CN115423132A

CN115423132A - Engineering machinery predictive maintenance method based on digital twinning

Info

Publication number: CN115423132A
Application number: CN202211217345.3A
Authority: CN
Inventors: 唐宏宾; 李志祥; 贺湘宇; 徐晓强; 何知义; 龚杨春; 唐一
Original assignee: Changsha University of Science and Technology
Current assignee: Changsha University of Science and Technology
Priority date: 2022-10-03
Filing date: 2022-10-03
Publication date: 2022-12-02

Abstract

The invention discloses a predictive maintenance method for engineering machinery based on digital twinning, which comprises the following steps: the physical layer consists of engineering machinery, an industrial personal computer and an external environment; the digital twin model layer realizes the comprehensive, real, objective and real-time mapping of the physical layer in the digital world; the data perception layer is used for acquiring, transmitting and storing real-time data and historical data generated in the operation of the engineering machinery and realizing the connection between the physical layer and the digital twin model layer; the model and data fusion layer is used for processing data generated by the physical layer and the digital twin model layer, realizing effective fusion of the model and the data and constructing a predictive maintenance algorithm; and the intelligent service layer is used for realizing intelligent predictive maintenance service of the engineering machinery according to the prediction result. The invention can effectively solve the problems of low fidelity of a digital model in the whole life cycle, insufficient fault data, poor fusion of the model and the data and low predictive maintenance precision in the current predictive maintenance of engineering machinery.

Description

Engineering machinery predictive maintenance method based on digital twinning

Technical Field

The invention belongs to the technical field of engineering machinery predictive maintenance intellectualization and digitization, and particularly relates to an engineering machinery predictive maintenance method based on digital twins.

Background

Engineering machinery is a core device in the fields of transportation construction, large-scale mine construction, industry, civil construction and the like. Under the background of industry 4.0, the degree of integration, automation and intelligence of engineering machinery is higher and higher, and the dependence of production activities on the engineering machinery is stronger and stronger. However, due to the badness of the working environment of the engineering machinery and the uncertainty of external interference, the engineering machinery is prone to performance degradation, damage failure and other behaviors, which often result in huge economic loss and casualties.

At present, fault prediction and health management strategies for complex mechanical equipment mainly comprise corrective, preventive and predictive maintenance methods. The predictive maintenance can effectively ensure the reliability and stability of the system and improve the fault-free running time. Predictive maintenance mainly includes: a physical model-based method, a data-driven-based method and a fusion-type predictive maintenance method. The method based on the physical model realizes the predictive maintenance service by establishing a mathematical model reflecting the performance degradation rule of the equipment, but the degradation mechanism of the complex equipment is not thoroughly understood at present. The data-driven method does not need to know specific degradation rules by mining a large amount of state monitoring data, but the method depends on the quantity and quality of data to a great extent. And the fusion type predictive maintenance can realize the performance complementation between the two methods, and effectively avoid the limitation of a single method.

However, how to construct a high-fidelity digital model in the whole life cycle of the equipment, acquire sufficient fault data, realize efficient fusion of the model and the data and realize intelligent predictive maintenance still remains a key technical problem to be solved. Therefore, a scientific and intelligent maintenance method is urgently needed to accurately predict the remaining service life and potential faults of the engineering machinery, so that the fault occurrence rate of the equipment is reduced.

Disclosure of Invention

The invention provides a digital twinning technology applied to predictive maintenance of engineering machinery, aiming at the problems of low fidelity of a digital model, insufficient fault data, poor fusion of the model and the data and low predictive maintenance precision in the whole life cycle in the predictive maintenance of the engineering machinery at present. Establishing a predictive maintenance system comprising a physical layer, a digital twin model layer, a data perception layer, a model and data fusion layer and an intelligent service layer; the method has the advantages that a multi-field unified modeling language Simscape is utilized, the construction of a high-fidelity digital twin model is achieved, and the problems that the existing model is low in fidelity and insufficient in fault data are effectively solved; the method has the advantages that the method combines the traditional machine learning and the deep learning, reduces the calculated amount, and fully excavates data; meanwhile, the advantages of transfer learning and particle filter algorithms are utilized, so that the high-efficiency fusion of the model and data is realized, and the accuracy and the applicability of a predictive maintenance algorithm are ensured; and finally, realizing the intelligent predictive maintenance service with high accuracy for the engineering machinery through an intelligent service layer.

In order to realize the purpose, the invention adopts the following technical scheme: a predictive maintenance method for a construction machine based on digital twinning, comprising:

1. a predictive maintenance method of engineering machinery based on digital twins comprises a physical layer, a digital twins model layer, a data perception layer, a model and data fusion layer and an intelligent service layer;

the data perception layer collects, transmits and stores real-time data and historical data generated by the physical layer, and transmits the data to the digital twin model layer in a wired/wireless mode to realize real-time simulation and update of the model; meanwhile, simulation data generated by the digital twin model layer is used for design optimization of the physical layer engineering machinery; simultaneously inputting data acquired by the data sensing layer and model simulation data generated by the digital twin model layer into the model and the data fusion layer, and after data processing, using the data to construct a data driving algorithm and a fusion algorithm; and finally, the intelligent service layer realizes the intelligent predictive maintenance service of the engineering machinery through a prediction result and a visualization technology.

2. The physical layer comprises engineering machinery, an industrial personal computer and an external environment; the industrial personal computer is equipment used for storing historical data in the actual operation process of the engineering machinery; the external environment refers to an external factor in actual operation of the construction machine.

3. The digital twin model layer comprises a multi-field coupling model, a behavior model, a rule model, a fault model and a model precision verification and updating mechanism.

4. The digital twin model layer establishing steps are as follows:

the method comprises the following steps that S1, aiming at the engineering machinery, the engineering machinery is divided into a mechanical part, an electrical part, a hydraulic part and a control part, and an object-oriented, incremental and multi-field unified modeling principle is adopted;

s2, establishing a mechanical model, an electrical model, a hydraulic model and a control model in an ideal state by adopting a multi-field unified modeling language Simscape, and realizing the coupling of the multi-field models by means of an energy interface provided by the Simscape;

s3, considering behaviors of the engineering machinery, such as evolution, performance degradation and the like, generated under the combined action of an external environment and interference and an internal operation mechanism under different time scales, and establishing a behavior model by adopting a Markov chain and a finite state machine without limitation;

s4, continuously mining the implicit knowledge and the related industry standards to generate new rules and knowledge, and establishing a rule model based on the technologies of traditional machine learning, deep learning and the like to enable the digital twin model to have the capabilities of self-growth, self-evolution and self-learning;

s5, establishing a corresponding fault model generated in the actual operation process of the engineering machinery according to a fault mechanism, and providing sufficient fault data for the construction of a predictive maintenance algorithm;

s6, collecting actual operation data of the engineering machinery, simulating a multi-field coupling model in an ideal state under the same working condition, comparing the data obtained by simulation with the actual data, and continuously increasing the details of the model to achieve the actual physical engineering mechanical characteristics;

and step S7, realizing real-time updating of the digital twin model through a model updating mechanism comprising a behavior model and a mapping relation between actual operation data and model simulation data.

5. The data perception layer comprises a sensor, an industrial communication protocol, a wired/wireless network and data storage.

6. The data perception layer is established by the following steps:

s1, acquiring required real-time operation data and environmental parameters through a large number of high-precision sensors, such as pressure, temperature, vibration, acoustic emission and the like; on the other hand, historical data is collected through equipment such as an industrial personal computer;

s2, constructing a server end and a client end of the OPC-UA (organic Process Control-Unified Architecture, OPC-UA) based on an OPC-UA (organic Process Control-Unified Architecture) engineering machinery scene perception data acquisition technology, wherein the server end is responsible for analyzing multi-sensor data and providing the data to the client end for Unified data, and solving the problems of a data interface form and a communication protocol;

s3, a storage mode combining distributed edge storage and centralized cloud storage is adopted, edge computing is introduced, data for real-time analysis are transmitted to computing equipment closest to a data source through a network for processing and analysis, local storage is realized, the pressure of the throughput of an equipment end and a cloud network is reduced, the data analysis load of cloud center equipment is relieved, and the data analysis response efficiency and the data storage safety are improved; and the data which is not sensitive to time delay and needs to be processed in a centralized way are transmitted to the cloud center through a special line, and the cloud center can access the edge computing data at any time.

7. The model and data fusion layer comprises data processing, a data driving algorithm and a fusion algorithm.

8. The model and data fusion layer is established by the following steps:

s1, performing missing value supplement on sensing data and model simulation data by adopting but not limited to mean interpolation, similar mean interpolation or manual interpolation; smoothing and denoising the data by adopting but not limited to an average method, and eliminating random interference in the original data so as to realize the pretreatment of the data;

s2, performing missing value supplement on the sensing data and the model simulation data by adopting, but not limited to, mean interpolation, similar mean interpolation or manual interpolation; smoothing and denoising the data by adopting but not limited to an average method, and eliminating random interference in the original data so as to realize data preprocessing;

s3, in order to reduce the calculated amount and improve the prediction performance, time domain, frequency domain and time-frequency domain analysis are carried out on the data, and characteristics including average value, standard deviation, kurtosis, skewness and the like are extracted to be used as multi-source input;

s4, in order to accelerate the construction of a data-driven algorithm model, constructing a data-driven algorithm which has simple extraction of fault or performance degradation data features and can meet diagnosis and prediction requirements by utilizing traditional machine learning based on the traditional machine learning;

s5, for faults or performance degradation behaviors which are difficult to extract and have complex data quantity, automatically extracting the features and constructing a data-driven algorithm model by utilizing the extremely strong nonlinear fitting capability of deep learning;

s6, for the condition that the engineering machinery fault data or performance degradation data are easy to obtain and mark, constructing a data driving model by using the markable fault data or performance degradation data, obtaining a model-based predicted value through a digital twin model, then realizing the fusion of the model predicted value and the data driving predicted value based on a filtering algorithm, and correcting the model-based predicted value by using a data driving result as a system observation value;

and S7, injecting corresponding faults into the engineering machinery fault data or performance degradation data which are difficult to obtain and mark through the fault model, simulating to obtain corresponding fault data and performance degradation data, training the data driving model by using the data obtained through simulation, transferring the trained model to an actual application scene, and inputting the collected real-time operation data to the transferred model, thereby quickly realizing predictive maintenance.

9. The intelligent service layer comprises state monitoring, fault early warning, service life prediction and an intelligent maintenance strategy.

10. The intelligent service layer is established by the following steps:

step S1, the real-time visualization of a physical space is realized by using Simulink 3D Animation, the interaction of a virtual world is realized by other hardware devices such as a force feedback control rod and a three-dimensional mouse, the three-dimensional virtual world is viewed in a stereoscopic vision immersion manner, and a real-time virtual model of the physical space is displayed on mobile devices such as a computer, a mobile phone and a tablet computer by using a Simulink 3D Animation Web viewer in a Web browser;

and S2, realizing service encapsulation by adopting MATLAB software, and meeting personalized services of different fields and different users in a mode of generating application software, mobile terminal App or web pages. The intelligent predictive maintenance service such as visual intelligent real-time state monitoring, visual online life prediction, remote visual operation guidance and professional maintenance guidance is realized.

Compared with the prior art, the invention has the following advantages:

(1) according to the invention, a high-fidelity digital model in the whole life cycle of equipment is constructed through a multi-field unified modeling language Simscape and a model precision verification and updating mechanism, so that the problems of low fidelity of the digital model and insufficient fault data in the current predictive maintenance are effectively solved;

(2) according to the invention, a storage mode combining distributed edge storage and centralized cloud storage is adopted, so that the data analysis load of the cloud center equipment is relieved, and the data response efficiency and the data storage safety are improved;

(3) the predictive maintenance algorithm is constructed by combining the traditional machine learning and the deep learning, and the advantages of the respective algorithms are utilized to realize the efficient mining of data;

(4) by adopting a fusion type predictive maintenance method and utilizing respective advantages of transfer learning and particle filter algorithms, the method realizes efficient fusion of the model and data, and ensures the accuracy and the applicability of the predictive maintenance algorithm;

(5) the intelligent service layer is constructed, and the service encapsulation is realized in the form of application software, mobile terminal App or webpage, so that more intelligent predictive maintenance service is provided for different users, different professional fields and different business requirements.

Drawings

Fig. 1 is a structural architecture diagram of the predictive maintenance method of construction machinery based on digital twin according to the present invention.

FIG. 2 is a diagram of multi-domain model coupling according to the present invention.

FIG. 3 is a diagram of a model accuracy verification method according to the present invention.

FIG. 4 is a diagram of a fusion algorithm implementation method according to the present invention.

Detailed Description

For a more detailed explanation of the objects, technical solutions and advantages of the present invention, the present application will be described in further detail below with reference to the accompanying drawings and examples, and it is to be understood that all technical and scientific terms used in the present invention have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

The structural architecture diagram of the predictive maintenance method of the engineering machinery based on the digital twin is shown in figure 1, and the specific implementation mode is as follows:

1. the digital twinning model layer in fig. 1 is implemented as follows:

the method comprises the following steps that S1, aiming at engineering mechanical equipment of a certain model, the engineering mechanical equipment is divided into a mechanical part, an electrical part, a hydraulic part and a control part, and an object-oriented, incremental and multi-field unified modeling principle is adopted;

s2, establishing a geometric model of the model by adopting SolidWorks software, importing the geometric model into a Simscape multi-body Link plug-in unit, adding an ideal kinematic pair for the geometric model, and defining material characteristics in the geometric model so as to establish an ideal mechanical model;

s3, establishing a hydraulic model and an Electrical model of the equipment in an ideal state by using a component library in a Simscape fluid module and a Simscape electric module, and finally designing a control model for the Electrical model and the hydraulic model by using Simulink;

step S4, as shown in FIG. 2, on the basis of establishing sub-field models such as a mechanical model, a hydraulic model, an electrical model, a control model and the like, coupling of multi-field models is realized by utilizing an energy interface provided by a Simscape modeling language;

s5, taking the behaviors of the equipment such as evolution, performance degradation and the like generated under the combined action of the external environment and interference and an internal operation mechanism under different time scales into consideration, and establishing a behavior model by adopting but not limited to a Markov chain and a finite state machine;

s6, continuously mining the implicit knowledge and the related industry standards to generate new rules and knowledge, and establishing a rule model based on the technologies of traditional machine learning, deep learning and the like to enable the digital twin model to have the capabilities of self-growth, self-evolution and self-learning;

s7, establishing a corresponding fault model generated in the actual operation process of the equipment according to a fault mechanism, and providing sufficient fault data for the construction of a predictive maintenance algorithm;

s8, introducing a Band-Limited White Noise and Zero-Order Hold module in Simulink into the model for simulating the noisy of the actual environment and the discrete sampling of the sensor;

step S9, as shown in FIG. 3, collecting actual operation data of the equipment, then simulating the multi-field coupling model in an ideal state under the same working condition, comparing the simulated data with the actual data, adjusting model parameters by using Simulink Design Optimization, and achieving the characteristics of the actual physical equipment by continuously increasing model details;

step S10 is to realize real-time updating of the digital twin model through a model updating mechanism comprising a behavior model and a mapping relation between actual operation data and model simulation data.

2. The data sensing layer in fig. 1 is implemented as follows:

s1, acquiring required real-time operation data and environmental parameters through a large number of high-precision sensors, such as pressure, temperature, vibration, acoustic emission and the like; on the other hand, historical data is stored through equipment such as an industrial personal computer;

s3, adopting a storage mode of combining distributed edge storage and centralized cloud storage, introducing edge computing, transmitting data for real-time analysis to computing equipment closest to a data source through 4G/5G/Wi-Fi and the like for processing analysis and realizing local storage, reducing the pressure of the throughput of an equipment end and a cloud network, relieving the data analysis load of cloud center equipment, and improving the data analysis response efficiency and the data storage safety; and the data which is not sensitive to time delay and needs to be processed in a centralized way are transmitted to the cloud center through a special line, and the cloud center can access the edge computing data at any time.

3. In the model and data fusion layer in fig. 1, the specific implementation of the data processing and data driving algorithm in this layer is as follows:

step S1, MATLAB is used for supplementing missing values to perception data and model simulation data through mean interpolation, similar mean interpolation or manual interpolation; smoothing and denoising the data by adopting but not limited to an average method, and eliminating random interference in the original data so as to realize data preprocessing;

s2, reducing the calculated amount to accelerate the construction of a data-driven algorithm model, constructing a data-driven algorithm which is simple in Feature extraction and can meet the diagnosis and prediction requirements based on the traditional machine learning, importing the data set obtained after the data processing in the step S1 into an MATLAB Diagnostic Feature Designer, carrying out time domain, frequency domain and time-frequency domain analysis on the data, and extracting features including an average value, a standard deviation, a kurtosis, a skewness and the like; meanwhile, the characteristic values are sorted, selected and fused, the obtained data set is imported into an MATLAB Classification receiver, 24 algorithms including a support vector machine, a decision tree and naive Bayes in a tool box are trained, and a machine learning method with the highest accuracy is selected from the algorithms;

and S3, for faults or performance degradation behaviors which are difficult to extract and large in data quantity, automatically extracting the features and constructing a data-driven algorithm model by utilizing the extremely strong nonlinear fitting capability of deep learning. And (3) constructing different Deep learning networks by using MATLAB Deep Network design, importing preprocessed data for training, comprehensively comparing prediction accuracy of different algorithms and the like, and selecting an optimal algorithm from the preprocessed data for constructing a data-driven algorithm model.

4. FIG. 4 is a diagram of a fusion algorithm implementation method in a model and data fusion layer, which is specifically implemented as follows:

s1, for the condition that equipment fault data or performance degradation data are easy to obtain and mark, constructing a data driving model by using the markable fault data or performance degradation data, obtaining a model-based predicted value through a digital twin model, then realizing the fusion of the model predicted value and the data driving predicted value based on a filtering algorithm, and correcting the predicted value based on the model by using a data driving result as a system observation value;

and S2, injecting corresponding faults into the condition that the equipment fault data or the performance degradation data are difficult to obtain and mark through the fault model, simulating to obtain corresponding fault data and performance degradation data, carrying out data-driven model training by using the data obtained through simulation, transferring the trained model to an actual application scene in a transfer learning mode, and inputting the collected real-time operation data to the transferred model, thereby quickly realizing predictive maintenance.

5. The intelligent service layer in fig. 1 is implemented as follows:

step S1, the real-time visualization of a physical space is realized by using Simulink 3D Animation, the interaction of a virtual world is realized by other hardware devices such as a force feedback control rod and a three-dimensional mouse, the three-dimensional virtual world is viewed in a stereoscopic vision immersion manner, and a real-time virtual model of the physical space is displayed on a computer, a mobile phone, a tablet computer and other mobile devices by using a Simulink 3D Animation Web viewer in a Web browser;

and S2, adopting MATLAB software to realize service encapsulation, and meeting personalized services of different fields and different users in a mode of generating application software, mobile terminal App or webpage. The intelligent predictive maintenance service such as visual intelligent real-time state monitoring, visual online life prediction, remote visual operation guidance and professional maintenance guidance is realized.

The specific embodiments described herein are merely illustrative of the present application and are not intended to be limiting of the present application. All other embodiments obtained by a person skilled in the art without inventive step are within the scope of protection of the present invention.

Claims

1. The predictive maintenance method of the engineering machinery based on the digital twin is characterized by comprising a physical layer, a digital twin model layer, a data perception layer, a model and data fusion layer and an intelligent service layer;

the data perception layer collects, transmits and stores real-time data and historical data generated by the physical layer, and transmits the data to the digital twin model layer in a wired/wireless mode to realize real-time simulation and update of the model; meanwhile, simulation data generated by the digital twin model layer is used for design optimization of the physical layer engineering machinery; simultaneously inputting the data acquired by the data sensing layer and the model simulation data generated by the digital twin model layer into the model and the data fusion layer, and after data processing, using the data to construct a data driving algorithm and a fusion algorithm; and finally, the intelligent service layer realizes the intelligent predictive maintenance service of the engineering machinery through a prediction result and a visualization technology.

2. The predictive maintenance method for construction machinery based on digital twin as claimed in claim 1, wherein the physical layer comprises construction machinery, industrial personal computer and external environment; the industrial personal computer is equipment used for storing historical data in the actual operation process of the engineering machinery; the external environment refers to an external factor in actual operation of the construction machine.

3. The predictive maintenance method for engineering machinery based on digital twinning as claimed in claim 1, wherein said digital twinning model layer comprises multi-domain coupling model, behavior model, rule model, fault model, model accuracy verification and update mechanism.

4. The predictive maintenance method for construction machinery based on digital twinning as claimed in claim 3, wherein the step of establishing the digital twinning model layer is as follows:

s2, establishing a mechanical model, an electrical model, a hydraulic model and a control model in an ideal state by using a multi-field unified modeling language Simscape, and realizing the coupling of the multi-field models by means of an energy interface provided by the Simscape;

s3, considering behaviors of the engineering machinery, such as evolution, performance degradation and the like, generated under the combined action of an external environment and interference and an internal operation mechanism under different time scales, and establishing a behavior model by adopting but not limited to a Markov chain and a finite state machine;

s5, establishing a corresponding fault model generated in the actual operation process of the engineering machinery according to a fault mechanism, and providing a reliable fault data source for the construction of a predictive maintenance algorithm;

and S7, realizing real-time updating of the digital twin model through a model updating mechanism comprising a behavior model, actual operation data and a model simulation data mapping relation.

5. The predictive maintenance method for construction machinery based on digital twin as claimed in claim 1, wherein said data sensing layer comprises sensors, industrial communication protocol, wired/wireless network, data storage.

6. The predictive maintenance method for engineering machinery based on the digital twin as claimed in claim 5, characterized in that the data sensing layer is established by the following steps:

s3, a storage mode combining distributed edge storage and centralized cloud storage is adopted, edge computing is introduced, data for real-time analysis are transmitted to computing equipment closest to a data source through a network for processing and analysis, local storage is realized, the pressure of the throughput of an equipment end and a cloud network is reduced, the data analysis load of cloud center equipment is relieved, and the data analysis response efficiency and the data storage safety are improved; and data which are insensitive to time delay and need to be processed in a centralized manner are transmitted to the cloud center through a special line, and the cloud center can access the edge computing data at any time.

7. The predictive maintenance method for construction machinery based on digital twin as claimed in claim 1, wherein the model and data fusion layer comprises data processing, data driven algorithm, fusion algorithm.

8. The predictive maintenance method for engineering machinery based on the digital twin as claimed in claim 7, characterized in that the model and the data fusion layer are established by the following steps:

s1, performing missing value supplement on sensing data and model simulation data by adopting but not limited to mean interpolation, similar mean interpolation or manual interpolation; smoothing and denoising the data by adopting but not limited to an average method, and eliminating random interference in the original data so as to realize data preprocessing;

s2, in order to reduce the calculated amount and improve the prediction performance, time domain, frequency domain and time-frequency domain analysis are carried out on the data, and characteristics including an average value, a standard deviation, a kurtosis, a skewness and the like are extracted to be used as multi-source input;

s3, accelerating the construction of a data-driven algorithm model, constructing a data-driven algorithm which is simple in fault or performance degradation data feature extraction and can meet diagnosis and prediction requirements by utilizing traditional machine learning based on traditional machine learning;

s4, for fault or performance degradation data with difficult feature extraction and complex data volume, automatically extracting features and constructing a data-driven algorithm model by utilizing the extremely strong nonlinear fitting capability of deep learning;

s5, for the condition that the engineering machinery fault data or performance degradation data are easy to obtain and mark, constructing a data driving model by using the markable fault data or performance degradation data, obtaining a model-based predicted value through a digital twin model, then realizing the fusion of the model predicted value and the data driving predicted value based on a filtering algorithm, and correcting the model-based predicted value by using a data driving result as a system observation value;

and S6, injecting corresponding faults into the engineering machinery fault data or performance degradation data which are difficult to obtain and mark through the fault model, simulating to obtain corresponding fault data and performance degradation data, training the data driving model by using the data obtained through simulation, transferring the trained model to an actual application scene, and inputting the collected real-time operation data to the transferred model, thereby quickly realizing predictive maintenance.

9. The predictive maintenance method for construction machinery based on the digital twin as claimed in claim 1, wherein the intelligent service layer comprises status monitoring, fault pre-warning, life prediction and intelligent maintenance strategy.

10. The predictive maintenance method for construction machinery based on digital twin as claimed in claim 9, wherein the intelligent service layer establishing step is as follows:

step S1, realizing real-time visualization of a physical space by using Simulink 3D Animation, realizing interaction of a virtual world by using other hardware equipment such as a force feedback control rod and a three-dimensional mouse, viewing the three-dimensional virtual world in a stereoscopic vision immersion manner, and displaying a real-time virtual model of the physical space on mobile equipment such as a computer, a mobile phone and a tablet personal computer by using a Simulink 3D Animation Web viewer in a Web browser;

and S2, adopting MATLAB software to realize service encapsulation, meeting personalized services of different fields and different users in a mode of generating application software, mobile terminal App or webpage, and realizing intelligent predictive maintenance services such as visual intelligent real-time state monitoring, visual online life prediction, remote visual operation guidance and professional maintenance guidance.