a historical healthy state; constructing an early warning model for the ship power equipment to be tested based on the degradation sensitive feature set and a Gaussian mixture model under the historical healthy state; inputting the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive features of the ship power equipment to be tested in the current operating state; inputting the degradation sensitive features in the current operating state into the early warning model to obtain the current state of the ship power equipment to be tested; the current state is a healthy state or a fault state; if the current state of the ship power equipment to be tested is a fault state, a warning is issued.
[0007] In the second aspect, the present application provides a ship power equipment early warning device, including: an acquisition module, used to collect a vibration time domain signal sample set of the ship power equipment to be tested operating in a historical healthy state; a training module, used to train a deep convolutional generative adversarial network using part of the samples in the vibration time domain signal sample set, and use the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested; a feature extraction module, used to input the remaining samples in the vibration time domain signal sample set into the feature extraction model to obtain a degradation sensitive feature set of the ship power equipment to be tested in a historical healthy state; an early warning model A construction module is used to construct an early warning model for the ship power equipment to be tested based on the degradation sensitive feature set and Gaussian mixture model under the historical healthy state; an application module is used to input the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive features of the ship power equipment to be tested in the current operating state; a state determination module is used to input the degradation sensitive features in the current operating state into the early warning model to obtain the current state of the ship power equipment to be tested; the current state is a healthy state or a fault state; an early warning module is used to issue an early warning if the current state of the ship power equipment to be tested is a fault state.
[0008] In a third aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described ship power equipment early warning methods.
[0009] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned ship power equipment early warning methods.
[0010] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned ship power equipment early warning methods.
[0011] According to the specific embodiments provided in this application, this application has the following technical effects:
[0012] The present application provides a method, device, equipment, medium and product for early warning of ship power equipment. By combining the data of the ship power equipment to be tested in a healthy state with a deep convolutional generative adversarial network, the distribution law of the data in a healthy state can be learned, and the automatic extraction of degradation sensitive features can be guided, which solves the problem of the large workload of researching the degradation mechanism and sensitive feature extraction of each device and each component one by one. The discriminator in the trained deep convolutional generative adversarial network can accurately extract the degradation sensitive features. On the basis of the extracted degradation sensitive features, the early warning model constructed by using the Gaussian mixture model can further accurately warn the current state of the ship power equipment to be tested. Therefore, the combination of the deep convolutional generative adversarial network and the Gaussian mixture model realizes the rapid and accurate early warning of ship power equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0014] Figure 1 A schematic diagram of a process flow of a ship power equipment early warning method provided in an embodiment of the present application;
[0015] Figure 2 A schematic diagram of the structure of a deep convolutional generative adversarial network provided in another embodiment of the present application;
[0016] Figure 3 A more specific flowchart of a ship power equipment early warning method provided by another embodiment of the present application;
[0017] Figure 4 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0018] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0019] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.
[0020] In an exemplary embodiment, Figure 1 As shown, a method for early warning of ship power equipment is provided, comprising the following steps 101 to 107. Among them:
[0021] Step 101: Collect a vibration time domain signal sample set of the power equipment of the ship to be tested running in a historical healthy state.
[0022] Step 102: Use some samples in the vibration time domain signal sample set to train a deep convolutional generative adversarial network, and use the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested.
[0023] Step 103: Input the remaining samples in the vibration time domain signal sample set into a feature extraction model to obtain a degradation sensitive feature set of the ship power equipment to be tested under a historical healthy state.
[0024] Step 104: Construct an early warning model for the power equipment of the ship to be tested based on the degradation sensitive feature set and the Gaussian mixture model under the historical health state.
[0025] Step 105: Input the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive characteristics of the ship power equipment to be tested in the current operating state.
[0026] Step 106: Input the degradation sensitive characteristics under the current operating state into the early warning model to obtain the current state of the ship power equipment to be tested; the current state is a healthy state or a fault state.
[0027] Step 107: If the current state of the power equipment of the ship to be tested is a fault state, an early warning is issued.
[0028] Implementing the above steps 101 to 107, the Deep Convolution GAN (DCGAN) is a network structure built based on the game theory confrontation idea. It generates sample data similar to the input sample distribution law by the confrontation between the generator G and the discriminator D, and has achieved good application results in the field of data expansion. By combining the health data of the ship power equipment to be analyzed with DCGAN, the distribution law of the health data can be learned and the automatic extraction of degradation sensitive features can be guided. The Gaussian Mixture Model (GMM) can realize the distribution fitting of multi-dimensional features and is often used in the construction of early warning models based on multi-dimensional features. Therefore, the combination of DCGAN and GMM is used to realize the intelligent early warning of ship power equipment.
[0029] In another exemplary embodiment of the present application, the training process of the deep convolutional generative adversarial network in the above step 102 can be replaced by the following steps 201 to 202:
[0030] Step 201: extract one third of the vibration time domain signal samples in the vibration time domain signal sample set, and train a deep convolutional generative adversarial network.
[0031] Step 202: When the discriminator output probability in the deep convolutional generative adversarial network stabilizes to 0.5, stop training to obtain a trained deep convolutional generative adversarial network.
[0032] Or stop training when the discriminator output probability in the deep convolutional generative adversarial network stabilizes to around 0.5.
[0033] In another exemplary embodiment of the present application, the structure of the deep convolutional generative adversarial network is as follows: Figure 2 As shown in Table 1, the structural parameters of the deep convolutional generative adversarial network are shown.
[0034] Table 1 DCGAN network structure parameters
[0035]
[0036]
[0037] The deep convolutional generative adversarial network includes: a generator and a discriminator. The generator includes the first convolution layer (convolution layer #1), the first upsampling layer (upsampling layer #1), the second convolution layer (convolution layer #2), the second upsampling layer (upsampling layer #2), the third convolution layer (convolution layer #3), the third upsampling layer (upsampling layer #3), the fourth convolution layer (convolution layer #4), the fourth upsampling layer (upsampling layer #4) and the fifth convolution layer (convolution layer #5) connected in sequence. The discriminator includes the sixth convolution layer (convolution layer #6), the first pooling layer (pooling layer #1), the seventh convolution layer (convolution layer #7), the second pooling layer (pooling layer #2), the eighth convolution layer (convolution layer #8), the third pooling layer (pooling layer #3), the ninth convolution layer (convolution layer #9), the fourth pooling layer (pooling layer #4) and the fully connected layer (fully connected layer #1) connected in sequence.
[0038] Among them, the number of convolution kernels of the sixth convolution layer, the seventh convolution layer, the eighth convolution layer, the ninth convolution layer and the fully connected layer are 32, 64, 128, 256 and 1024 respectively. The convolution kernel sizes of the sixth convolution layer, the first pooling layer, the seventh convolution layer, the second pooling layer, the eighth convolution layer, the third pooling layer, the ninth convolution layer and the fourth pooling layer are 32×32, 2×1, 5×5, 2×1, 5×5, 2×1, 5×5 and 2×1 respectively. The step sizes of the sixth convolution layer, the first pooling layer, the seventh convolution layer, the second pooling layer, the eighth convolution layer, the third pooling layer, the ninth convolution layer and the fourth pooling layer are 4, 2, 4, 2, 4, 2, 4 and 2 respectively. The activation functions of the sixth convolutional layer, the seventh convolutional layer, the eighth convolutional layer, the ninth convolutional layer and the fully connected layer are LeakyReLU, LeakyReLU, LeakyReLU, LeakyReLU and Sigmoid, respectively.
[0039] In another exemplary embodiment of the present application, if one-third of the vibration time domain signal samples in the vibration time domain signal sample set are extracted to train the deep convolutional generative adversarial network, then in step 103, two-thirds of the samples in the vibration time domain signal sample set are input into the feature extraction model to automatically extract sensitive features reflecting the health status of the ship equipment to be tested, thereby completing the construction of a multi-dimensional feature sample of the health status.
[0040] In another exemplary embodiment of the present application, the above step 104 may be replaced by the following steps 301 to 304:
[0041] Step 301: Divide the degradation sensitive feature set under the historical health state into training samples and threshold learning samples.
[0042] Step 302: Use the training samples to train a Gaussian mixture model to obtain a trained Gaussian mixture model.
[0043] Step 303: Input the threshold learning sample into the trained Gaussian mixture model, and obtain the self-learning alarm threshold according to the 3σ criterion.
[0044] Step 304: The trained Gaussian mixture model and the self-learning alarm threshold constitute an early warning model for the power equipment of the ship to be tested.
[0045] GMM is a mixed density function distribution model based on Bayesian probability statistics. The distribution of vector features in probability space is represented by the weighted sum of multiple Gaussian probability density functions. M mixed numbers are linear combinations of M single Gaussian distributions, namely:
[0046]
[0047] Where: x is the signal characteristic sequence of D-dimensional vector; a i(i=1,2,…,M)—the weight of the i-th single Gaussian distribution, and p i —Single Gaussian probability distribution function; λ = {a i , μ i ,Σ i};μ i —mean of a single Gaussian distribution function; Σ i —The covariance matrix of the single Gaussian distribution function.
[0048] The Gaussian probability density distribution function is shown below:
[0049]
[0050] In the formula, D is the dimension describing the system state characteristics, which is the same as the characteristic dimension proposed in S3.
[0051] The KL distance is used to measure the difference between the current state and the normal state. The smaller the KL distance, the smaller the difference between the two distribution functions. It is defined as follows:
[0052]
[0053] Where: p(x i )—target distribution; q(x i )—matched distribution; x i —Discrete random variable; N—length of distribution, the GMM alarm threshold is set according to the 3σ criterion.
[0054] In another exemplary embodiment of the present application, referring to Figure 3 , the above step 106 can be replaced by the following steps 401 to 404:
[0055] Step 401: Input the degradation sensitive features in the current operating state into the trained Gaussian mixture model, and output the KL distance between the current operating state and the healthy state.
[0056] Step 402: Compare the KL distance with the self-learning alarm threshold.
[0057] Step 403: If the KL distance is greater than the self-learning alarm threshold, it is determined that the current state of the ship power equipment to be tested is a fault state.
[0058] Step 404: If the KL distance is less than or equal to the self-learning alarm threshold, it is determined that the current state of the ship power equipment to be tested is a healthy state.
[0059] Based on the same inventive concept, the embodiment of the present application also provides a ship power equipment early warning device for implementing the ship power equipment early warning method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more ship power equipment early warning device embodiments provided below can refer to the limitations of the ship power equipment early warning method above, and will not be repeated here.
[0060] In an exemplary embodiment, a ship power equipment early warning device is provided, including: an acquisition module, a training module, a feature extraction module, an early warning model construction module, an application module, a state determination module and an early warning module.
[0061] The acquisition module is used to acquire a vibration time domain signal sample set of the ship power equipment to be tested running in a historical healthy state. The training module is used to train a deep convolutional generative adversarial network using some samples in the vibration time domain signal sample set, and use the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested. The feature extraction module is used to input the remaining samples in the vibration time domain signal sample set into the feature extraction model to obtain the degradation sensitive feature set of the ship power equipment to be tested in a historical healthy state. The early warning model construction module is used to construct an early warning model for the ship power equipment to be tested based on the degradation sensitive feature set in the historical healthy state and the Gaussian mixture model. The application module is used to input the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive features of the ship power equipment to be tested in the current operating state. The state determination module is used to input the degradation sensitive features in the current operating state into the early warning model to obtain the current state of the ship power equipment to be tested; the current state is a healthy state or a fault state. The early warning module is used to issue an early warning if the current state of the power equipment of the ship to be tested is a fault state.
[0062] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 4As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store the current state of the ship power equipment to be tested. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a ship power equipment early warning method is implemented.
[0063] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components. In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.
[0064] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0065] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.
[0066] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0067] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0068] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.
[0069] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0070] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.
Claims
1. A method for early warning of ship power equipment, characterized in that: include: Collect vibration time domain signal sample sets of the power equipment of the ship to be tested running in the historical healthy state; Using some samples in the vibration time domain signal sample set to train a deep convolutional generative adversarial network, and using the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested; Inputting the remaining samples in the vibration time domain signal sample set into a feature extraction model to obtain a degradation sensitive feature set of the power equipment of the ship to be tested under a historical healthy state; According to the degradation sensitive feature set and Gaussian mixture model under the historical health status, the early warning model of the ship power equipment to be tested is constructed; Inputting the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive characteristics of the ship power equipment to be tested in the current operating state; Inputting the degradation sensitive characteristics under the current operating state into the early warning model to obtain the current state of the power equipment of the ship to be tested; the current state is a healthy state or a fault state; If the current state of the power equipment of the ship to be tested is a fault state, an early warning is issued.
2. The ship power equipment early warning method according to claim 1 is characterized in that: Using some samples in the vibration time domain signal sample set to train a deep convolutional generative adversarial network specifically includes: Extracting one third of the vibration time domain signal samples from the vibration time domain signal sample set to train a deep convolutional generative adversarial network; When the discriminator output probability in the deep convolutional generative adversarial network stabilizes to 0.5, the training is stopped to obtain the trained deep convolutional generative adversarial network.
3. The ship power equipment early warning method according to claim 1 is characterized in that: The deep convolutional generative adversarial network includes: a generator and a discriminator; The generator includes a first convolution layer, a first upsampling layer, a second convolution layer, a second upsampling layer, a third convolution layer, a third upsampling layer, a fourth convolution layer, a fourth upsampling layer and a fifth convolution layer connected in sequence; The discriminator includes a sixth convolutional layer, a first pooling layer, a seventh convolutional layer, a second pooling layer, an eighth convolutional layer, a third pooling layer, a ninth convolutional layer, a fourth pooling layer and a fully connected layer, which are connected in sequence.
4. The ship power equipment early warning method according to claim 3 is characterized in that: The number of convolution kernels in the sixth convolutional layer, the seventh convolutional layer, the eighth convolutional layer, the ninth convolutional layer and the fully connected layer are 32, 64, 128, 256 and 1024 respectively; The convolution kernel sizes of the sixth convolution layer, the first pooling layer, the seventh convolution layer, the second pooling layer, the eighth convolution layer, the third pooling layer, the ninth convolution layer, and the fourth pooling layer are 32×32, 2×1, 5×5, 2×1, 5×5, 2×1, 5×5, and 2×1 respectively; The step sizes of the sixth convolutional layer, the first pooling layer, the seventh convolutional layer, the second pooling layer, the eighth convolutional layer, the third pooling layer, the ninth convolutional layer, and the fourth pooling layer are 4, 2, 4, 2, 4, 2, 4, and 2, respectively; The activation functions of the sixth convolutional layer, the seventh convolutional layer, the eighth convolutional layer, the ninth convolutional layer and the fully connected layer are LeakyReLU, LeakyReLU, LeakyReLU, LeakyReLU and Sigmoid, respectively.
5. The ship power equipment early warning method according to claim 1, characterized in that: According to the degradation sensitive feature set and Gaussian mixture model under the historical health state, an early warning model of the ship power equipment to be tested is constructed, which includes: Divide the degradation-sensitive feature set under the historical health state into training samples and threshold learning samples; Using the training samples to train a Gaussian mixture model to obtain a trained Gaussian mixture model; Input the threshold learning samples into the trained Gaussian mixture model, and obtain the self-learning alarm threshold according to the 3σ criterion; The trained Gaussian mixture model and self-learning alarm threshold constitute the early warning model of the ship power equipment to be tested.
6. The ship power equipment early warning method according to claim 5 is characterized in that: Inputting the degradation sensitive features under the current operating state into the early warning model to obtain the current state of the power equipment of the ship to be tested specifically includes: Input the degradation-sensitive features of the current operating state into the trained Gaussian mixture model and output the KL distance between the current operating state and the healthy state; Comparing the KL distance with a self-learning alarm threshold; If the KL distance is greater than the self-learning alarm threshold, it is determined that the current state of the power equipment of the ship to be tested is a fault state; If the KL distance is less than or equal to the self-learning alarm threshold, it is determined that the current state of the power equipment of the ship to be tested is a healthy state.
7. A ship power equipment early warning device, characterized in that: The ship power equipment early warning device comprises: The acquisition module is used to collect a set of vibration time domain signal samples of the power equipment of the ship to be tested running in a historical healthy state; A training module, used to train a deep convolutional generative adversarial network using some samples in the vibration time domain signal sample set, and use the discriminator in the trained deep convolutional generative adversarial network as a feature extraction model for the ship power equipment to be tested; A feature extraction module, used for inputting the remaining samples in the vibration time domain signal sample set into a feature extraction model to obtain a degradation sensitive feature set of the power equipment of the ship to be tested under a historical health state; The early warning model building module is used to build an early warning model of the ship power equipment to be tested based on the degradation sensitive feature set and Gaussian mixture model under the historical health state; An application module is used to input the vibration time domain signal of the current operating state of the ship power equipment to be tested into the feature extraction model to obtain the degradation sensitive characteristics of the ship power equipment to be tested in the current operating state; A state determination module is used to input the degradation sensitive characteristics under the current operating state into the early warning model to obtain the current state of the ship power equipment to be tested; the current state is a healthy state or a fault state; The early warning module is used to issue an early warning if the current state of the power equipment of the ship to be tested is a fault state.
8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the ship power equipment early warning method according to any one of claims 1 to 6.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the ship power equipment early warning method 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 ship power equipment early warning method described in any one of claims 1 to 6 is implemented.