Digital twin modeling and fault prediction method for large hydraulic cylinder

Abstract

The present application belongs to the technical field of digital twinning, and particularly relates to a digital twinning modeling and fault prediction method for a large hydraulic cylinder. The method comprises the following steps: S1: a parameterized geometric model of a hydraulic cylinder assembly is established by using a three-dimensional modeling software. S2: a conversion plug-in is used to import the hydraulic cylinder mechanical system into Simulink, and an equivalent multi-body dynamics model of the hydraulic cylinder is established. S3: a hydraulic subsystem is created in Simulink; the equivalent multi-body dynamics model of the hydraulic cylinder and the hydraulic subsystem are fused to obtain a digital twinning model of the hydraulic cylinder. S4: a digital twinning virtual-real interaction system including the digital twinning model of the hydraulic cylinder is constructed. S5: the digital twinning model of the hydraulic cylinder is tested for accuracy and optimized in the digital twinning virtual-real interaction system. S6: state monitoring is performed by using the digital twinning virtual-real interaction system, and a fault prediction model is trained by taking the monitoring data as samples. The present application solves the problem that the operating state and component wear of the hydraulic cylinder are difficult to detect directly.

Description

Technical Field

[0001] This invention belongs to the field of digital twin technology, specifically relating to a digital twin modeling and fault prediction method for a large hydraulic cylinder, a digital twin virtual-real interaction system, and a hydraulic cylinder visualization monitoring device based on the digital twin system. Background Technology

[0002] A hydraulic cylinder is a mechanical actuator, part of a hydraulic system, and is widely used in various industries such as construction, manufacturing, aerospace, and marine oil and gas. It can achieve linear motion, lifting operations, and more. The normal operation of the hydraulic system is crucial to the successful implementation of numerous engineering applications; therefore, it is necessary to provide technical means to monitor the operating status of hydraulic cylinders and other mechanisms, as well as the wear and tear of their components.

[0003] Existing methods for monitoring hydraulic systems primarily involve embedding numerous sensors within the system and analyzing the internal state of the hydraulic cylinders based on the sensor data. Vibration, pressure, and acoustic emission signals are the main characteristic parameters used to monitor the operation of the hydraulic system. However, state analysis based on signal characteristics inherently suffers from latency. Furthermore, monitoring models obtained through data-driven approaches are highly dependent on the quantity and quality of the data, thus exhibiting certain limitations.

[0004] Digital twin technology is an emerging technology that can create a virtual model of a target entity in a computer and use this virtual model to perform a high degree of simulation of the target entity. Currently, digital twin technology has been applied to the design, construction, assembly, and testing of large-scale engineering projects or electrical systems, helping technicians to better complete projects. Building on this, how to utilize digital twin technology to solve the dilemma of the difficulty in directly monitoring hydraulic systems has become a pressing technical challenge for those skilled in the art. Summary of the Invention

[0005] To address the challenges of directly detecting component wear and tear in hydraulic cylinders, which results in high detection costs, insufficient accuracy, and inadequate real-time performance, this invention provides a digital twin modeling and fault prediction method for large hydraulic cylinders, along with corresponding systems and equipment.

[0006] This invention is achieved using the following technical solution:

[0007] A digital twin modeling and fault prediction method for large hydraulic cylinders includes the following steps:

[0008] S1: Based on the physical characteristics, material parameters and actual working conditions of the hydraulic cylinder, a parametric geometric model of the hydraulic cylinder assembly is established using 3D modeling software, and a mechanical model of the hydraulic cylinder with a digital twin virtual-real interaction interface is generated.

[0009] S2: Use the conversion plugin to import the hydraulic cylinder mechanical system into Simulink and establish an equivalent multibody dynamics model of the hydraulic cylinder.

[0010] S3: Create a hydraulic subsystem in Simulink; merge the equivalent multibody dynamics model of the hydraulic cylinder and the hydraulic subsystem to realize the physical connection and drive control between the components, and obtain the required digital twin model of the hydraulic cylinder.

[0011] S4: Construct a digital twin virtual-real interaction system that includes a digital twin model of the hydraulic cylinder, a hardware system, and a control system. The digital twin virtual-real interaction system is driven by the acquired drive signals of the hydraulic cylinder entity, and performs real-time simulation of the hydraulic cylinder entity's operating state in the digital twin model.

[0012] S5: Set a preset threshold for simulation accuracy, and perform accuracy testing and optimization on the digital twin model of the hydraulic cylinder created in the digital twin virtual-real interaction system.

[0013] (1) When the simulation accuracy reaches the threshold, the corresponding model parameters are saved;

[0014] (2) When the simulation accuracy is lower than the threshold, return to step S3 to correct the parameters of the created hydraulic subsystem.

[0015] S6: Utilizing a digital twin virtual-real interaction system that meets accuracy requirements, the operating status of the hydraulic system is visualized and monitored. Using the real-time acquired monitoring data as samples, a fault prediction model based on machine learning algorithms is trained to achieve fault prediction of the hydraulic system.

[0016] As a further improvement to the present invention, the detailed steps of creating the hydraulic cylinder mechanical model in step S1 are as follows:

[0017] S11: The structure of the hydraulic cylinder assembly is simplified to include the following parts: sealing ring, guide sleeve, piston rod, cylinder barrel, piston sealing ring, piston, and cylinder bottom.

[0018] S12: Define the shape, size, structure and constraint relationships of the parts in 3D modeling software, and perform part modeling and assembly to obtain a parametric model of the assembly.

[0019] As a further improvement to the present invention, the process of creating the equivalent multibody dynamics model of the hydraulic cylinder is as follows:

[0020] S21: The parameterized model of the hydraulic cylinder assembly is converted into an XML file using the Simscape Multibody Link model conversion plugin. The file stores the following information and constraints: cylinder inner diameter, piston rod diameter, piston stroke, piston area, and seal area.

[0021] S22: Identify and transform the model using MATLAB / Simulink commands, converting the XML file into an SLX program file. Use a multi-domain modeling language to describe the constructed hydraulic cylinder assembly model, and correct information such as the coordinate system and connection relationships of the model.

[0022] S23: Use the Prismatic Joint module and Translational Multibody Interface module in Matlab / Simulink software to realize signal conversion between the hydraulic machinery model and the hydraulic subsystem.

[0023] As a further improvement of the present invention, in step S21, the parameterized model of the assembly provides a parameter adjustment interface to realize the adjustment of the hydraulic cylinder structure.

[0024] In step S22, the XML file provides transformation nodes that can be recognized by the multi-domain modeling language for the established hydraulic cylinder assembly model.

[0025] In step S23, the interface information transmitted between the hydraulic mechanical model and the hydraulic subsystem includes position, velocity, and force information.

[0026] As a further improvement to the present invention, the detailed steps of creating the digital twin model of the hydraulic cylinder in step S3 are as follows:

[0027] S31: Create a hydraulic subsystem in Simulink that includes a hydraulic pump, angular velocity source, relief valve, first proportional valve, second proportional valve, double-acting hydraulic cylinder, oil tank, pressure sensor, and flow sensor.

[0028] S32: The hydraulic subsystem is used as the driving part of the equivalent multibody dynamics model of the hydraulic cylinder, thus forming the hydraulic system.

[0029] S34: Define and configure the control parameters of the hydraulic system; the hydraulic system generates a corresponding driving force through the hydraulic cylinder according to the input control parameters, realizing the kinematic simulation of the digital twin model of the hydraulic cylinder.

[0030] The control parameters include the first proportional valve control signal, the second proportional valve control signal, and the hydraulic pump control signal.

[0031] As a further improvement of the present invention, the control signal for the first proportional valve is a voltage control signal, which is electrically connected to the first proportional valve to control the opening degree of the first proportional valve. The working port of the first proportional valve is connected to the rodless chamber of the double-acting hydraulic cylinder, the inlet port of the first proportional valve is connected to the outlet of the hydraulic pump, and the return port of the first proportional valve is connected to the oil tank.

[0032] The control signal for the second proportional valve is a voltage control signal, which is connected to the second proportional valve to control its opening degree. The working port of the second proportional valve is connected to the rod chamber of the double-acting hydraulic cylinder, the inlet port of the second proportional valve is connected to the outlet of the hydraulic pump, and the return port of the second proportional valve is connected to the oil tank.

[0033] The hydraulic pump control signal is used to control the speed of the servo motor connected to the hydraulic pump, and its signal output interface is connected to the Translational Multibody Interface (TMI) conversion module.

[0034] As a further improvement of the present invention, in step S4, the hardware system includes a hydraulic cylinder body, a pressure sensor, a displacement sensor, a data acquisition card, and a communication cable. The hydraulic cylinder body includes a cylinder, a servo motor, hydraulic lines, and a drive assembly. The pressure sensor and the displacement sensor are mounted on the hydraulic cylinder body; the former is used to acquire the rod chamber pressure and rodless chamber pressure of the hydraulic cylinder body, and the latter is used to acquire the piston displacement and movement speed. The data acquisition card is used to acquire the detection data from the pressure sensor and the displacement sensor, and communicates with a host computer containing the control system via the communication cable.

[0035] The control system employs Simulink Real-Time to execute the control of the hydraulic system. It then acquires digital or analog control signals from the hydraulic cylinder entity via a signal acquisition terminal, and controls the simulation operation of the hydraulic cylinder's digital twin model. The digital signals originate from signal sampling of the drive components during the cylinder's downward and return strokes; the analog signals are derived from the detection results of displacement and pressure sensors. In the hydraulic cylinder's digital twin model, the digital signals are used to control the on / off switching of the solenoid valve, while the analog signals are used to adjust the valve opening of the proportional valve and the speed of the servo motor.

[0036] As a further improvement of this invention, in step S5, the digital twin model of the hydraulic cylinder operates synchronously with the physical hydraulic cylinder. The displacement d and velocity v of the piston, as well as the pressure p1 in the rod chamber and p2 in the rodless chamber, are simultaneously sampled during their operation to form a test vector A: A = {d, v, p1, p2}. The fitting degree R is then calculated using the following formula. 2 To evaluate the simulation accuracy of the digital twin model of the hydraulic cylinder:

[0037]

[0038] In the above formula: i represents the sampling time of different data points; n is the sample size in the test data; y i Y is the test vector composed of the operating parameters of the hydraulic cylinder at time i; i The test vector is composed of the operating parameters of the digital twin model of the hydraulic cylinder at time i; This represents a vector composed of the average values ​​of various operating data of the digital twin model of the hydraulic cylinder.

[0039] As a further improvement of the present invention, in step S6, the visualization monitoring process based on the digital twin model of the hydraulic cylinder is as follows:

[0040] (1) Extract feature data that characterizes the operating state of the digital twin model, and reduce the dimensionality by linearly combining the high-dimensional feature data through the transformation of the feature data.

[0041] (2) Adjustable throttle orifices are connected to the inlet and outlet of the hydraulic cylinder in the twin model. Changing the size of the throttle orifice causes changes in pressure and flow rate, simulating a pipeline blockage fault module. By adjusting the parameters of the sealing ring, including the elastic modulus and friction coefficient, a sealing ring wear fault module is set. A sliding resistance module is added to simulate a piston wear fault module that hinders piston movement. The inlet and outlet oil lines are connected through throttle orifices, and the size of the throttle orifice is adjusted to simulate an internal leakage fault at the piston of the hydraulic cylinder. The twin simulation model is run, and the corresponding fault data is imported into the MATLAB workspace. Principal component analysis is used to extract the main features reflecting the fault type, and the correlation between the main feature parameters and the fault type is established.

[0042] (3) Using the main features of wear-related parameter changes as samples, the simulation results of the digital twin model of the hydraulic cylinder are used to extract training fault data through FFT. A machine learning algorithm is designed and trained to predict typical faults, including seal wear, pipeline blockage, piston wear, and internal leakage. Fault prediction is performed using the trained machine learning algorithm.

[0043] (4) Based on the sampling data of the operation process of the digital twin model of the hydraulic cylinder, establish the curve of the main feature of the hydraulic system entity output changing with time point, and display the curve of the fault parameter changing with time point in the fault parameter prediction coordinate area; establish the state space model based on the changing trend of the parameter curve, estimate the system state at future time points and realize the state monitoring function.

[0044] (5) Use any visualization method to visualize the process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system.

[0045] This invention also includes a digital twin virtual-real interaction system, which is created using the aforementioned digital twin modeling and fault prediction method for large hydraulic cylinders. The digital twin virtual-real interaction system is used to synchronously and realistically simulate the operating state of the hydraulic cylinder entity, thereby assisting in the state monitoring and fault prediction of the hydraulic cylinder entity. The digital twin virtual-real interaction system includes: a digital twin model of the hydraulic cylinder, a signal acquisition module, a control system, and a human-machine interaction module.

[0046] A digital twin model of a hydraulic cylinder, serving as a virtual model of the hydraulic cylinder entity, can be used to simulate the operating state of the hydraulic cylinder entity in real time.

[0047] The signal acquisition module is used to acquire the operating status parameters of the hydraulic cylinder and upload them to the host computer. The signal acquisition module includes a signal acquisition unit, a pressure sensor, a displacement sensor, a data acquisition card, and a communication cable. The signal acquisition unit acquires the electrical signals transmitted by the drive components of the hydraulic cylinder. The pressure sensor and displacement sensor are mounted on the hydraulic cylinder; the former acquires the rod chamber pressure and rodless chamber pressure, while the latter acquires the piston displacement and velocity. The data acquisition card acquires the detection data from the signal acquisition unit, pressure sensor, and displacement sensor. The data acquisition card communicates with the host computer via the communication cable.

[0048] The control system is used to synchronously drive the digital twin model of the hydraulic cylinder based on electrical signals collected from the drive components of the hydraulic cylinder entity, and to extract data characterizing the operating state of the digital twin model to achieve condition monitoring and fault prediction.

[0049] The human-computer interaction module is used to visualize the process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system.

[0050] The host computer is an industrial control computer, and the control system runs within the industrial control computer. The two together form a host-server system, which uses a digital twin model of the hydraulic cylinder to synchronously simulate the operating state of the hydraulic cylinder entity and visualize it in the connected human-computer interaction module.

[0051] The present invention also includes a hydraulic cylinder visualization monitoring device based on a digital twin system, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it creates a virtual part in the aforementioned digital twin virtual-real interaction system. The virtual part interacts with the physical part, thereby realizing the status monitoring and fault prediction of the hydraulic cylinder entity.

[0052] The technical solution provided by this invention has the following beneficial effects:

[0053] This invention creates a 3D model of a hydraulic cylinder using a 3D modeling tool, and then uses visualization and simulation tools such as Simulink to create a digital twin model of the hydraulic cylinder that can fully simulate the operation of the physical hydraulic cylinder. A signal link is then established between the physical hydraulic cylinder and its twin model via a data interface to ensure synchronous operation, thus forming a digital twin virtual-physical interactive system.

[0054] This invention utilizes a digital twin model of a hydraulic cylinder within a created digital twin virtual-real interaction system as the monitoring object. It collects and analyzes data on the cylinder's operational status, enabling indirect detection of the cylinder's operational status and component damage. Furthermore, it supports the prediction of hydraulic cylinder failures using equivalent data collected from the digital twin model. Simultaneously, the digital twin virtual-real interaction system created by this invention also supports visualization of the status monitoring and failure prediction process. Attached Figure Description

[0055] Figure 1 This is a flowchart illustrating the steps of a digital twin modeling and fault prediction method for a large hydraulic cylinder provided in Embodiment 1 of the present invention.

[0056] Figure 2 This is a schematic diagram of a digital twin modeling and fault prediction method for a large hydraulic cylinder provided in Embodiment 1 of the present invention.

[0057] Figure 3 This is the parametric geometric model of the hydraulic cylinder constructed in Embodiment 1 of the present invention.

[0058] Figure 4 A simplified diagram of the assembly relationships of parts in the parametric geometric model of a hydraulic cylinder.

[0059] Figure 5 This is a schematic diagram illustrating the process of creating the equivalent multibody dynamics model of the hydraulic cylinder in Embodiment 1 of the present invention.

[0060] Figure 6 This is a schematic diagram of the hydraulic subsystem created in Embodiment 1 of the present invention.

[0061] Figure 7 This is a schematic diagram of the principle architecture of a digital twin virtual-real interaction system created in Embodiment 2 of the present invention. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0063] Example 1

[0064] This embodiment provides a digital twin modeling and fault prediction method for large hydraulic cylinders, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0065] S1: Based on the physical characteristics, material parameters and actual working conditions of the hydraulic cylinder, a parametric geometric model of the hydraulic cylinder assembly is established using 3D modeling software, and a mechanical model of the hydraulic cylinder with a digital twin virtual-real interaction interface is generated.

[0066] The detailed steps for creating a hydraulic cylinder mechanical model are as follows:

[0067] S11: Simplify the structure of the hydraulic cylinder, such as... Figure 3 As shown, the simplified hydraulic cylinder assembly includes the following parts: sealing ring, guide sleeve, piston rod, cylinder barrel, piston sealing ring, piston, and cylinder bottom.

[0068] S12: Define the shape, size, structure and constraint relationships of the parts in the 3D modeling software SolidWorks to model the parts, and assemble the parts according to the spatial assembly relationships (such as parallel, coaxial, overlapping) and motion characteristics to obtain the parametric model of the assembly.

[0069] The constraint relationships between the various parts of the hydraulic cylinder created in this embodiment are as follows: Figure 4 As shown, it includes the following:

[0070] (1) The guide sleeve and the cylinder are in a concentric and overlapping fit;

[0071] (2) The guide ring and the sealing ring are in a concentric fit;

[0072] (3) The cylinder and piston rod are in a concentric fit and can move relative to each other along the axial direction;

[0073] (4) The piston rod and the cylinder bottom are in a concentric fit;

[0074] (5) The piston rod and the piston are in a concentric and overlapping fit;

[0075] (6) The piston and piston seal are in a concentric and overlapping fit.

[0076] S2: Use a conversion plugin to import the hydraulic cylinder mechanical system into Simulink and establish an equivalent multibody dynamics model of the hydraulic cylinder. In this embodiment, as shown... Figure 5 As shown, the process of creating the equivalent multibody dynamics model of the hydraulic cylinder is as follows:

[0077] S21: The parameterized model of the hydraulic cylinder assembly is converted into an XML file using the Simscape Multibody Link model conversion plugin. The file stores the following information and constraints: cylinder inner diameter, piston rod diameter, piston stroke, piston area, and seal area.

[0078] The parametric model of the assembly provides a parameter adjustment interface to adjust the structure of the hydraulic cylinder. When it is necessary to adjust the specifications or model of the hydraulic cylinder, it can be done simply by modifying the specified parameters of the parametric model.

[0079] S22: The XML file provides transformation nodes recognizable by the multi-domain modeling language for the established hydraulic cylinder assembly model. The model is identified and transformed using MATLAB / Simulink execution commands, converting the XML file into an SLX program file. The constructed hydraulic cylinder assembly model is described using the multi-domain modeling language, and information such as the coordinate system and connection relationships of the model are corrected.

[0080] In this embodiment, rigid bodies, joints, forces / torques, connectors, etc., are used to connect the components of the mechanical parts, and a spatial coordinate system is established. In the established spatial coordinate system, the x-axis and z-axis are horizontal, and the y-axis points upwards.

[0081] S23: Use the Prismatic Joint module and Translational Multibody Interface module in Matlab / Simulink software to implement signal conversion between the hydraulic machinery model and the hydraulic subsystem. The interface information transmitted between the hydraulic machinery model and the hydraulic subsystem includes position, velocity, and force information.

[0082] Thus, the hydraulic cylinder mechanical model established in this embodiment can output time-series data such as force, displacement, velocity, and acceleration through simulation analysis, while the constructed equivalent multibody dynamics model of the hydraulic cylinder has a digital twin virtual-real interaction functional interface.

[0083] S3: Create a hydraulic subsystem in Simulink; fuse the equivalent multibody dynamics model of the hydraulic cylinder with the hydraulic subsystem, realizing the physical connection and drive control between the components, to obtain the required digital twin model of the hydraulic cylinder. The detailed steps of creating the digital twin model of the hydraulic cylinder are as follows:

[0084] S31: Create a hydraulic subsystem in Simulink that includes a hydraulic pump, angular velocity source, relief valve, first proportional valve, second proportional valve, double-acting hydraulic cylinder, oil tank, pressure sensor, and flow sensor.

[0085] S32: The hydraulic subsystem is used as the driving part of the equivalent multibody dynamics model of the hydraulic cylinder, thus forming the hydraulic system.

[0086] S34: Define and configure the control parameters of the hydraulic system; based on the input control parameters, the hydraulic system generates a corresponding driving force through the hydraulic cylinder, realizing the kinematic simulation of the digital twin model of the hydraulic cylinder. The control parameters include the first proportional valve control signal, the second proportional valve control signal, and the hydraulic pump control signal.

[0087] like Figure 6 As shown, in this embodiment, the first proportional valve control signal is a voltage control signal, which is connected to the first proportional valve to control its opening degree. The working port of the first proportional valve is connected to the rodless chamber of the double-acting hydraulic cylinder, the inlet port of the first proportional valve is connected to the hydraulic pump outlet, and the return port of the first proportional valve is connected to the oil tank. The second proportional valve control signal is a voltage control signal, which is connected to the second proportional valve to control its opening degree. The working port of the second proportional valve is connected to the rod chamber of the double-acting hydraulic cylinder, the inlet port of the second proportional valve is connected to the hydraulic pump outlet, and the return port of the second proportional valve is connected to the oil tank. The hydraulic pump control signal is used to control the speed of the servo motor connected to the hydraulic pump, and its signal output interface is connected to the Translational Multibody Interface motion interface conversion module.

[0088] S4: Construct a digital twin virtual-real interaction system that includes a digital twin model of the hydraulic cylinder, a hardware system, and a control system. The digital twin virtual-real interaction system is driven by the acquired drive signals of the hydraulic cylinder entity, and performs real-time simulation of the hydraulic cylinder entity's operating state in the digital twin model.

[0089] The hardware system includes a hydraulic cylinder body, pressure sensors, displacement sensors, a data acquisition card, and communication cables. The hydraulic cylinder body comprises a cylinder, a servo motor, hydraulic lines, and drive components. Pressure and displacement sensors are mounted on the hydraulic cylinder body; the former acquires the pressure in the rod-side chamber and the pressure in the rodless chamber, while the latter acquires the piston's displacement and velocity. The data acquisition card acquires the detection data from the pressure and displacement sensors and communicates with a host computer containing the control system via the communication cable.

[0090] The control system employs Simulink Real-Time to execute the control of the hydraulic system. It then acquires digital or analog control signals from the hydraulic cylinder entity via a signal acquisition terminal, and controls the simulation operation of the hydraulic cylinder's digital twin model. The digital signals originate from signal sampling of the drive components during the cylinder's downward and return strokes; the analog signals are derived from the detection results of displacement and pressure sensors. In the hydraulic cylinder's digital twin model, the digital signals are used to control the on / off switching of the solenoid valve, while the analog signals are used to adjust the valve opening of the proportional valve and the speed of the servo motor.

[0091] S5: Set a preset threshold for simulation accuracy, and perform accuracy testing and optimization on the digital twin model of the hydraulic cylinder created in the digital twin virtual-real interaction system.

[0092] (1) When the simulation accuracy reaches the threshold, the corresponding model parameters are saved;

[0093] (2) When the simulation accuracy is lower than the threshold, return to step S3 to correct the parameters of the created hydraulic subsystem.

[0094] Specifically, in this embodiment, the accuracy test of the digital twin model of the hydraulic cylinder is achieved through the following method:

[0095] First, when the digital twin model of the hydraulic cylinder operates synchronously with the physical hydraulic cylinder, the displacement d and velocity v of the piston, as well as the pressure p1 in the rod chamber and the pressure p2 in the rodless chamber, are sampled synchronously to form a test vector A: A = {d, v, p1, p2}.

[0096] Secondly, test vectors associated with the digital twin model and the physical hydraulic cylinder are synchronously acquired from the hydraulic cylinder according to a preset acquisition frequency, denoted as y. i and Y i .

[0097] Finally, the goodness of fit R is calculated using a series of test vectors collected during the testing period. 2 The simulation accuracy of the digital twin model of the hydraulic cylinder is evaluated by the goodness of fit. A goodness of fit closer to 1 indicates higher accuracy of the created digital twin model. The goodness of fit Rfit is... 2 The calculation formula is as follows:

[0098]

[0099] In the above formula: i represents the sampling time of different data points; n is the sample size in the test data; y i Y is the test vector composed of the operating parameters of the hydraulic cylinder at time i; i The test vector is composed of the operating parameters of the digital twin model of the hydraulic cylinder at time i; This represents a vector composed of the average values ​​of various operating data of the digital twin model of the hydraulic cylinder.

[0100] S6: Utilizing a digital twin virtual-real interaction system that meets accuracy requirements, the operating status of the hydraulic system is visualized and monitored. Using the real-time acquired monitoring data as samples, a fault prediction model based on machine learning algorithms is trained to achieve fault prediction of the hydraulic system.

[0101] In this embodiment, the visualization and monitoring process based on the digital twin model of the hydraulic cylinder is as follows:

[0102] (1) Real-time pressure and displacement characteristic data are collected by sensors on the actual hydraulic platform, and piston rod driving force data are obtained from the hydraulic cylinder twin model. The twin data, composed of actual sensor data and simulation data including normal operation data and fault simulation data, is used to expand the dataset. First, the collected feature dataset is transformed, and the high-dimensional feature data is linearly combined to transform it into a more analyzable form. Independent main features are constructed to achieve dimensionality reduction with minimal loss of original information.

[0103] For example, in this embodiment, the collected feature data is rewritten into an m×n dimensional matrix form. The feature data matrix collects m sets of data, each set containing n features, and the matrix form is as follows:

[0104]

[0105] To ensure the accuracy of subsequent principal component analysis, data standardization is performed. Each feature value is converted into a standard score relative to the mean and standard deviation of that feature, eliminating differences in dimensions and magnitudes between different features, thus ensuring the data are on a similar scale. The specific standardization formula is as follows:

[0106]

[0107] In the above formula, x ij This represents the elements in the data sample matrix, where μ represents the mean of the sample data. σ represents the standard deviation of the sample data.

[0108] Next, we calculate the covariance matrix Ω of the dataset. The covariance matrix is ​​used to measure the correlation and variation between different features. We then use the covariance matrix to find its eigenvalues ​​and eigenvectors.

[0109] The formula for covariance is as follows:

[0110] c ij =Cov(X) i X j )=E{[X i -E(X i )][X j -E(X j )]}=E[X i X j ]-E[X i ][X j ]

[0111] i, j = 1, 2, ..., n

[0112] covariance matrix

[0113] In the above formula, X i and X j Cov(X) represents a vector composed of samples of any two feature data types. i X j ) represents X i and X j covariance; E(X) i ) represents vector X i The expected value; E(X) j ) represents vector X j Expected value;

[0114] The formulas for eigenvalues ​​and eigenvectors are as follows:

[0115] |Ω-λI|=0

[0116] (Ω-λI)A=0

[0117] In the above formula, Ω is the covariance matrix, λ is the eigenvalue, I is the identity matrix, and A is the eigenvector. The principal features for analyzing the changes in the loss-type parameters are selected based on the calculated eigenvalues. The magnitude of the eigenvalues ​​reflects the proportion of variance of the principal components in the overall features. The principal component contribution rate δ and the cumulative contribution rate are also considered. The calculation formula is as follows:

[0118]

[0119]

[0120] In the above formula, λ1, λ2,…,λ m These are the corresponding 1st, 2nd, ..., mth (m≤n) principal components.

[0121] If the cumulative contribution rate of its eigenvalues ​​reaches more than 85%, then a combined feature can be constructed from its corresponding eigenvectors as the main feature. The larger the eigenvalue, the more information the corresponding principal component explains in the overall variance. This feature is selected as the main feature for analyzing the changes in the loss-type parameter.

[0122] (2) Adjustable throttle orifices are connected to the inlet and outlet of the hydraulic cylinder in the twin model. Changing the size of the throttle orifice causes changes in pressure and flow rate, simulating a pipeline blockage fault module. By adjusting the parameters of the sealing ring, including the elastic modulus and friction coefficient, a sealing ring wear fault module is set. A sliding resistance module is added to simulate a piston wear fault module that hinders piston movement. The inlet and outlet oil lines are connected through throttle orifices, and the size of the throttle orifice is adjusted to simulate an internal leakage fault at the piston of the hydraulic cylinder. The twin simulation model is run, and the corresponding fault data is imported into the MATLAB workspace. Principal component analysis is used to extract the main features reflecting the fault type, and the correlation between the main feature parameters and the fault type is established.

[0123] (3) Using the main features of wear-related parameter changes as samples, the simulation results of the digital twin model of the hydraulic cylinder are used to extract the training fault data by FFT, and the time domain data is converted into frequency domain data to obtain the features of different fault types in the frequency domain. A machine learning algorithm is designed and trained to predict typical faults including seal wear, pipeline blockage, piston wear, and internal leakage. Fault prediction is performed through the trained machine learning algorithm.

[0124] (4) Based on the sampling data of the operation process of the digital twin model of the hydraulic cylinder, establish the curve of the main characteristic of the hydraulic system entity output changing with time, and display the curve of the fault parameter changing with time in the fault parameter prediction coordinate area; establish a state space model based on the changing trend of the parameter curve, estimate the system state at future time points, and realize the state monitoring function. Furthermore, store the real-time physical data in the fault dataset, and use the real physical fault data to update the simulation fault data.

[0125] (6) The process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system are visualized using any visualization method. Among them, the analysis process of visualizing the digital twin model of the hydraulic cylinder is implemented in the Mechanics Explorer module of Matlab / Simulink.

[0126] Example 2

[0127] Based on the scheme in Example 1, this embodiment provides a digital twin virtual-real interaction system, which is created using the digital twin modeling and fault prediction method for large hydraulic cylinders as described in Example 1. The digital twin virtual-real interaction system is used to synchronously and fully simulate the operating state of the hydraulic cylinder entity, thereby assisting in the state monitoring and fault prediction of the hydraulic cylinder entity. Figure 7 As shown, the digital twin virtual-real interaction system includes: a digital twin model of the hydraulic cylinder, a signal acquisition module, a control system, and a human-computer interaction module.

[0128] A digital twin model of a hydraulic cylinder, serving as a virtual model of the hydraulic cylinder entity, can be used to simulate the operating state of the hydraulic cylinder entity in real time.

[0129] The signal acquisition module is used to acquire the operating status parameters of the hydraulic cylinder and upload them to the host computer. The signal acquisition module includes a signal acquisition unit, a pressure sensor, a displacement sensor, a data acquisition card, and a communication cable. The signal acquisition unit acquires the electrical signals transmitted by the drive components of the hydraulic cylinder. The pressure sensor and displacement sensor are mounted on the hydraulic cylinder; the former acquires the rod chamber pressure and rodless chamber pressure, while the latter acquires the piston displacement and velocity. The data acquisition card acquires the detection data from the signal acquisition unit, pressure sensor, and displacement sensor. The data acquisition card communicates with the host computer via the communication cable.

[0130] The control system is used to synchronously drive the digital twin model of the hydraulic cylinder based on electrical signals collected from the drive components of the hydraulic cylinder entity, and to extract data characterizing the operating state of the digital twin model to achieve condition monitoring and fault prediction.

[0131] The human-computer interaction module is used to visualize the process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system.

[0132] The host computer is an industrial control computer, and the control system runs within the industrial control computer. The two together form a host-server system, which uses a digital twin model of the hydraulic cylinder to synchronously simulate the operating state of the hydraulic cylinder entity and visualize it in the connected human-computer interaction module.

[0133] Example 3

[0134] This embodiment provides a hydraulic cylinder visualization monitoring device based on a digital twin system, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it creates a virtual part in the digital twin virtual-real interaction system as shown in Embodiment 2. The virtual part interacts with the physical part, thereby realizing status monitoring and fault prediction of the hydraulic cylinder entity.

[0135] In this embodiment, the hydraulic cylinder visualization monitoring device based on a digital twin system is essentially a computer device for data processing and instruction generation. It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. Its front-end device is the actual hydraulic cylinder, and its back-end device is a human-computer interaction module, such as a mouse, keyboard, and monitor, that can visualize the status monitoring process and fault prediction results.

[0136] The computer device provided in this embodiment may be a smart terminal, tablet computer, laptop computer, desktop computer, rack server, blade server, tower server, or cabinet server (including independent servers or server clusters composed of multiple servers) capable of executing programs. The computer device in this embodiment includes, but is not limited to, a memory and a processor that can be interconnected via a system bus.

[0137] In this embodiment, the memory (i.e., the readable storage medium) includes flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory may be an internal storage unit of a computer device, such as the hard disk or RAM of the computer device.

[0138] In other embodiments, the memory can also be an external storage device of the computer device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Of course, the memory can also include both internal storage units and external storage devices of the computer device. In this embodiment, the memory is typically used to store the operating system and various application software installed on the computer device. Furthermore, the memory can also be used to temporarily store various types of data that have been output or will be output.

[0139] In some embodiments, a processor may be a central processing unit (CPU), a graphics processing unit (GPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor is typically used to control the overall operation of a computer device. In this embodiment, the processor is used to run program code stored in memory or process data.

[0140] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A digital twin modeling and fault prediction method for large hydraulic cylinders, characterized in that, It includes the following steps: S1: Based on the physical characteristics, material parameters and actual working conditions of the hydraulic cylinder, a parametric geometric model of the hydraulic cylinder assembly is established using 3D modeling software, and a mechanical model of the hydraulic cylinder with a digital twin virtual-real interaction interface is generated. S2: Use the conversion plugin to import the hydraulic cylinder mechanical model into Simulink and establish an equivalent multibody dynamics model of the hydraulic cylinder; S3: Create a hydraulic subsystem in Simulink; fuse the equivalent multibody dynamics model of the hydraulic cylinder with the hydraulic subsystem to achieve physical connection and drive control between the components, obtaining the required digital twin model of the hydraulic cylinder; the detailed steps of creating the digital twin model of the hydraulic cylinder are as follows: S31: Create a hydraulic subsystem in Simulink that includes a hydraulic pump, angular velocity source, relief valve, first proportional valve, second proportional valve, double-acting hydraulic cylinder, oil tank, pressure sensor, and flow sensor. S32: The hydraulic subsystem is used as the driving part of the equivalent multibody dynamics model of the hydraulic cylinder to form a hydraulic system; S33: Define and configure the control parameters of the hydraulic system; the hydraulic system generates a corresponding driving force through the hydraulic cylinder according to the input control parameters, realizing the kinematic simulation of the digital twin model of the hydraulic cylinder; The control parameters include a first proportional valve control signal, a second proportional valve control signal, and a hydraulic pump control signal; The first proportional valve control signal is a voltage control signal, which is connected to the first proportional valve to control the opening degree of the first proportional valve; the working port of the first proportional valve is connected to the rodless chamber of the double-acting hydraulic cylinder, the oil inlet of the first proportional valve is connected to the outlet of the hydraulic pump, and the oil return port of the first proportional valve is connected to the oil tank. The control signal for the second proportional valve is a voltage control signal, which is connected to the second proportional valve to control the opening degree of the second proportional valve; the working port of the second proportional valve is connected to the rod chamber of the double-acting hydraulic cylinder, the inlet port of the second proportional valve is connected to the outlet of the hydraulic pump, and the return port of the second proportional valve is connected to the oil tank. The hydraulic pump control signal is used to control the speed of the servo motor connected to the hydraulic pump, and its signal output interface is connected to the Translational Multibody Interface motion interface conversion module. S4: Construct a digital twin virtual-real interaction system including a digital twin model of a hydraulic cylinder, a hardware system, and a control system; the digital twin virtual-real interaction system is driven by the acquired driving signals of the hydraulic cylinder entity, and performs real-time simulation of the operating state of the hydraulic cylinder entity in the digital twin model; S5: Set a preset threshold for simulation accuracy, and perform accuracy testing and optimization on the digital twin model of the hydraulic cylinder created in the digital twin virtual-real interaction system: (1) When the simulation accuracy reaches the threshold, the corresponding model parameters are saved; (2) When the simulation accuracy is lower than the threshold, return to step S3 to correct the parameters of the created hydraulic subsystem; S6: Utilize a digital twin virtual-real interaction system that meets accuracy requirements to visualize and monitor the operating status of the hydraulic system, and use the real-time acquired monitoring data as samples to train a fault prediction model based on machine learning algorithms to achieve fault prediction of the hydraulic system.

2. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 1, characterized in that: In step S1, the detailed steps of creating the hydraulic cylinder mechanical model are as follows: S11: The structure of the hydraulic cylinder assembly is simplified to include the following parts: sealing ring, guide sleeve, piston rod, cylinder barrel, piston sealing ring, piston, and cylinder bottom; S12: Define the shape, size, structure and constraint relationships of the parts in 3D modeling software, and perform part modeling and assembly to obtain a parametric model of the assembly.

3. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 1, characterized in that: The process of creating the equivalent multibody dynamics model of the hydraulic cylinder is as follows: S21: The parameterized model of the hydraulic cylinder assembly is converted into an XML file using the Simscape Multibody Link model conversion plugin. The file's stored content and constraint information include: cylinder inner diameter, piston rod diameter, piston stroke, piston area, and seal area. S22: Identify and transform the model using MATLAB / Simulink execution commands, converting XML files into SLX program files; describe the constructed hydraulic cylinder assembly model using a multi-domain modeling language, and correct the coordinate system and connection relationships of the above model; S23: Use the Prismatic Joint module and Translational Multibody Interface module in Matlab / Simulink software to realize signal conversion between the hydraulic machinery model and the hydraulic subsystem.

4. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 3, characterized in that: In step S21, the parameter adjustment interface is provided in the parameterized model of the assembly; In step S22, the XML file provides transformation nodes that can be recognized by a multi-domain modeling language for the established hydraulic cylinder assembly model; In step S23, the interface information transmitted between the hydraulic cylinder mechanical model and the hydraulic subsystem includes position, speed, and force information.

5. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 1, characterized in that: In step S4, the hardware system includes a hydraulic cylinder body, a pressure sensor, a displacement sensor, a data acquisition card, and a communication cable; the hydraulic cylinder body includes an oil cylinder, a servo motor, hydraulic lines, and a drive assembly; the pressure sensor and the displacement sensor are mounted on the hydraulic cylinder body, the former is used to acquire the rod chamber pressure and rodless chamber pressure of the hydraulic cylinder body, and the latter is used to acquire the piston displacement and movement speed; the data acquisition card is used to acquire the detection data of the pressure sensor and the displacement sensor, and communicates with the host computer containing the control system through the communication cable; The control system uses Simulink Real-Time to execute the control of the hydraulic system; it then acquires digital or analog control signals from the hydraulic cylinder entity through a signal acquisition terminal and controls the simulation operation of the digital twin model of the hydraulic cylinder; the digital signals come from signal sampling of the drive components during the downward and return strokes of the hydraulic cylinder; the analog signals come from displacement and pressure sensors; in the digital twin model of the hydraulic cylinder, the digital signals are used to control the on / off switching of the solenoid valve, and the analog signals are used to control the valve opening of the proportional valve and the speed of the servo motor.

6. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 1, characterized in that: In step S5, the digital twin model of the hydraulic cylinder operates synchronously with the physical hydraulic cylinder. The displacement d and velocity v of the piston, as well as the pressure p1 in the rod chamber and p2 in the rodless chamber, are simultaneously sampled during their operation to form a test vector A: A = {d, v, p1, p2}. The fitting degree R is then calculated using the following formula. 2 To evaluate the simulation accuracy of the digital twin model of the hydraulic cylinder: In the above formula: i represents the sampling time of different data points; n is the sample size in the test data; y i Y is the test vector composed of the operating parameters of the hydraulic cylinder at time i; i The test vector is composed of the operating parameters of the digital twin model of the hydraulic cylinder at time i; This represents a vector composed of the average values ​​of various operating data of the digital twin model of the hydraulic cylinder.

7. The digital twin modeling and fault prediction method for large hydraulic cylinders as described in claim 1, characterized in that: The detailed process of step S6 is as follows: (1) Real-time acquisition of feature data through sensors on the actual hydraulic platform, transformation of the acquired feature dataset, linear combination of high-dimensional feature data, conversion into a more analyzable form, construction of mutually independent main features, and dimensionality reduction with minimal loss of original information. (2) An adjustable throttle orifice is connected to the inlet and outlet of the hydraulic cylinder in the twin model. Changing the size of the throttle orifice causes changes in pressure and flow rate, simulating the pipeline blockage fault module; by adjusting the parameters of the sealing ring, including the elastic modulus and friction coefficient of the sealing ring, a sealing ring wear fault module is set; by adding a sliding resistance module, a piston wear fault module that hinders piston movement is simulated; by connecting the inlet and outlet oil pipelines through the throttle orifice and adjusting the size of the throttle orifice, an internal leakage fault at the piston of the hydraulic cylinder is simulated; the twin simulation model is run, the corresponding fault data is selected and imported into the MATLAB workspace, and the main features reflecting the fault type are extracted through principal component analysis to establish the correlation between the main feature parameters and the fault type; (3) Using the main features of wear-type parameter changes as samples, the simulation results of the digital twin model of the hydraulic cylinder are used to extract the training fault data by FFT; a machine learning algorithm is designed and trained to predict typical faults including seal wear, pipeline blockage, piston wear and internal leakage, and fault prediction is performed by the trained machine learning algorithm. (4) Based on the sampling data of the operation process of the digital twin model of the hydraulic cylinder, establish the curve of the main feature of the hydraulic system entity output as a function of time, and display the curve of the fault parameter as a function of time in the fault parameter prediction coordinate area; establish the state space model based on the trend of the parameter curve, estimate the system state at future time points and realize the state monitoring function. (5) Visualization techniques are used to visualize the process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system.

8. A digital twin virtual-real interaction system, characterized in that: It is created using the digital twin modeling and fault prediction method for large hydraulic cylinders as described in any one of claims 1-6; the digital twin virtual-real interaction system is used to realize the synchronous full simulation of the operating state of the hydraulic cylinder entity, thereby assisting in the state monitoring and fault prediction of the hydraulic cylinder entity. The digital twin virtual-real interaction system includes: A digital twin model of a hydraulic cylinder, which serves as a virtual model of the hydraulic cylinder entity, can be used to simulate the operating state of the hydraulic cylinder entity in real time. A signal acquisition module is used to acquire the operating status parameters of the hydraulic cylinder entity and upload them to a host computer. The signal acquisition module includes a signal acquisition unit, a pressure sensor, a displacement sensor, a data acquisition card, and a communication cable. The signal acquisition unit acquires the electrical signals transmitted by the drive components of the hydraulic cylinder entity. The pressure sensor and displacement sensor are mounted on the hydraulic cylinder entity; the former acquires the rod chamber pressure and rodless chamber pressure, and the latter acquires the piston displacement and velocity. The data acquisition card acquires the detection data from the signal acquisition unit, pressure sensor, and displacement sensor. The data acquisition card is connected to the host computer via the communication cable. The control system is used to synchronously drive the digital twin model of the hydraulic cylinder according to the electrical signals collected from the drive components of the hydraulic cylinder entity, and to extract data characterizing the operating status of the digital twin model to achieve status monitoring and fault prediction. The human-computer interaction module is used to visualize the process and results of status monitoring and fault prediction in the digital twin virtual-real interaction system; The host computer is an industrial control computer, and the control system runs within the industrial control computer. The two together form a host-server system, which uses a digital twin model of the hydraulic cylinder to synchronously simulate the operating state of the hydraulic cylinder entity and visualize it in the connected human-computer interaction module.

Images