Method and system for estimating remaining service life of underwater throttle valve based on digital twin

Abstract

The present invention relates to the technical field of underwater throttle valve maintenance, and specifically, to a method and system for estimating the remaining service life of an underwater throttle valve based on digital twins. The method includes: establishing a mathematical model of the erosion state of an underwater throttle valve; measuring and obtaining complex and diverse data of the erosion environment; substituting real-time data into the above model to construct a digital twin model of an underwater throttle valve; simulating and calculating the operating state and erosion state of the underwater throttle valve; establishing a digital twin model of the performance degradation of an underwater throttle valve; predicting the remaining service life; comparing with the remaining service life threshold and grading the degree of erosion; completing evaluation and prediction, and completing maintenance or repair according to the results. The design of the present invention can improve the prediction accuracy of the underwater throttle valve by establishing a fault prediction and erosion degradation model of the underwater throttle valve, thereby facilitating the early arrangement of production and operation safety maintenance of the underwater throttle valve, reducing shutdown costs, and improving the accuracy of the health index and remaining service life distribution prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of underwater throttle valve maintenance, and in particular to a method and system for estimating the remaining service life of an underwater throttle valve based on digital twins. Background Art

[0002] Underwater throttle valves operate underwater for a long time, especially those used in the ocean. They are very susceptible to corrosion damage caused by external factors, such as seawater pressure, flow rate, temperature, carbon dioxide partial pressure and underwater disturbance of seawater, which can cause corrosion of the underwater throttle valve body; at the same time, the inner wall of most underwater throttle valves is eroded, corroded and accelerated by the flow of oil and water mixed with sand, resulting in thinning of the underwater throttle valve wall structure. Under the action of erosion, the geometric contour shape of the wall of the underwater throttle valve will change with the increase of solid particle erosion exposure. Therefore, the change in surface contour will change the flow field, thereby redirecting the particles to another erosion "hotspot". The underwater pipeline structure is complex and covers a wide area. If the underwater throttle valves are inspected and maintained manually one by one in daily life, not only is the workload huge and time-consuming, but it is also very easy to make omissions and misjudgments, which brings trouble to the maintenance of the underwater throttle valves. If the underwater throttle valves with problems cannot be discovered, maintained or repaired and replaced in time, it may cause huge losses.

[0003] Digital twins make full use of data such as physical models, sensor updates, and operation history, integrate multi-disciplinary, multi-physical, multi-scale, and multi-probability simulation processes, and complete mapping in virtual space, thereby reflecting the entire life cycle of the corresponding physical equipment. If it is possible to combine digital twin technology based on a large amount of real data and experimental data, it is expected to predict and track the erosion of underwater throttle valves and evaluate their remaining service life, so that maintenance personnel can perform targeted maintenance or repairs on underwater throttle valves in the corresponding areas. However, there is currently no technology that uses digital twins to estimate the remaining service life of underwater throttle valves.

[0004] In view of this, we proposed a method and system for estimating the remaining service life of underwater throttle valves based on digital twins. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for estimating the remaining service life of an underwater throttle valve based on digital twins to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, one of the purposes of the present invention is to provide a method for estimating the remaining service life of an underwater throttle valve based on digital twins, comprising the following steps:

[0007] S1. Obtain multiple sets of valve body metal material resistance values ​​and corresponding erosion amounts of valve body metal resistance values ​​through underwater throttle valve erosion experiments. A nonlinear regression algorithm is used based on the data obtained from the metal corrosion experiments to establish a mathematical model of underwater throttle valve erosion state.

[0008] S2. The complex and diverse data of the underwater throttle valve erosion environment are measured by the data acquisition module and transmitted to the database of the mobile workstation for storage and management, so as to realize the real-time data-driven finite element simulation model and build a physical model database;

[0009] S3, substituting the real-time data of the measured environmental data and erosion measurement data into the mathematical model of the erosion state of the underwater throttle valve to construct a digital twin model of the underwater throttle valve;

[0010] S4. The operation state and erosion state of the underwater throttle valve are simulated and calculated by using the underwater throttle valve digital twin model. The operation state data and erosion state data of the underwater throttle valve obtained by the simulation calculation are stored in the twin data unit of the database.

[0011] S5. A dynamic evolution model of underwater throttle valve performance degradation is established through a dynamic Bayesian model, which is implemented through a particle filter PF algorithm to establish a complete digital twin model of underwater throttle valve performance degradation;

[0012] S6. Collecting the underwater throttle valve direct inspection process data through the data acquisition module and saving it in the evaluation data unit of the database; predicting the underwater throttle valve remaining life using the underwater throttle valve remaining life prediction algorithm according to the data in the evaluation data unit, and saving the remaining life prediction results obtained in the evaluation data unit;

[0013] S7, calling the remaining life threshold data of the underwater throttle valve preset in the database, and then comparing the remaining life calculation result with the remaining life threshold. If the remaining life calculation result is lower than 10%, 50%, and 80% of the remaining life threshold, the erosion degree of the underwater throttle valve is divided into three levels: A-mild erosion, B-moderate erosion, and C-severe erosion respectively;

[0014] S8. Output the erosion level of the underwater throttle valve in time, complete the evaluation and prediction of the erosion condition of the underwater throttle valve foundation; find the corrosion area according to the positioning of the monitoring point, and complete the maintenance or repair of the area according to the evaluation results.

[0015] As a further improvement of the technical solution, in S3, the real-time data of the measured environmental data and erosion measurement data are substituted into the underwater throttle valve erosion state mathematical model, and the specific steps of constructing the underwater throttle valve digital twin model are as follows:

[0016] S3.1. Use nonlinear regression algorithm to establish a mathematical model of underwater throttle valve erosion state;

[0017] S3.2. A common characterization model of the underwater throttle valve performance degradation process is established by using a nonlinear regression algorithm. The evolution law of the state space model over time is described by a dynamic Bayesian network, and a dynamic evolution model of the underwater throttle valve performance degradation is established.

[0018] S3.3, establishing a dynamic evolution model of underwater throttle valve based on dynamic Bayesian neural network, that is, a dynamic evolution model of underwater throttle valve performance degradation established based on DBN model;

[0019] S3.4. Use the particle filter PF algorithm to track the degradation state and predict the remaining life of the underwater throttle valve, thereby completing the establishment of a digital twin model of the underwater throttle valve performance degradation;

[0020] S3.5. Establish a digital twin model for underwater throttle valve performance degradation and remaining life prediction based on particle filter PF approximate inference algorithm: The basic idea of ​​PF algorithm is to extract a series of weighted particles to realize the posterior evaluation of the quantity to be estimated. In this process, the weight of each particle is iteratively updated by continuously comparing the difference between the estimated value and the observed value, and finally approximate the true posterior distribution;

[0021] S3.6. Use reinforcement learning algorithms to compare perception data with simulation data, and adjust parameters to maintain consistency between the virtual and the real.

[0022] As a further improvement of the technical solution, in S3.1, the specific method of establishing the mathematical model of the underwater throttle valve erosion state by using a nonlinear regression algorithm comprises the following steps:

[0023] S3.1.1. Obtain multiple sets of experimental data through underwater throttle valve erosion experiments, the experimental data being target metal resistance values ​​and erosion amounts corresponding to the target metal resistance values;

[0024] S3.1.2, correct the experimental data obtained in step S3-2-1 to eliminate systematic errors;

[0025] S3.1.3. Based on the data corrected in step S3-2-2, a fitting function model is constructed using a nonlinear regression algorithm; an interpolation method is used to solve the model and obtain a fitting function curve.

[0026] As a further improvement of the technical solution, in S3.2, the specific method for establishing the commonality characterization model is:

[0027] Since the external causes of underwater throttle valve corrosion are mostly factors such as seawater pressure, flow rate, temperature, carbon dioxide partial pressure and underwater disturbance in seawater, an external impact model of underwater throttle valve in marine environment is established, and the external impact process of underwater throttle valve in marine environment is modeled as a Poisson process. 2 ≥0, there are:

[0028]

[0029] Where n is the number of times the underwater throttle valve is impacted by the marine environment, N c (t 1 +t 2 ) is t 1 +t 2 The number of external impacts of the marine environment in a time period, N c (t 1 ) is t 1 The number of external impacts of the marine environment in a period of time, P{N c (t 1 +t 2 )-N c (t 1 )=n} is the 2 The probability of n external shocks to the marine environment occurring in a time period, λ is the parameter of the Poisson distribution, which is determined by historical data;

[0030] The inner wall of the underwater throttle valve is eroded, corroded and accelerated by the flow of oil and water mixed with sand, which causes the wall structure of the underwater throttle valve to become thinner. Under the action of erosion, the geometric shape of the wall of the underwater throttle valve will change with the increase of the erosion exposure of solid particles. Therefore, the change of surface contour will change the flow field, thereby redirecting the particles to another erosion "hot spot".

[0031] In addition, the local impact angle of the particles will change with time, which may lead to a large change in the local erosion rate. Therefore, the amplitude of the thinning of the underwater throttle valve body wall caused by erosion can be used as an indicator of the performance degradation of the underwater throttle valve based on digital twins.

[0032] As a further improvement of the technical solution, in S3.3, the purpose of establishing a DBN-based dynamic evolution model is: given the model parameter θ k =(A k ,B k ) and a set of underwater throttle valve observation data Q 0:k =(q 0 ,q 1 ,…,q k ) is used to estimate the posterior distribution of the degradation state of the underwater throttle valve. The calculation formula is as follows:

[0033]

[0034] in, represents the i-th particle; δ(·) is the Dirac function; represents the weight corresponding to the i-th particle, which can be calculated by the following formula:

[0035]

[0036]

[0037] In order to overcome the particle degradation problem caused by too small weights of some particles, a resampling strategy is used to reset the weights of each particle:

[0038]

[0039] is equivalent to:

[0040]

[0041] In the formula, represents the particle after resampling at time k; n i Represents particles After resampling, a new particle set is generated The number of times the number of copies is made;

[0042] Further predict the degradation state at the future time step k+t (t=1, 2, ...)

[0043]

[0044] in, Available through We get: where V is the uncertainty of the model;

[0045] The RUL of the underwater throttle valve can be defined as the time interval between the moment when the predicted future degradation state first reaches the preset failure threshold and the current moment under the conditions of the current degradation state and model parameters. The RUL prediction of the turbine disk can be achieved by extending the degradation trajectory to reach the preset failure threshold. The calculation formula is as follows:

[0046]

[0047] is the RUL prediction value of the ith particle; t is the time step, ε is the failure threshold;

[0048] Finally, the predicted value of RUL can be obtained by weighted averaging:

[0049]

[0050] As a further improvement of the technical solution, in S3.6, the specific method of using a reinforcement learning algorithm to compare the perception data with the simulation data and adjusting parameters to achieve virtual-real consistency includes the following steps:

[0051] S3.6.1. By compiling the physical property model interface, the injection of perception data and the output of data sets are realized to build a two-way interactive mechanism, and the production data of the underwater throttle valve, such as throttle valve material, throttle valve opening, fluid mass flow rate, fluid density, solid particle mass flow rate and other parameters, are used as the input of the model;

[0052] S3.6.2. Initialize the digital twin model, use the reinforcement learning algorithm in the correction space, compare the perception data with the simulation data, and maintain consistency through parameter adjustment;

[0053] S3.6.3. Initialize the digital twin model, manually select the correction parameters and target variables of the digital twin model in the correction space, conduct multiple groups of experiments, obtain the sensitivity relationship between each correction parameter and the target variable, and then screen the correction parameters to determine the parameters with higher sensitivity as the correction parameters;

[0054] S3.6.4. With the correction parameter as the independent variable of the test and the target variable as the dependent variable of the test, based on the idea of ​​experimental design, multiple regression or difference fitting is performed on the correction parameter and the target variable to construct a response surface model. The response surface model is a mathematical model that expresses the relationship between the independent variable and the dependent variable in the form of a mathematical function, thereby constructing the objective function of the target variable.

[0055] S3.6.5. Find the correction parameters that can meet the objective function in the response surface space, replace the original twin model parameters with the corrected parameters, realize the mapping of the twin model, establish the fault prediction and erosion degradation model of the underwater throttle valve, improve the accuracy of the health index and remaining life distribution prediction, and finally realize the establishment and consistency maintenance of the two-way interactive twin model of the underwater production system.

[0056] As a further improvement of the technical solution, in S4, the operation state and erosion state of the underwater throttle valve are simulated and calculated by the underwater throttle valve digital twin model, and the simulation calculation includes but is not limited to the erosion rate calculation, wherein the erosion rate algorithm formula is as follows:

[0057]

[0058] Where: R erosion is the erosion rate, (kg / (m 2 ·s)); Nparticle is the number of solid particles colliding with the wall, dimensionless; is the mass flow rate of solid particles, kg / s; f(α) is a function of the solid particle collision angle α; v is the velocity of the solid particles relative to the wall, m / s; b(v) is a function of the relative velocity of the solid particles; C(d p ) is a function of the solid particle size, d p is the diameter of the particle, m; A face is the unit area of ​​the impact wall;

[0059] And then:

[0060] C(d p )=1.599HB -0.59 F s ×10 -7 ;

[0061]

[0062] The second object of the present invention is to provide a system for estimating the remaining service life of an underwater throttle valve based on digital twins, which is used to implement the operation process of the above-mentioned method for estimating the remaining service life of an underwater throttle valve based on digital twins, including a data acquisition module, a database and a numerical simulation unit; the signal output end of the data acquisition module is connected to the signal input end of the database, and the database is connected to the numerical simulation unit signal; wherein:

[0063] The data acquisition module is used to acquire data such as multiple sets of valve body metal material resistance values ​​and valve body metal resistance values ​​corresponding to erosion amounts obtained through underwater throttle valve erosion experiments, data on the complex and diverse underwater throttle valve erosion environment, and underwater throttle valve direct inspection process data, and can store such data in the database;

[0064] The database is used to classify and store various types of data collected and acquired by the data acquisition module, various types of simulation data acquired by the numerical simulation unit, and user presets;

[0065] The numerical simulation unit is used to construct relevant mathematical models and digital twin models of the underwater throttle valve based on the real data of the underwater throttle valve obtained by measurement and the experimental data obtained through the underwater throttle valve erosion experiment, using certain algorithms and models, and to estimate the remaining service life of the underwater throttle valve through numerical simulation and deep learning.

[0066] As a further improvement of the technical solution, the database includes a physical model database, a twin data unit and an evaluation data unit; the physical model database, the twin data unit and the evaluation data unit run in parallel and are sequentially communicated; wherein:

[0067] The physical model database is used to store and manage the complex and diverse data of underwater throttle valve erosion environment including environmental data and erosion measurement data obtained by the data acquisition module;

[0068] The twin data unit is used to store and manage the operation status data and erosion status data of the underwater throttle valve obtained by simulation calculation by the numerical simulation unit;

[0069] The evaluation data unit is used to save and manage preset remaining life threshold data, direct inspection process data of the underwater throttle valve collected by the data acquisition module, and remaining life prediction result data obtained by predicting the remaining life of the underwater throttle valve using the underwater throttle valve remaining life prediction algorithm through the numerical simulation unit. By comparing the remaining life calculation result with the preset remaining life threshold data, the evaluation and prediction of the erosion condition of the underwater throttle valve foundation can be realized.

[0070] As a further improvement of the technical solution, the numerical simulation unit includes an underwater throttle valve erosion state mathematical model and an underwater throttle valve digital twin model; the underwater throttle valve erosion state mathematical model and the underwater throttle valve digital twin model run in parallel and are signal connected; wherein:

[0071] The underwater throttle valve erosion state mathematical model is used to perform mathematical simulation on the physical state of the underwater throttle valve based on experimental data;

[0072] The underwater throttle valve erosion state mathematical model includes a common characterization model and an underwater throttle valve performance degradation dynamic evolution model; the common characterization model is used to simulate the common characterization state of the underwater throttle valve in combination with external causes that may cause corrosion of the underwater throttle valve body, so as to determine an indicator value that can be used as an underwater throttle valve performance degradation; the underwater throttle valve performance degradation dynamic evolution model is used to describe the variation law of degradation state and parameters over time and the transmission process of model uncertainty;

[0073] The underwater throttle valve digital twin model is constructed by substituting the real-time data of the measured environmental data and erosion measurement data into the underwater throttle valve erosion state mathematical model, and is used to simulate and calculate the operating state and erosion state of the underwater throttle valve;

[0074] The underwater throttle valve digital twin model includes an underwater throttle valve performance degradation digital twin model and an underwater throttle valve performance degradation and remaining life prediction digital twin model; the underwater throttle valve performance degradation digital twin model is used to simulate and calculate the performance degradation state of the underwater throttle valve; the underwater throttle valve performance degradation and remaining life prediction digital twin model is used to track the degradation state and predict the remaining life of the underwater throttle valve using a particle filter PF algorithm.

[0075] The third object of the present invention is to provide a system operation platform device, including a processor, a memory, and a computer program stored in the memory and running on the processor, and the processor is used to implement the steps of the above-mentioned digital twin-based underwater throttle valve remaining service life estimation system and method when executing the computer program.

[0076] The fourth object of the present invention is to provide a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, it implements the steps of the above-mentioned underwater throttle valve remaining service life estimation system and method based on digital twin.

[0077] Compared with the prior art, the present invention has the following beneficial effects:

[0078] In the underwater throttle valve remaining service life estimation method and system based on digital twin, by establishing the fault prediction and erosion degradation model of the underwater throttle valve, the prediction accuracy of the underwater throttle valve can be improved, so as to facilitate the advance arrangement of the production operation safety maintenance of the underwater throttle valve, reduce the shutdown cost, and improve the accuracy of the health index and the remaining service life distribution prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] Figure 1 It is an exemplary overall method flow diagram of the present invention;

[0080] Figure 2 is a schematic diagram of the dynamic evolution of an exemplary underwater throttle valve performance degradation dynamic evolution model of the present invention;

[0081] Figure 3 This is a schematic diagram of the principle flow chart for maintaining virtual-real consistency in the present invention;

[0082] Figure 4 is a schematic diagram of an exemplary system device structure in the present invention;

[0083] Figure 5 It is a schematic diagram of the structure of an exemplary electronic computer platform device in the present invention.

[0084] In the figure:

[0085] 1. Data acquisition module;

[0086] 2. Database; 21. Physical model database; 22. Twin data unit; 23. Evaluation data unit;

[0087] 3. Numerical simulation unit; 31. Mathematical model of underwater throttle valve erosion state; 311. Common characterization model; 312. Dynamic evolution model of underwater throttle valve performance degradation; 32. Digital twin model of underwater throttle valve; 321. Digital twin model of underwater throttle valve performance degradation; 322. Digital twin model of underwater throttle valve performance degradation and remaining life prediction. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0089] Example 1

[0090] like Figure 1-Figure 5 As shown, this embodiment provides a method for estimating the remaining service life of an underwater throttle valve based on digital twins, comprising the following steps:

[0091] S1. Obtain multiple sets of valve body metal material resistance values ​​and valve body metal resistance values ​​corresponding to erosion amounts through underwater throttle valve erosion experiments. Based on the data obtained from the metal corrosion experiments, a nonlinear regression algorithm is used to establish an underwater throttle valve erosion state mathematical model 3;

[0092] S2, obtaining complex and diverse data of underwater throttle valve erosion environment through data acquisition module 1 and transmitting it to database 2 of mobile workstation for storage and management, realizing real-time data-driven finite element simulation model, and constructing physical model database 21 at the same time;

[0093] S3, substituting the real-time data of the measured environmental data and erosion measurement data into the underwater throttle valve erosion state mathematical model 31, and constructing the underwater throttle valve digital twin model 32;

[0094] S4, simulating and calculating the operating state and erosion state of the underwater throttle valve through the underwater throttle valve digital twin model 32, and storing the operating state data and erosion state data of the underwater throttle valve obtained by the simulation and calculation in the twin data unit 22 of the database 2;

[0095] S5. Establishing a dynamic evolution model 312 of underwater throttle valve performance degradation through a dynamic Bayesian model, and implementing it through a particle filter PF algorithm, thereby establishing a complete underwater throttle valve performance degradation digital twin model 321;

[0096] S6, collecting the underwater throttle valve direct inspection process data through the data acquisition module 1 and saving it in the evaluation data unit 23 of the database 2; using the underwater throttle valve remaining life prediction algorithm to predict the underwater throttle valve remaining life according to the data in the evaluation data unit 23, and the remaining life prediction results obtained are all saved in the evaluation data unit 23;

[0097] S7, calling the remaining life threshold data of the underwater throttle valve preset in the database 2, and then comparing the remaining life calculation result with the remaining life threshold. If the remaining life calculation result is lower than 10%, 50%, and 80% of the remaining life threshold, the erosion degree of the underwater throttle valve is divided into three levels: A-mild erosion, B-moderate erosion, and C-severe erosion respectively;

[0098] S8. Output the erosion level of the underwater throttle valve in time, complete the evaluation and prediction of the erosion condition of the underwater throttle valve foundation; find the corrosion area according to the positioning of the monitoring point, and complete the maintenance or repair of the area according to the evaluation results.

[0099] Among them, in S2, the complex and diverse data of underwater throttle valve erosion environment include environmental data and erosion measurement data;

[0100] In S5, the underwater throttle valve performance degradation dynamic evolution model 312 is used to describe the variation of degradation state and parameters over time and the transmission process of model uncertainty.

[0101] Specifically, the overall estimation method can be described as follows: a nonlinear regression algorithm is used based on the data obtained from the metal material erosion experiment to establish a mathematical model of the erosion state; the real-time data of the metal resistance at each measuring point on the underwater throttle valve is obtained through the data acquisition module; the real-time data is substituted into the erosion state mathematical model to obtain the amount of metal erosion at each measuring point; then an erosion digital twin model is established based on the underwater throttle valve body data and environmental data; the application module uses the pipeline twin data model to perform simulation operation to obtain operation status data and corrosion status data; the erosion digital twin model is used to evaluate the basic erosion level of the underwater throttle valve; the degree of erosion of the underwater throttle valve is divided into levels according to the amount of erosion of the underwater throttle valve; and the underwater throttle valve is output in a timely manner. The throttle valve erosion grade is used to evaluate and predict the underwater throttle valve erosion condition. The application module is used for preliminary evaluation, and the remaining life of the underwater throttle valve is predicted based on the preliminary evaluation results. When the calculated result or the predicted result is less than the threshold, the underwater throttle valve erosion grade assessment result is used to make maintenance or repair decisions based on the evaluation and prediction information of the monitoring point. Finally, the reinforcement learning algorithm is used to compare the perception data with the simulation data, and the consistency is maintained through parameter adjustment. The correction parameters and target variables are subjected to multivariate regression or difference fitting, and a response surface model is constructed. The error between the twin model and the experimental model is reduced through the response surface construction technology to realize the mapping of the twin model, thereby achieving the virtual-real consistency maintenance of the two-way interactive digital twin of the underwater throttle valve.

[0102] In this embodiment, in S3, the real-time data of the measured environmental data and erosion measurement data are substituted into the underwater throttle valve erosion state mathematical model 31, and the specific steps of constructing the underwater throttle valve digital twin model 32 are as follows:

[0103] S3.1. Use nonlinear regression algorithm to establish a mathematical model of underwater throttle valve erosion state 31;

[0104] S3.2, using a nonlinear regression algorithm to establish a common characterization model 311 of the underwater throttle valve performance degradation process, using a dynamic Bayesian network to describe the evolution of the state space model over time, and establishing a dynamic evolution model 312 of the underwater throttle valve performance degradation;

[0105] S3.3, establishing a dynamic evolution model of an underwater throttle valve based on a dynamic Bayesian neural network, that is, an underwater throttle valve performance degradation dynamic evolution model 312 established based on a DBN model;

[0106] S3.4, using the particle filter PF algorithm to track the degradation state and predict the remaining life of the underwater throttle valve, thereby completing the establishment of the underwater throttle valve performance degradation digital twin model 321;

[0107] S3.5. Establish a digital twin model for underwater throttle valve performance degradation and remaining life prediction based on particle filter PF approximate inference algorithm 322: The basic idea of ​​PF algorithm is to extract a series of weighted particles to realize the posterior evaluation of the quantity to be estimated. In this process, the weight of each particle is iteratively updated by continuously comparing the difference between the estimated value and the observed value, and finally approximate the true posterior distribution;

[0108] S3.6. Use reinforcement learning algorithms to compare perception data with simulation data, and adjust parameters to maintain consistency between the virtual and the real.

[0109] Among them, in S3.3, the DBN model is a deep learning_DBN model or a deep belief network (DBN).

[0110] Furthermore, in S3.1, the specific method of establishing the underwater throttle valve erosion state mathematical model 31 using the nonlinear regression algorithm includes the following steps:

[0111] S3.1.1. Obtain multiple sets of experimental data through underwater throttle valve erosion experiments, the experimental data being target metal resistance values ​​and erosion amounts corresponding to the target metal resistance values;

[0112] S3.1.2, correct the experimental data obtained in step S3-2-1 to eliminate systematic errors;

[0113] S3.1.3. Based on the data corrected in step S3-2-2, a fitting function model is constructed using a nonlinear regression algorithm; an interpolation method is used to solve the model and obtain a fitting function curve.

[0114] In this embodiment, in S3.2, the specific method of establishing the commonality representation model 311 is:

[0115] Since the external causes of underwater throttle valve corrosion are mostly factors such as seawater pressure, flow rate, temperature, carbon dioxide partial pressure and underwater disturbance in seawater, an external impact model of underwater throttle valve in marine environment is established, and the external impact process of underwater throttle valve in marine environment is modeled as a Poisson process. 1, , t 2 ≥0, there are:

[0116]

[0117] Where n is the number of times the underwater throttle valve is impacted by the marine environment, N c (t 1 +t 2 ) is t 1 +t 2 The number of external impacts of the marine environment in a time period, N c (t 1 ) is t1 The number of external impacts of the marine environment in a period of time, P{N c (t 1 +t 2 )-N c (t 1 )=n} is the 2 The probability of n external shocks to the marine environment occurring in a time period, λ is the parameter of the Poisson distribution, which is determined by historical data;

[0118] The inner wall of the underwater throttle valve is eroded, corroded and accelerated by the flow of oil and water mixed with sand, which causes the wall structure of the underwater throttle valve to become thinner. Under the action of erosion, the geometric shape of the wall of the underwater throttle valve will change with the increase of the erosion exposure of solid particles. Therefore, the change of surface contour will change the flow field, thereby redirecting the particles to another erosion "hot spot".

[0119] In addition, the local impact angle of the particles will change with time, which may lead to a large change in the local erosion rate. Therefore, the amplitude of the thinning of the underwater throttle valve body wall caused by erosion can be used as an indicator of the performance degradation of the underwater throttle valve based on digital twins.

[0120] In this embodiment, in S3.3, a dynamic evolution model 312 of underwater throttle valve performance degradation is established based on the DBN model.

[0121] Specifically, the model uses observation nodes, random nodes and functional nodes to comprehensively express the observed quantities, uncertainty factors, degradation models, etc. in the performance degradation of underwater throttle valves;

[0122] The underwater throttle valve performance degradation dynamic evolution model 312 contains two important parameters A and B, which are determined by material properties. During the operation of the underwater throttle valve, parameters A and B change due to factors such as the working environment and damage state. Random nodes are used to describe parameters A and B. In addition, the dotted line connection between different time steps represents the posterior information of the previous moment as the prior information of the next moment. The degradation state X of the underwater throttle valve is calculated through two functional nodes. In this process, the degradation state is considered to be affected by the uncertainty of the model and the measurement system. The degradation state is also expressed in the form of random nodes, and the posterior distribution of the degradation state in the previous time step is used as the prior distribution of the next time step. With the continuous collection of observation data Y, the model parameters A and B are first updated and corrected, and then the degradation state of the turbine disk is evaluated and dynamically evolved along the time step, such as Figure 2 shown.

[0123] Furthermore, in S3.3, the purpose of establishing a DBN-based dynamic evolution model is to: k =(A k ,Bk ) and a set of underwater throttle valve observation data Q 0:k =(q 0 ,q 1 ,…,q k ) is used to estimate the posterior distribution of the degradation state of the underwater throttle valve. The calculation formula is as follows:

[0124]

[0125] in, represents the i-th particle; δ(·) is the Dirac function; represents the weight corresponding to the i-th particle, which can be calculated by the following formula:

[0126]

[0127]

[0128] In order to overcome the particle degradation problem caused by too small weights of some particles, a resampling strategy is used to reset the weights of each particle:

[0129]

[0130] is equivalent to:

[0131]

[0132] In the formula, represents the particle after resampling at time k; n i Represents particles After resampling, a new particle set is generated The number of times the number of copies is made;

[0133] Further predict the degradation state at the future time step k+t (t=1, 2, ...)

[0134]

[0135] in, Available through We get: where V is the uncertainty of the model;

[0136] The RUL of the underwater throttle valve can be defined as the time interval between the moment when the predicted future degradation state first reaches the preset failure threshold and the current moment under the conditions of the current degradation state and model parameters. The RUL prediction of the turbine disk can be achieved by extending the degradation trajectory to reach the preset failure threshold. The calculation formula is as follows:

[0137]

[0138] is the RUL prediction value of the ith particle; t is the time step, ε is the failure threshold;

[0139] Finally, the predicted value of RUL can be obtained by weighted averaging:

[0140]

[0141] In this embodiment, Figure 3 As shown, in S3.6, the specific method of using the reinforcement learning algorithm to compare the perception data with the simulation data and adjusting the parameters to achieve the consistency of virtuality and reality includes the following steps:

[0142] S3.6.1. By compiling the physical property model interface, the injection of perception data and the output of data sets are realized to build a two-way interactive mechanism, and the production data of the underwater throttle valve, such as throttle valve material, throttle valve opening, fluid mass flow rate, fluid density, solid particle mass flow rate and other parameters, are used as the input of the model;

[0143] S3.6.2. Initialize the digital twin model, use the reinforcement learning algorithm in the correction space, compare the perception data with the simulation data, and maintain consistency through parameter adjustment;

[0144] S3.6.3. Initialize the digital twin model, manually select the correction parameters and target variables of the digital twin model in the correction space, conduct multiple groups of experiments, obtain the sensitivity relationship between each correction parameter and the target variable, and then screen the correction parameters to determine the parameters with higher sensitivity as the correction parameters;

[0145] S3.6.4. With the correction parameter as the independent variable of the test and the target variable as the dependent variable of the test, based on the idea of ​​experimental design, multiple regression or difference fitting is performed on the correction parameter and the target variable to construct a response surface model. The response surface model is a mathematical model that expresses the relationship between the independent variable and the dependent variable in the form of a mathematical function, thereby constructing the objective function of the target variable.

[0146] S3.6.5. Find the correction parameters that can meet the objective function in the response surface space, replace the original twin model parameters with the corrected parameters, realize the mapping of the twin model, establish the fault prediction and erosion degradation model of the underwater throttle valve, improve the accuracy of the health index and remaining life distribution prediction, and finally realize the establishment and consistency maintenance of the two-way interactive twin model of the underwater production system.

[0147] In this embodiment, in S4, the operation state and erosion state of the underwater throttle valve are simulated and calculated by the underwater throttle valve digital twin model 4, and the simulation calculation includes but is not limited to the erosion rate calculation, wherein the erosion rate algorithm formula is as follows:

[0148]

[0149] Where: R erosion is the erosion rate, (kg / (m 2 ·s)); N particle is the number of solid particles colliding with the wall, dimensionless; is the mass flow rate of solid particles, kg / s; f(α) is a function of the solid particle collision angle α; v is the velocity of the solid particles relative to the wall, m / s; b(v) is a function of the relative velocity of the solid particles; C(d p ) is a function of the solid particle size, d p is the diameter of the particle, m; A face is the unit area of ​​the impact wall;

[0150] And then:

[0151] C(d p )=1.599HB -0.59 F s ×10 -7 ;

[0152]

[0153] like Figure 4-Figure 5 As shown, this embodiment also provides a system for estimating the remaining service life of an underwater throttle valve based on digital twins, which is used to implement the operation process of the above-mentioned method for estimating the remaining service life of an underwater throttle valve based on digital twins, including a data acquisition module 1, a database 2 and a numerical simulation unit 3; the signal output end of the data acquisition module 1 is connected to the signal input end of the database 2, and the database 2 is signal-connected to the numerical simulation unit 3; wherein:

[0154] The data acquisition module 1 is used to acquire data on multiple sets of valve body metal material resistance values ​​and corresponding erosion values ​​of valve body metal resistance values ​​obtained through underwater throttle valve erosion experiments, data on the complex and diverse underwater throttle valve erosion environment, and underwater throttle valve direct inspection process data, and can store such data in the database 2;

[0155] The database 2 is used to classify and store various types of data collected by the data acquisition module 1, various types of simulation data obtained by the numerical simulation unit 3, and user presets;

[0156] The numerical simulation unit 3 is used to construct the relevant mathematical model and digital twin model of the underwater throttle valve based on the real data of the underwater throttle valve obtained by measurement and the experimental data obtained through the underwater throttle valve erosion experiment, using certain algorithms and models, and to estimate the remaining service life of the underwater throttle valve through numerical simulation and deep learning.

[0157] In this embodiment, the database 2 includes a physical model database 21, a twin data unit 22 and an evaluation data unit 23; the physical model database 21, the twin data unit 22 and the evaluation data unit 23 run in parallel and are sequentially connected in communication; wherein:

[0158] The physical model database 21 is used to store and manage the complex and diverse data of the underwater throttle valve erosion environment including environmental data and erosion measurement data obtained by the data acquisition module 1;

[0159] The twin data unit 22 is used to store and manage the operation state data and erosion state data of the underwater throttle valve obtained by simulation calculation of the numerical simulation unit 3;

[0160] The evaluation data unit 23 is used to save and manage the preset remaining life threshold data, the underwater throttle valve direct inspection process data collected by the data acquisition module 1, and the remaining life prediction result data obtained by predicting the remaining life of the underwater throttle valve using the underwater throttle valve remaining life prediction algorithm through the numerical simulation unit 3. By comparing the remaining life calculation result with the preset remaining life threshold data, the evaluation and prediction of the erosion condition of the underwater throttle valve foundation can be realized.

[0161] In this embodiment, the numerical simulation unit 3 includes an underwater throttle valve erosion state mathematical model 31 and an underwater throttle valve digital twin model 32; the underwater throttle valve erosion state mathematical model 31 and the underwater throttle valve digital twin model 32 run in parallel and are signal connected; wherein:

[0162] The underwater throttle valve erosion state mathematical model 31 is used to mathematically simulate the physical state of the underwater throttle valve based on experimental data;

[0163] The underwater throttle valve erosion state mathematical model 31 includes a common characterization model 311 and an underwater throttle valve performance degradation dynamic evolution model 312; the common characterization model 311 is used to simulate the common characterization state of the underwater throttle valve in combination with external causes that may cause corrosion of the underwater throttle valve body, so as to determine the index value that can be used as the underwater throttle valve performance degradation; the underwater throttle valve performance degradation dynamic evolution model 312 is used to describe the variation law of degradation state and parameters over time and the transmission process of model uncertainty;

[0164] The underwater throttle valve digital twin model 32 is constructed by substituting the real-time data of the measured environmental data and erosion measurement data into the underwater throttle valve erosion state mathematical model 31, and is used to simulate and calculate the operating state and erosion state of the underwater throttle valve;

[0165] The underwater throttle valve digital twin model 32 includes an underwater throttle valve performance degradation digital twin model 321 and an underwater throttle valve performance degradation and remaining life prediction digital twin model 322; the underwater throttle valve performance degradation digital twin model 321 is used to simulate the performance degradation state of the underwater throttle valve; the underwater throttle valve performance degradation and remaining life prediction digital twin model 322 is used to track the degradation state and predict the remaining life of the underwater throttle valve using a particle filter PF algorithm.

[0166] like Figure 5 As shown, this embodiment also provides a system operation platform device, which includes a processor, a memory, and a computer program stored in the memory and running on the processor.

[0167] The processor includes one or more processing cores. The processor is connected to the memory through a bus. The memory is used to store program instructions. When the processor executes the program instructions in the memory, the steps of the above-mentioned underwater throttle valve remaining service life estimation system and method based on digital twin are implemented.

[0168] Optionally, the memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0169] In addition, the present invention also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the steps of the above-mentioned digital twin-based underwater throttle valve remaining service life estimation system and method.

[0170] Optionally, the present invention also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the steps of the above-mentioned system and method for estimating the remaining useful life of an underwater throttle valve based on digital twins.

[0171] A person of ordinary skill in the art can understand that the process of implementing all or part of the steps of the above-mentioned embodiments can be completed by hardware, or can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0172] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. Remaining service life estimation method of underwater throttle valve based on digital twin, Features: The steps include: S1. Obtain multiple sets of valve body metal material resistance values ​​and corresponding erosion amounts of valve body metal resistance values ​​through underwater throttle valve erosion experiments. Use a nonlinear regression algorithm based on the data obtained from the metal corrosion experiments to establish a mathematical model of underwater throttle valve erosion state (3). S2, measuring the complex and diverse data of the underwater throttle valve erosion environment through the data acquisition module (1) and transmitting it to the database (2) of the mobile workstation for storage and management, realizing a real-time data-driven finite element simulation model, and constructing a physical model database (21); S3, substituting the measured environmental data and the real-time data of the erosion measurement data into the underwater throttle valve erosion state mathematical model (31), and constructing the underwater throttle valve digital twin model (32); S4, using the underwater throttle valve digital twin model (32) to simulate and calculate the operating state and erosion state of the underwater throttle valve, and the operating state data and erosion state data of the underwater throttle valve obtained by the simulation calculation are stored in the twin data unit (22) of the database (2); S5. Establishing a dynamic evolution model of underwater throttle valve performance degradation (312) through a dynamic Bayesian model, and implementing it through a particle filter PF algorithm, thereby establishing a complete underwater throttle valve performance degradation digital twin model (321); S6, collecting the underwater throttle valve direct inspection process data through the data collection module (1) and storing it in the evaluation data unit (23) of the database (2); using the underwater throttle valve remaining life prediction algorithm to predict the underwater throttle valve remaining life according to the data in the evaluation data unit (23), and storing the remaining life prediction results obtained in the evaluation data unit (23); S7, calling the remaining life threshold data of the underwater throttle valve preset in the database (2), and then comparing the remaining life calculation result with the remaining life threshold. If the remaining life calculation result is lower than 10%, 50%, or 80% of the remaining life threshold, the erosion degree of the underwater throttle valve is divided into three levels: A-mild erosion, B-moderate erosion, and C-severe erosion respectively; S8. Output the erosion level of the underwater throttle valve in time, complete the evaluation and prediction of the erosion condition of the underwater throttle valve foundation; find the corrosion area according to the positioning of the monitoring point, and complete the maintenance or repair of the area according to the evaluation results.

2. According to the method for estimating the remaining service life of an underwater throttle valve based on digital twins according to claim 1, Features: In S3, the real-time data of the measured environmental data and erosion measurement data are substituted into the underwater throttle valve erosion state mathematical model (31), and the specific steps of constructing the underwater throttle valve digital twin model (32) are as follows: S3.

1. Using nonlinear regression algorithm to establish a mathematical model of underwater throttle valve erosion state (31); S3.2, using a nonlinear regression algorithm to establish a common characterization model (311) of the underwater throttle valve performance degradation process, using a dynamic Bayesian network to describe the evolution of the state space model over time, and establishing a dynamic evolution model (312) of the underwater throttle valve performance degradation; S3.3, establishing a dynamic evolution model of an underwater throttle valve based on a dynamic Bayesian neural network, that is, a dynamic evolution model of underwater throttle valve performance degradation established based on a DBN model (312); S3.4, using the particle filter PF algorithm to track the degradation state of the underwater throttle valve and predict the remaining life, thereby completing the establishment of the underwater throttle valve performance degradation digital twin model (321); S3.

5. Establish a digital twin model for underwater throttle valve performance degradation and remaining life prediction based on particle filter PF approximate reasoning algorithm (322); S3.

6. Use reinforcement learning algorithms to compare perception data with simulation data, and adjust parameters to maintain consistency between the virtual and the real.

3. According to the method for estimating the remaining service life of an underwater throttle valve based on digital twins according to claim 2, Features: In S3.1, the specific method of establishing the underwater throttle valve erosion state mathematical model (31) by using the nonlinear regression algorithm comprises the following steps: S3.1.

1. Obtain multiple sets of experimental data through underwater throttle valve erosion experiments, the experimental data being target metal resistance values ​​and erosion amounts corresponding to the target metal resistance values; S3.1.2, correct the experimental data obtained in step S3-2-1 to eliminate systematic errors; S3.1.

3. Based on the data corrected in step S3-2-2, a fitting function model is constructed using a nonlinear regression algorithm; an interpolation method is used to solve the model and obtain a fitting function curve.

4. According to claim 2, the method for estimating the remaining service life of an underwater throttle valve based on digital twins, Features: In S3.2, the specific method of establishing the commonality representation model (311) is: Since the external causes of underwater throttle valve corrosion are mostly factors such as seawater pressure, flow rate, temperature, carbon dioxide partial pressure and underwater disturbance in seawater, an external impact model of underwater throttle valve in marine environment is established, and the external impact process of underwater throttle valve in marine environment is modeled as a Poisson process. 1, , t 2 ≥0, there are: Where n is the number of times the underwater throttle valve is impacted by the marine environment, N c (t 1 +t 2 ) is t 1 +t 2 The number of external impacts of the marine environment in a time period, N c (t 1 ) is t 1 The number of external impacts of the marine environment in a period of time, P{N c (t 1 +t 2 )-N c (t 1 )=n} is the 2 The probability of n external shocks to the marine environment occurring in a time period, λ is the parameter of the Poisson distribution, which is determined by historical data.

5. According to claim 2, the method for estimating the remaining service life of an underwater throttle valve based on digital twins, Features: In S3.3, the purpose of establishing a DBN-based dynamic evolution model is to: k =(A k ,B k ) and a set of underwater throttle valve observation data Q 0:k =(q 0 ,q 1 ,…,q k ) is used to estimate the posterior distribution of the degradation state of the underwater throttle valve. The calculation formula is as follows: in, represents the i-th particle; δ(·) is the Dirac function; represents the weight corresponding to the i-th particle, which can be calculated by the following formula: In order to overcome the particle degradation problem caused by too small weights of some particles, a resampling strategy is used to reset the weights of each particle: is equivalent to: In the formula, represents the particle after resampling at time k; n i Represents particles After resampling, a new particle set is generated The number of times the number of copies is made; Further predict the degradation state at the future time step k+t (t=1, 2, ...) in, Available through We get: where V is the uncertainty of the model; The RUL of the underwater throttle valve can be defined as the time interval between the moment when the predicted future degradation state first reaches the preset failure threshold and the current moment under the conditions of the current degradation state and model parameters. The RUL prediction of the turbine disk can be achieved by extending the degradation trajectory to reach the preset failure threshold. The calculation formula is as follows: is the RUL prediction value of the ith particle; t is the time step, ε is the failure threshold; Finally, the predicted value of RUL can be obtained by weighted averaging:

6. The method for estimating the remaining service life of an underwater throttle valve based on digital twin according to claim 2, Features: In S3.6, the specific method of using a reinforcement learning algorithm to compare the perception data with the simulation data and adjusting parameters to achieve virtual-real consistency includes the following steps: S3.6.

1. By compiling the physical property model interface, the injection of perception data and the output of data sets are realized to build a two-way interaction mechanism, and the production data of the underwater throttle valve is used as the input of the model; S3.6.

2. Initialize the digital twin model, use the reinforcement learning algorithm in the correction space, compare the perception data with the simulation data, and maintain consistency through parameter adjustment; S3.6.

3. Initialize the digital twin model, manually select the correction parameters and target variables of the digital twin model in the correction space, conduct multiple groups of experiments, obtain the sensitivity relationship between each correction parameter and the target variable, and then screen the correction parameters to determine the parameters with higher sensitivity as the correction parameters; S3.6.

4. Take the correction parameter as the independent variable of the test and the target variable as the dependent variable of the test. Based on the idea of ​​experimental design, perform multiple regression or difference fitting on the correction parameter and the target variable, construct a response surface model, and thus construct the objective function of the target variable. S3.6.

5. Find the correction parameters that can satisfy the objective function in the response surface space, replace the original twin model parameters with the corrected parameters, realize the mapping of the twin model, establish the fault prediction and erosion degradation model of the underwater throttle valve, and finally realize the establishment and consistency maintenance of the two-way interactive twin model of the underwater production system.

7. The method for estimating the remaining service life of an underwater throttle valve based on digital twin according to claim 1, Features: In S4, the operation state and erosion state of the underwater throttle valve are simulated and calculated by the underwater throttle valve digital twin model (4), and the simulation calculation includes but is not limited to the erosion rate calculation, wherein the erosion rate algorithm formula is as follows: Where: R erosion is the erosion rate, (kg / (m 2 ·s)); N particle is the number of solid particles colliding with the wall, dimensionless; is the mass flow rate of solid particles, kg / s; f(α) is a function of the solid particle collision angle α; v is the velocity of the solid particles relative to the wall, m / s; b(v) is a function of the relative velocity of the solid particles; C(d p ) is a function of the solid particle size, d p is the diameter of the particle, m; A face is the unit area of ​​the impact wall; And then: C(d p )=1.599HB -0.59 F s ×10 -7 ; 8. A system for estimating the remaining useful life of an underwater throttle valve based on digital twins, the system being used to implement the operation process of the method for estimating the remaining useful life of an underwater throttle valve based on digital twins as described in any one of claims 1 to 7, Features: The invention comprises a data acquisition module (1), a database (2) and a numerical simulation unit (3); the signal output end of the data acquisition module (1) is connected to the signal input end of the database (2), and the database (2) is signal-connected to the numerical simulation unit (3); wherein: The data acquisition module (1) is used to acquire data on multiple groups of valve body metal material resistance values ​​and valve body metal resistance values ​​corresponding to erosion amounts, data on the complex and diverse underwater throttle valve erosion environment, underwater throttle valve direct inspection process data, etc. obtained through underwater throttle valve erosion experiments, and can store the data in the database (2); The database (2) is used to classify and store various types of data collected and acquired by the data acquisition module (1), various types of simulation data acquired by the numerical simulation unit (3), and user presets; The numerical simulation unit (3) is used to construct a relevant mathematical model and a digital twin model of the underwater throttle valve based on the real data of the underwater throttle valve obtained by measurement and the experimental data obtained through the underwater throttle valve erosion experiment, using a certain algorithm and model, and to estimate the remaining service life of the underwater throttle valve through numerical simulation and deep learning.

9. The underwater throttle valve remaining service life estimation system based on digital twin according to claim 8, Features: The database (2) comprises a physical model database (21), a twin data unit (22) and an evaluation data unit (23); the physical model database (21), the twin data unit (22) and the evaluation data unit (23) are operated in parallel and are sequentially connected in communication; wherein: The physical model database (21) is used to store and manage complex and diverse data of underwater throttle valve erosion environment, including environmental data and erosion measurement data, obtained by the data acquisition module (1); The twin data unit (22) is used to store and manage the operation state data and erosion state data of the underwater throttle valve obtained by simulation calculation of the numerical simulation unit (3); The evaluation data unit (23) is used to store and manage preset remaining life threshold data, underwater throttle valve direct inspection process data collected by the data collection module (1), and remaining life prediction result data obtained by predicting the remaining life of the underwater throttle valve using an underwater throttle valve remaining life prediction algorithm through the numerical simulation unit (3), and to evaluate and predict the erosion condition of the underwater throttle valve foundation by comparing the remaining life calculation result with the preset remaining life threshold data.

10. The underwater throttle valve remaining service life estimation system based on digital twin according to claim 8, Features: The numerical simulation unit (3) comprises an underwater throttle valve erosion state mathematical model (31) and an underwater throttle valve digital twin model (32); the underwater throttle valve erosion state mathematical model (31) and the underwater throttle valve digital twin model (32) run in parallel and are signal-connected; wherein: The underwater throttle valve erosion state mathematical model (31) is used to perform mathematical simulation on the physical state of the underwater throttle valve based on experimental data; The underwater throttle valve erosion state mathematical model (31) includes a common characterization model (311) and an underwater throttle valve performance degradation dynamic evolution model (312); the common characterization model (311) is used to simulate the common characterization state of the underwater throttle valve in combination with external causes that may cause corrosion of the underwater throttle valve body, so as to determine an indicator value that can be used as an underwater throttle valve performance degradation; the underwater throttle valve performance degradation dynamic evolution model (312) is used to describe the variation law of degradation state and parameters over time and the transmission process of model uncertainty; The underwater throttle valve digital twin model (32) is constructed by substituting the real-time data of the measured environmental data and erosion measurement data into the underwater throttle valve erosion state mathematical model (31), and is used to simulate and calculate the operating state and erosion state of the underwater throttle valve; The underwater throttle valve digital twin model (32) comprises an underwater throttle valve performance degradation digital twin model (321) and an underwater throttle valve performance degradation and remaining life prediction digital twin model (322); the underwater throttle valve performance degradation digital twin model (321) is used to simulate and calculate the performance degradation state of the underwater throttle valve; the underwater throttle valve performance degradation and remaining life prediction digital twin model (322) is used to track the degradation state and predict the remaining life of the underwater throttle valve using a particle filter PF algorithm.

Images