Centrifugal pump fault diagnosis method, system and equipment based on digital twinning

Abstract

The invention discloses a centrifugal pump fault diagnosis method, system and equipment based on digital twinning, and the method comprises the steps: obtaining historical operation data of a centrifugal pump, and obtaining original sample data; determining feature parameters of the original sample data by adopting a feature selection method based on multi-objective optimization; establishing a digital twin model of the centrifugal pump according to the characteristic parameters, and generating first new sample data similar to the original sample data; using a WGAN model with gradient penalty to generate second new sample data approaching Nash equilibrium; and feature extraction is carried out on the mixed sample data based on the convolutional neural network, an intelligent fault diagnosis model is obtained, and intelligent fault diagnosis of the centrifugal pump is achieved. According to the method, the fault sample sets which are sufficient in number and approximate to Nash equilibrium in quality are generated based on the digital twinning technology, the problem that the model generalization ability is limited due to sample scarcity in the prior art is solved, and a new solution is provided for centrifugal pump fault diagnosis under the condition of few fault samples.

Description

Technical Field

[0001] The present invention relates to the field of computer integrated manufacturing technology, and in particular to a centrifugal pump fault diagnosis method, system, electronic device and storage device based on digital twin. Background Art

[0002] Centrifugal pumps are common key equipment in the machinery industry and are widely used in complex product systems such as large nuclear power energy, water supply and drainage, and ship power. The efficient and reliable operation of centrifugal pumps is an important guarantee for the safety and stability of complex product systems. Therefore, research on the fault diagnosis method of centrifugal pumps is of great significance to ensure the safe and efficient operation of complex product systems, and has broad engineering application prospects.

[0003] With the popularization of deep learning technology, deep learning-based diagnostic models have been applied in the fault diagnosis of centrifugal pumps. However, the existing centrifugal pump fault diagnosis methods have certain limitations. This is because most of the existing diagnostic models are generally supervised learning, which requires sufficient labeled data for iterative training. If the distribution of fault samples is biased or the number is insufficient, supervised learning will not have sufficient generalization and diagnostic performance, and it will be difficult to meet the abnormal type diagnosis needs of centrifugal pumps in the actual production environment.

[0004] Therefore, it is necessary to propose a centrifugal pump fault diagnosis method, system and equipment based on digital twins, which can solve the problems faced by centrifugal pumps in the fault diagnosis process, such as lack of fault data, heavy reliance on expert experience, and poor performance of the diagnostic model. Summary of the invention

[0005] The present invention provides a centrifugal pump fault diagnosis method, system and equipment based on digital twin, which are used to solve the technical problem that the existing centrifugal pump fault diagnosis method based on supervised learning model lacks fault data, resulting in insufficient generalization and diagnostic performance of the diagnosis model, and is difficult to meet the requirements of centrifugal pump abnormality type diagnosis in a real production environment.

[0006] In order to solve the above problems, the present invention provides a centrifugal pump fault diagnosis method based on digital twin, comprising:

[0007] Obtain historical operation data of the centrifugal pump and obtain original sample data;

[0008] Using a feature selection method based on multi-objective optimization to determine the feature parameters of the original sample data;

[0009] Establishing a digital twin model of the centrifugal pump according to the characteristic parameters, and generating first new sample data similar to the original sample data;

[0010] Using the WGAN model with gradient penalty, generating second new sample data close to the Nash equilibrium according to the original sample data and the first new sample data;

[0011] Based on a convolutional neural network, feature extraction is performed on mixed sample data consisting of the original sample data, the first new sample data, and the second new sample data to obtain an intelligent fault diagnosis model;

[0012] The operation data of the centrifugal pump is acquired in real time and input into the intelligent fault diagnosis model to perform real-time fault diagnosis on the centrifugal pump.

[0013] Furthermore, a feature selection method based on multi-objective optimization is used to determine the feature parameters of the original sample data, including:

[0014] Taking minimizing the number of characteristic parameters and maximizing the fault association accuracy as a multi-objective function, generating an initial solution set of the multi-objective function;

[0015] Calculating the multi-objective function value corresponding to each individual in the initial solution set, selecting candidate individuals according to the multi-objective function value, performing crossover and mutation operations on the candidate individuals, and generating new offspring individuals;

[0016] Iteratively performing selection, crossover and mutation operations on the new offspring individuals until a preset number of iterations is reached to obtain an optimal solution set;

[0017] The dominance relationship between individuals in the optimal solution set is evaluated, and a non-dominated solution set tending to the Pareto front is extracted according to the dominance relationship, wherein the non-dominated solution set is a characteristic parameter of the original sample data.

[0018] Further, a digital twin model of the centrifugal pump is established according to the characteristic parameters to generate first new sample data similar to the original sample data, including:

[0019] Determine static information according to the physical structure of the centrifugal pump, and select dynamic information of the actual operating environment of the centrifugal pump from the characteristic parameters;

[0020] A centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved using a finite element analysis tool to obtain first new sample data similar to the original sample data.

[0021] Furthermore, a centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved using a finite element analysis tool, including:

[0022] Creating a three-dimensional model of a centrifugal pump in a preset modeling software and importing it into a finite element analysis tool to obtain a geometric model, and defining properties for each part of the geometric model;

[0023] Using the finite element analysis tool to divide the geometric model into a plurality of basic units;

[0024] According to the actual operating condition of the centrifugal pump, boundary conditions and loads are applied to the geometric model after the defined attributes and segmentation to ensure that it is consistent with the actual operating condition of the centrifugal pump;

[0025] The response data of the model is solved according to the boundary conditions and the load to obtain first new sample data.

[0026] Further, using the WGAN model with gradient penalty, second new sample data close to the Nash equilibrium is generated according to the original sample data and the first new sample data, including:

[0027] Input random noise into the generative network of the WGAN model with gradient penalty to obtain generated data;

[0028] Randomly sampling the original sample data and the first new sample data to obtain real data;

[0029] Inputting the generated data and the real data into the discriminant network of the WGAN model to obtain the classification error of the discriminant network;

[0030] updating the parameters of the discriminant network according to the classification error, and optimizing the generative network based on the updated discriminant network;

[0031] When the loss functions of the discriminant network and the generative network converge, a trained WGAN model is obtained, and second new sample data close to the Nash equilibrium is generated based on the trained WGAN model.

[0032] Furthermore, the loss function of the generative network is expressed as:

[0033]

[0034] Among them, D(G(z(t))) represents the judgment probability output by the discriminant network, E pG(z) Represents the expected value of the true data distribution pG(z).

[0035] Furthermore, the loss function of the discriminant network is expressed as:

[0036]

[0037] Among them, x represents the original sample, z represents the input random noise, and L gpand λ represent the gradient penalty term and its weight respectively, E pG(z) Represents the expected value of the discriminant network on the generated data G(z).

[0038] The present invention also provides a centrifugal pump fault diagnosis system based on digital twin, comprising:

[0039] A data acquisition module is used to acquire historical operation data of the centrifugal pump and obtain original sample data;

[0040] A feature extraction module, used to determine the feature parameters of the original sample data by adopting a feature selection method based on multi-objective optimization;

[0041] A first generating module, used for establishing a finite element model according to the characteristic parameters, and generating first new sample data similar to the original sample data;

[0042] A second generating module is used to generate second new sample data close to the Nash equilibrium according to the original sample data and the new sample data by using the WGAN model with gradient penalty;

[0043] A diagnosis model building module, used for extracting features from mixed sample data consisting of the original sample data, the first new sample data and the second new sample data based on a convolutional neural network to obtain an intelligent fault diagnosis model;

[0044] The analysis module is used to obtain the operation data of the centrifugal pump in real time and input the data into the intelligent fault diagnosis model to perform intelligent fault diagnosis on the centrifugal pump.

[0045] The present invention also provides an electronic device, including a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the centrifugal pump fault diagnosis method based on digital twins as described in any of the above technical solutions is implemented.

[0046] The present invention also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the centrifugal pump fault diagnosis method based on digital twins as described in any of the above technical solutions is implemented.

[0047] Compared with the prior art, the beneficial effects of the present invention include: the present invention fully considers the actual situation that centrifugal pumps are mostly in normal state in actual engineering applications and lack equipment fault sample data, generates new samples similar to the original sample data through digital twin technology, enhances the diversity of training data sets when real data is scarce, and helps to improve the generalization ability of the model; through the WGAN model with gradient penalty, the sample generation process is better handled to ensure higher quality of generated samples, reduce the impact of fault sample imbalance on fault diagnosis results, and alleviate the impact of few fault samples on diagnostic efficiency. The present invention expands the number of fault samples of centrifugal pumps, reduces the impact of fault sample imbalance on fault diagnosis, improves the accuracy and efficiency of centrifugal pump fault diagnosis, and has extremely strong engineering practical value. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic flow chart of a centrifugal pump fault diagnosis method based on digital twin provided by the present invention;

[0049] Figure 2 A schematic diagram of the process of establishing the intelligent fault diagnosis model provided by the present invention;

[0050] Figure 3 A schematic structural diagram of an embodiment of a centrifugal pump fault diagnosis system based on digital twin provided by the present invention;

[0051] Figure 4 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0052] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0053] The present embodiment provides a centrifugal pump fault diagnosis method, system, electronic device and computer-readable storage device based on digital twins, which mainly include the steps of collecting operating data, selecting characteristic parameters, expanding fault samples, and building a fault diagnosis model.

[0054] like Figure 1 As shown, Figure 1 is a flow chart of the centrifugal pump fault diagnosis method based on digital twin, the method comprising:

[0055] Step S101: Acquire historical operation data of a centrifugal pump to obtain original sample data;

[0056] Step S102: using a feature selection method based on multi-objective optimization to determine feature parameters of the original sample data;

[0057] Step S103: establishing a digital twin model of the centrifugal pump according to the characteristic parameters, and generating first new sample data similar to the original sample data;

[0058] Step S104: using the WGAN model with gradient penalty, generating second new sample data that approximates the Nash equilibrium according to the original sample data and the first new sample data;

[0059] Step S105: extracting features from the mixed sample data consisting of the original sample data, the first new sample data and the second new sample data based on a convolutional neural network to obtain an intelligent fault diagnosis model;

[0060] Step S106: acquiring the operation data of the centrifugal pump in real time and inputting the data into the intelligent fault diagnosis model to perform real-time fault diagnosis on the centrifugal pump.

[0061] The centrifugal pump fault diagnosis method based on digital twin provided in this embodiment, first, obtains the operation data of the centrifugal pump, such as temperature, pressure, vibration, etc.; secondly, adopts a feature selection method based on multi-objective optimization (taking into account the accuracy and the number of characteristic parameters) to identify the most relevant characteristic parameters, and reduces the influence of redundant data on fault diagnosis; thirdly, in order to generate new samples similar to the original samples, a finite element model is established using finite element analysis software. At the same time, in order to overcome the problems of gradient vanishing or mode collapse in the traditional GAN ​​method, a generative adversarial network with Wasserstein gradient penalty (WGANwith GP) is proposed to effectively extract the original sample features and generate high-quality samples with high rationality and diversity; finally, a convolutional neural network is used to extract features from the mixed samples to realize the fault diagnosis of the centrifugal pump. The method of this embodiment generates a sufficient number of fault sample sets with a quality close to Nash equilibrium based on digital twin technology, which overcomes the problem of limited model generalization ability caused by the scarcity of samples in the traditional method, and provides a new solution for the fault diagnosis of centrifugal pumps with few fault samples.

[0062] As a preferred embodiment, in step S101, the historical operating data of the centrifugal pump is obtained by installing vibration sensors on both sides of the centrifugal pump unit on the basis of the original plant SCADA system (supervisory control and data acquisition system), and using OPC UA technology (open platform communication unified architecture) to collect the temperature, pressure, vibration and other operating data of the centrifugal pump.

[0063] As a specific embodiment, the specific method in the data acquisition process is as follows: first, a vibration displacement sensor (model LZDSL1-930) is installed on the rolling bearing seats on both sides of the centrifugal pump unit; then, the OPC UA technology is used to collect the vertical and horizontal vibration signals of the equipment and extract the pressure signal in the SCADA system; finally, the data is collected once every 1 second and stored as a sample, and each sample records 10 measured vibration data, thereby obtaining the original sample data reflecting the actual operation of the centrifugal pump.

[0064] As a preferred embodiment, in step S102, a feature selection method based on multi-objective optimization is used to determine the feature parameters of the original sample data, including:

[0065] Taking minimizing the number of characteristic parameters and maximizing the fault association accuracy as a multi-objective function, generating an initial solution set of the multi-objective function;

[0066] Calculating the multi-objective function value corresponding to each individual in the initial solution set, selecting candidate individuals according to the multi-objective function value, performing crossover and mutation operations on the candidate individuals, and generating new offspring individuals;

[0067] Iteratively performing selection, crossover and mutation operations on the new offspring individuals until a preset number of iterations is reached to obtain an optimal solution set;

[0068] The dominance relationship between individuals in the optimal solution set is evaluated, and a non-dominated solution set tending to the Pareto front is extracted according to the dominance relationship, wherein the non-dominated solution set is a characteristic parameter of the original sample data.

[0069] As a specific embodiment, first, the initialization algorithm is used to obtain the initial solution set Pop of the objective function. Assuming there are w objective functions, FS' w Represents the parameter combination corresponding to the wth target Next, the initial population is optimized and the function is evaluated. represents the function value corresponding to the wth individual; then, the selection operator is used to select the next generation of individuals participating in the evolution; the selection operator here can be roulette selection, tournament selection or ranking selection, etc. Repeat the above steps until the number of iterations is reached, and output the best solution set of the problem; finally, through continuous optimization, the solution set of the problem constantly approaches the ideal Pareto frontier, and the non-dominated solution set (NDSs) of the problem is obtained, that is, the characteristic parameters of the original sample data.

[0070] The most relevant feature parameters are identified using a feature selection method based on multi-objective optimization to reduce the impact of redundant data on fault diagnosis.

[0071] After determining the characteristic parameters that affect the centrifugal pump failure, a digital twin model of the centrifugal pump is established to simulate the operation of the centrifugal pump, and new sample data is generated to expand the fault samples. The sample features are further extracted through the Wasserstein generative adversarial network (GAN) algorithm with gradient penalty, and high-quality samples with high rationality and diversity are generated.

[0072] As a preferred embodiment, in step S103, a digital twin model of the centrifugal pump is established according to the characteristic parameters to generate first new sample data similar to the original sample data, including:

[0073] Determine static information according to the physical structure of the centrifugal pump, and select dynamic information of the actual operating environment of the centrifugal pump from the characteristic parameters;

[0074] A centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved using a finite element analysis tool to obtain first new sample data similar to the original sample data.

[0075] Furthermore, a centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved using a finite element analysis tool, including:

[0076] Creating a three-dimensional model of a centrifugal pump in a preset modeling software and importing it into a finite element analysis tool to obtain a geometric model, and defining properties for each part of the geometric model;

[0077] Using the finite element analysis tool to divide the geometric model into a plurality of basic units;

[0078] According to the actual operating condition of the centrifugal pump, boundary conditions and loads are applied to the geometric model after the defined attributes and segmentation to ensure that it is consistent with the actual operating condition of the centrifugal pump.

[0079] As a specific embodiment, the preset modeling software is Solidworks, and the finite element analysis tool is Comsol. The specific process of using the finite element analysis tool to solve the response data of the centrifugal pump digital twin model is as follows:

[0080] The first step is to create a geometric model: import the 3D model created in Solidworks into Comsol (finite element software) and define material properties for each part in the geometric model, such as elastic modulus, Poisson's ratio, density, thermal conductivity, etc.

[0081] The second step is meshing: the geometric model is automatically divided into a series of small units (such as triangles, quadrilaterals, tetrahedrons or hexahedrons) using Comsol software, and the size, shape and distribution of the mesh are manually adjusted to ensure accuracy and computational efficiency in key areas.

[0082] Step 3: Boundary conditions and loads: According to the actual working conditions (operation data collected in step S101), appropriate boundary conditions and loads are applied to the model to ensure that they conform to the actual conditions and simulate the state of the analysis object as accurately as possible.

[0083] The fourth step is to generate new samples: according to the defined model, material properties, mesh division, boundary conditions and load information, the response of the model (such as displacement, stress, temperature, etc.) is solved, and a new sample similar to the original sample is generated to obtain the first new sample data.

[0084] As a preferred embodiment, in step S104, in order to overcome the problems of gradient vanishing or mode collapse in the traditional GAN ​​method, this embodiment uses a generative adversarial network with Wasserstein gradient penalty (WGAN with GP) to generate a sufficient number of fault sample sets with quality close to Nash equilibrium. The specific method is:

[0085] Input random noise into the generative network of the WGAN model with gradient penalty to obtain generated data;

[0086] Randomly sampling the original sample data and the first new sample data to obtain real data;

[0087] Inputting the generated data and the real data into the discriminant network of the WGAN model to obtain the classification error of the discriminant network;

[0088] updating the parameters of the discriminant network according to the classification error, and optimizing the generative network based on the updated discriminant network;

[0089] When the loss functions of the discriminant network and the generative network converge, a trained WGAN model is obtained, and second new sample data close to the Nash equilibrium is generated based on the trained WGAN model.

[0090] It should be noted here that WGAN (Wasserstein Generative Adversarial Network) is a variant of the Generative Adversarial Network (GAN). WGAN uses the Wasserstein distance (or Earth Mover distance) to measure the difference between the generated sample distribution and the real sample distribution. This distance has better mathematical properties and can provide a smoother loss function, which helps improve the model training process.

[0091] As a specific embodiment, the sample generation process of the generative adversarial network with Wasserstein gradient penalty (WGAN with GP) is as follows:

[0092] Step 1: The generator generates "fake" samples: randomly sample from the noise data distribution, input the generator G, and obtain a set of fake data, denoted as G(z).

[0093] The loss function of the generator G during training is shown in formula (1). Compared with the traditional GAN ​​method, the generator with constrained Wasserstein distance can obtain more similar samples.

[0094]

[0095] In formula (1), D(G(z)) is the judgment probability output by the discriminator D.

[0096] Loss Function The optimization goal is to continuously reduce G(z), that is, to make G(z) deceive the discriminator D as much as possible, so that the discriminator thinks it is a real sample.

[0097] In order to solve some problems in WGAN (such as weight clipping), compared with traditional GAN ​​methods, we introduced gradient penalty in WGAN to help ensure the Lipschitz (used to describe the smoothness of a function within its domain) continuity of the discriminator and further improve the stability of the system.

[0098] The loss function of the discriminant network is expressed as:

[0099]

[0100] In formula (2), x represents the original sample, z represents the input random noise, and L gp and λ represent the gradient penalty term and its weight respectively, E pG(z) It represents the expected value of the discriminant network on the generated data G(z), which indicates the ability of the discriminator to identify "fake" samples. The optimization goal of the discriminant network is to maximize the score of the real data and minimize the score of the generated data. At the same time, by penalizing the gradient of the discriminator, the smoothness of its change in the input space is ensured, thereby improving the stability of the model.

[0101] Step 2: Distinguish “true” and “false” samples: Randomly sample from the real data distribution as real data. The previously generated data is used as the input of the discriminant network. Therefore, the input of the discriminant model is two types of data, “true” data and generated “false” data. The output value of the discriminant network is the probability that the input belongs to the real data, “true” is recorded as 1, and “false” is 0.

[0102] Step 3: Update the parameters of the discriminant network: According to the classification error, update the parameters of the discriminant network to improve its classification accuracy.

[0103] Step 4: Generate realistic false data: The generation network updates its own parameters based on the feedback information from the discriminant network to generate more realistic false data.

[0104] As a preferred embodiment, in step S105, a convolutional neural network (CNN) is used to construct an intelligent fault diagnosis model, and feature extraction is performed on the mixed sample data consisting of the original sample data, the first new sample data and the second new sample data, including extracting local features by using a convolution kernel; reducing the feature size through a pooling layer to retain important information while reducing computational complexity; introducing nonlinearity through a ReLU activation function, etc., to finally obtain an intelligent fault diagnosis model that realizes automatic diagnosis of centrifugal pump faults.

[0105] like Figure 2 As shown, Figure 2 The processing of each of the above steps is presented in detail. The establishment process of the above intelligent fault diagnosis model includes four steps: collecting operating data, selecting characteristic parameters, expanding fault samples, and building a fault diagnosis model. By introducing digital twin technology, the number of fault samples of the centrifugal pump is expanded, the impact of the imbalance of fault samples on fault diagnosis is reduced, and the reliability of the centrifugal pump unit is improved.

[0106] This embodiment also provides a centrifugal pump fault diagnosis system based on digital twins, such as Figure 3 As shown, the centrifugal pump fault diagnosis system 300 based on digital twin includes:

[0107] The data acquisition module 301 is used to acquire the historical operation data of the centrifugal pump and obtain the original sample data;

[0108] A feature extraction module 302 is used to determine the feature parameters of the original sample data by using a feature selection method based on multi-objective optimization;

[0109] A first generating module 303, used to establish a finite element model according to the characteristic parameters, and generate first new sample data similar to the original sample data;

[0110] A second generating module 304 is used to generate second new sample data close to the Nash equilibrium according to the original sample data and the first new sample data by using the WGAN model with gradient penalty;

[0111] A diagnosis model building module 305 is used to extract features from the mixed sample data consisting of the original sample data, the first new sample data and the second new sample data based on a convolutional neural network to obtain an intelligent fault diagnosis model;

[0112] The analysis module 306 is used to obtain the operation data of the centrifugal pump in real time and input the data into the intelligent fault diagnosis model to perform intelligent fault diagnosis on the centrifugal pump.

[0113] like Figure 4 As shown, the above-mentioned centrifugal pump fault diagnosis method based on digital twin, the present invention also provides an electronic device 400 accordingly, which can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer and a server. The electronic device includes a processor 401, a memory 402 and a display 403.

[0114] In some embodiments, the memory 402 may be an internal storage unit of a computer device, such as a hard disk or memory of a computer device. In other embodiments, the memory 402 may also be an external storage device of a computer device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the computer device. Further, the memory 402 may also include both an internal storage unit of a computer device and an external storage device. The memory 402 is used to store application software and various types of data installed on the computer device, such as program codes for installing the computer device. The memory 402 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a centrifugal pump fault diagnosis method program 404 based on digital twins is stored on the memory 402, and the centrifugal pump fault diagnosis method program 404 based on digital twins can be executed by the processor 401, thereby realizing a centrifugal pump fault diagnosis method based on digital twins in each embodiment of the present invention.

[0115] In some embodiments, the processor 401 can be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 402, such as executing a centrifugal pump fault diagnosis method program based on digital twins.

[0116] In some embodiments, the display 403 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 403 is used to display information on the computer device and to display a visual user interface. The components 401-403 of the computer device communicate with each other through a system bus.

[0117] An embodiment of the present invention also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the centrifugal pump fault diagnosis method based on digital twins as described in any of the above technical solutions is implemented.

[0118] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A centrifugal pump fault diagnosis method based on digital twin, characterized in that: include; Obtain historical operation data of the centrifugal pump and obtain original sample data; Using a feature selection method based on multi-objective optimization to determine the feature parameters of the original sample data; Establishing a digital twin model of the centrifugal pump according to the characteristic parameters, and generating first new sample data similar to the original sample data; Using the WGAN model with gradient penalty, generating second new sample data close to the Nash equilibrium according to the original sample data and the first new sample data; Based on a convolutional neural network, feature extraction is performed on mixed sample data consisting of the original sample data, the first new sample data, and the second new sample data to obtain an intelligent fault diagnosis model; The operation data of the centrifugal pump is acquired in real time and input into the intelligent fault diagnosis model to perform real-time fault diagnosis on the centrifugal pump.

2. The centrifugal pump fault diagnosis method based on digital twin according to claim 1 is characterized in that: The feature selection method based on multi-objective optimization is used to determine the feature parameters of the original sample data, including: Taking minimizing the number of characteristic parameters and maximizing the fault association accuracy as a multi-objective function, generating an initial solution set of the multi-objective function; Calculating the multi-objective function value corresponding to each individual in the initial solution set, selecting candidate individuals according to the multi-objective function value, performing crossover and mutation operations on the candidate individuals, and generating new offspring individuals; Iteratively performing selection, crossover and mutation operations on the new offspring individuals until a preset number of iterations is reached to obtain an optimal solution set; The dominance relationship between individuals in the optimal solution set is evaluated, and a non-dominated solution set tending to the Pareto front is extracted according to the dominance relationship, wherein the non-dominated solution set is a characteristic parameter of the original sample data.

3. The centrifugal pump fault diagnosis method based on digital twin according to claim 1 is characterized in that: Establishing a digital twin model of the centrifugal pump according to the characteristic parameters, and generating first new sample data similar to the original sample data, including: Determine static information according to the physical structure of the centrifugal pump, and select dynamic information of the actual operating environment of the centrifugal pump from the characteristic parameters; A centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved using a finite element analysis tool to obtain first new sample data similar to the original sample data.

4. The centrifugal pump fault diagnosis method based on digital twin according to claim 3 is characterized in that: A centrifugal pump digital twin model is established based on the static information and the dynamic information, and the response data of the centrifugal pump digital twin model is solved by using a finite element analysis tool, including: Creating a three-dimensional model of a centrifugal pump in a preset modeling software and importing it into a finite element analysis tool to obtain a geometric model, and defining properties for each part of the geometric model; Using the finite element analysis tool to divide the geometric model into a plurality of basic units; According to the actual operating condition of the centrifugal pump, boundary conditions and loads are applied to the geometric model after the defined attributes and segmentation to ensure that it is consistent with the actual operating condition of the centrifugal pump; The response data of the model is solved according to the boundary conditions and the load to obtain first new sample data.

5. The centrifugal pump fault diagnosis method based on digital twin according to claim 1, characterized in that: Using the WGAN model with gradient penalty, generating second new sample data close to the Nash equilibrium according to the original sample data and the first new sample data, including: Input random noise into the generative network of the WGAN model with gradient penalty to obtain generated data; Randomly sampling the original sample data and the first new sample data to obtain real data; Inputting the generated data and the real data into the discriminant network of the WGAN model to obtain the classification error of the discriminant network; updating the parameters of the discriminant network according to the classification error, and optimizing the generative network based on the updated discriminant network; When the loss functions of the discriminant network and the generative network converge, a trained WGAN model is obtained, and second new sample data close to the Nash equilibrium is generated based on the trained WGAN model.

6. The centrifugal pump fault diagnosis method based on digital twin according to claim 5 is characterized in that: The loss function of the generative network is expressed as: Among them, D(G(z(t))) represents the judgment probability output by the discriminant network, E pG(z) Represents the expected value of the true data distribution pG(z).

7. The centrifugal pump fault diagnosis method based on digital twin according to claim 5 is characterized in that: The loss function of the discriminant network is expressed as: Among them, x represents the original sample, z represents the input random noise, and L gp and λ represent the gradient penalty term and its weight respectively, E pG(z) Represents the expected value of the discriminant network on the generated data G(z).

8. A centrifugal pump fault diagnosis system based on digital twin, characterized in that: include: A data acquisition module is used to acquire historical operation data of the centrifugal pump and obtain original sample data; A feature extraction module, used to determine the feature parameters of the original sample data by adopting a feature selection method based on multi-objective optimization; A first generating module, used for establishing a finite element model according to the characteristic parameters, and generating first new sample data similar to the original sample data; A second generating module is used to generate second new sample data close to Nash equilibrium according to the original sample data and the first new sample data by using a WGAN model with gradient penalty; A diagnosis model building module, used for extracting features from mixed sample data consisting of the original sample data, the first new sample data and the second new sample data based on a convolutional neural network to obtain an intelligent fault diagnosis model; The analysis module is used to obtain the operation data of the centrifugal pump in real time and input the data into the intelligent fault diagnosis model to perform intelligent fault diagnosis on the centrifugal pump.

9. An electronic device, characterized in that: It includes a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the centrifugal pump fault diagnosis method based on digital twins as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the centrifugal pump fault diagnosis method based on digital twins as described in any one of claims 1 to 7 is implemented.

Images