Digital platform architecture of ship electric power system
By designing a digital platform architecture in the ship power system, integrating multiple databases and simulation platforms, combining big data analysis and three-dimensional visualization, the problem of insufficient evaluation capabilities of traditional monitoring and control systems is solved, and intelligent operation and maintenance and efficient fault diagnosis are achieved. Improves system operation reliability.
Patent Information
- Application Number
- CN202410981676.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-12
AI Technical Summary
The traditional monitoring and control system of existing ship power systems has weak ability to evaluate the system's operating status and predict life, complex system failures cannot be traced and repaired in time, and the development of intelligent operation and maintenance and health management technologies is lagging behind.
A digital platform architecture for ship power system was designed, integrating domestic cache databases, timing databases, relational databases, power system simulation platform and operating system, combining built-in system mechanism model and big data analysis algorithm to realize data acquisition, storage and intelligent transportation Dimensional function, which can display the evaluation and diagnosis results through three-dimensional visualization to improve the system operation reliability.
It realizes the intelligent maintenance function of the ship's power system, improves the reliability of system operation and fault diagnosis capabilities, supports flexible configuration of advanced application algorithms, and realizes digital twin functions and efficient data management.
Smart Images

Figure CN118941235A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a ship power system, and in particular to a digital platform architecture of a ship power system. Background Art
[0002] The ship power system supports the normal operation of various equipment and components of the ship in the form of electrical energy, and is an important part of the ship's operation. At present, the ship power system still uses the traditional monitoring and control system, that is, the unit's operating status monitoring is achieved by judging whether the sensor data exceeds the threshold. Due to the complexity of the power system, there are many state parameters, and the operating range and attributes of each parameter are different. The ability to evaluate the system's operating status and predict its life by using this method is weak, and complex system failures cannot be traced and repaired in time. The research on operation and maintenance guarantees and equipment health diagnosis technologies such as fault diagnosis, state perception, and life assessment is still in its early stages. Advanced technology research such as the digital platform architecture platform for ship power systems for intelligent operation and maintenance has not yet been carried out, which is far from the development requirements of condition-based maintenance and health management. Summary of the invention
[0003] In view of the problem that the traditional monitoring and control system in the current ship power system has weak ability to evaluate the system operation status and predict the life, a digital platform architecture for the ship power system is proposed. By integrating domestic cache database, domestic time series database, domestic relational database, domestic power system simulation platform, domestic operating system and domestic computing platform and other software and hardware as the basic module of the digital platform for the ship power system. In view of the intelligent operation and maintenance functions such as fault diagnosis, state perception, and life evaluation required by the current ship power system, a multi-layer digital platform is constructed to realize the data collection and storage of the ship power system. The built-in system mechanism model, big data analysis algorithm and other intelligent application functions are combined to realize the system status evaluation, fault diagnosis, life prediction and other functions. At the same time, according to the evaluation and diagnosis combined with the expert database, the decision-making suggestions for the corresponding faults and the relationship with the spare parts library are clarified. Finally, the evaluation and diagnosis results are well displayed through three-dimensional visualization to realize the intelligent maintenance function of the ship power system, thereby improving the reliability of the system operation.
[0004] The technical solution of the present invention is:
[0005] A digital platform architecture for a ship power system, the platform as a whole is divided into a network layer, a platform hardware layer, an operating environment layer, a data management layer, a data interface layer, a business layer, a service layer, a unified authentication layer, a front-end interface, and a display terminal; at the same time, the digital platform has the function of managing data, models, algorithms, and communication services, provides an operating environment for model and algorithm applications, and serves as a carrier to realize the centralized operation of digital functions;
[0006] Platform architecture:
[0007] (1) Network layer
[0008] The network layer indicates the communication method and protocol between the acquisition and control layer and the platform. The two use Ethernet for communication and TCP / IP, UDP, MQTT, and Modbus TCP protocols for data exchange.
[0009] (2) Platform hardware layer
[0010] The hardware of the platform hardware layer includes:
[0011] CPU: Feiteng 2000+, 64 cores, 2.2GHz,
[0012] Graphics card: Jingjiawei JM7200,
[0013] Memory: Changxin Storage 128GB DDR4,
[0014] Hard disk: V-KOOL 4T MSATA,
[0015] Operating system: Kylin V10;
[0016] (3) Operational environment layer
[0017] The operating environment of the digital platform includes the operating system, language, and model algorithm dependency framework; the operating system uses Kylin as the system foundation;
[0018] For the implementation of models, standardized algorithms and advanced algorithm applications, the digital platform needs to have corresponding software development tools. The digital twin model simulation tools, standardized algorithms and advanced algorithm application development tools are summarized as follows:
[0019] Electrical simulation software: CloudPSS;
[0020] Thermodynamic simulation software: MWorks;
[0021] Vibration dynamics simulation software: MWorks;
[0022] Algorithm development software: Python;
[0023] Platform software development language environment: C / C++, Python, JAVA;
[0024] (4) Data management layer
[0025] The data management layer is divided into relational database, message queue and time series database according to the different data requirements of the business. The relational database is used to store data with relational models such as device attributes and alarm information; the message queue is used to cache real-time data to meet the business's demand for real-time data; the time series data is used to store sensor data uploaded by the device in real time to improve the business's data utilization efficiency; the relational database and memory database use the domestic database DAMO; the time series database is planned to use the domestic TDengine, DolphinDB, and CeresDB;
[0026] (5) Data interface layer
[0027] The data interface layer opens up the data channel between the data management layer and the business layer, providing the business layer with relevant data services including real-time data and historical data, and can also realize data pre-processing functions;
[0028] (6) Business layer
[0029] The business layer includes software framework technology, model algorithm interface, twins and applications, and management services;
[0030] Software framework technology includes the software framework and technology stack used in the platform development process. Based on the mature technology framework, the stability and convenience of the system can be guaranteed during the software development process.
[0031] The model algorithm interface is used to provide a standard access interface and control method for model algorithms developed by third parties, so as to realize the unified control and management of models and algorithms by the platform;
[0032] Twins and applications realize twins of systems / devices in digital platforms and display advanced applications;
[0033] Management services enable the platform to manage all information about devices, protocols, models, algorithms, and users;
[0034] At the same time, during the operation of the business layer, user authority control is implemented and corresponding log records are made for the steps in the operation process;
[0035] (7) Service Layer
[0036] The service layer uses TCP, Http, Https, WebSocket, and Restful API interfaces to represent the relevant data transmission methods that can be provided when the front-end interface accesses the back-end service;
[0037] (8) Unified Authentication
[0038] The platform adopts B / S architecture, and there is a need for multiple front-ends to access the business layer at the same time. To ensure the security and reliability of the front-end interface's access to the business layer, it is necessary to perform identity authentication on the front-end access and then assign the corresponding permissions to the business;
[0039] (9) Front-end interface
[0040] The front-end interface includes two modes: the WEB end and the large screen. The WEB end is used to implement the management and control operations of all businesses in the digital platform; the large screen is used to display the final display status of the business, so the two use different technical frameworks;
[0041] (10) Display terminal
[0042] The display end refers to the device used to display the front-end interface. Since the platform adopts B / S architecture, the platform can be displayed differently on PCs and mobile terminal devices;
[0043] 2. Platform Functions
[0044] 2.1 Model Management
[0045] (1) Visual model configuration
[0046] The 3D visual model file corresponding to the real device is uploaded to the platform, and data interaction with the real device is achieved by binding with the real device. The 3D model can be decomposed into different important components, each of which can be displayed and analyzed separately, or it can be configured and managed to meet various visual display needs through assembly and splicing. At the same time, combined with the processing of advanced algorithms and models, it can display relevant parameter evaluations of real projects, such as: evaluation of current health status, evaluation of future life, fault evaluation, and treatment opinions.
[0047] (2) Model hierarchical management
[0048] The platform can decompose and manage projects according to the hierarchical classification of systems, equipment, and modules, and bind models to real systems / equipment / modules, thereby achieving a one-to-one correspondence between the underlying interactions and implementation logic of models at different levels of digital projects, and completing the corresponding operation status management and control;
[0049] 2.2 Algorithm Management
[0050] (1) Algorithm configuration
[0051] According to the needs of digitalization, the implementation of the algorithm can be either a locally running thread or a remotely called service. The platform can realize the rapid configuration of the algorithm in the form of configuration according to the level of the system or device corresponding to the algorithm itself, and in accordance with the functional logic and interaction system of the algorithm itself, so as to realize the overall management and overall application of the algorithm.
[0052] (2) Standard Algorithm Tools
[0053] The platform provides standard algorithm tools, forming a typical algorithm library with sample functions; by providing an algorithm framework through application tools, it can quickly realize the classic algorithm solution of the mechanism model and the artificial intelligence algorithm implementation of the data-driven model;
[0054] 2.3 Data Management
[0055] (1) Database management
[0056] The platform manages multi-source heterogeneous data through a database; the relational database is used to store data with relational models such as device attributes and alarm information; the message queue is used to cache real-time data to meet the business's demand for real-time data; and the time series data is used to store sensor data uploaded by the device in real time to improve the business's efficiency in using data.
[0057] (2) Data query management
[0058] The data query function of the platform is divided into historical data and real-time data query, and data export can be realized, which is convenient for offline analysis or real-time preview of the platform or corresponding project data; at the same time, with mature chart components, graphical display of data and standardized report generation can be realized;
[0059] 2.4 Communication Management
[0060] (1) Protocol Management
[0061] The data message communication protocols transmitted by hardware from different equipment manufacturers connected to the digital platform may be different. The platform needs to support the parsing configuration of messages with different communication protocols. At the same time, if the background management configuration of communication protocol parsing messages is not supported, the platform needs to support uploading custom scripts to implement communication message parsing.
[0062] (2) Network management
[0063] The lower-layer networks connected to the digital platform include Ethernet, CAN network, and RS485. Different devices may have different network configurations when transmitting data with the platform. At the same time, in order not to affect the independence of data transmission, the same device may also have different network configurations. Therefore, the platform coordinates network configuration management to achieve access to different network forms.
[0064] 2.5 Other Services
[0065] (1) Safety management
[0066] The platform supports the custom addition, deletion, password management and maintenance of basic user information. By assigning different permissions to different user roles and unlocking different levels of service content, it can achieve secure isolation of different user operations.
[0067] (2) Knowledge base management
[0068] The platform provides structured, easy-to-operate, easy-to-use, and comprehensive knowledge cluster services. It stores solutions to common problems or certain faults in advance and provides auxiliary decision-making suggestions, so that users can perform preventive maintenance and predictive regulation on the results predicted by the digital platform.
[0069] In order to realize the functional design content of the digital platform architecture of the ship power system described above, the implementation plan of important functions is further introduced;
[0070] 3. Model Algorithm Interface
[0071] 3.1 Data Interaction Methods
[0072] The data interaction between the platform and the model and algorithm program is planned to adopt the HTTP interface; the calling method is for the model / algorithm program to provide HTTP services and provide the corresponding interface address for the platform to call. If the model / algorithm program itself does not have this capability, it is encapsulated in other languages supported by the model to ultimately achieve the ability to provide HTTP services;
[0073] The data transmission of the platform is divided into input parameters and output parameters according to the type, each of which has two solutions. In the future, we plan to select the appropriate solution to solve the data interaction problem based on the further development of the software;
[0074] The solution to the problem is:
[0075] (1) The platform transmits all parameters required by the model / algorithm program through JSON packages;
[0076] (2) The platform transmits the basic parameters for starting the model, and calls the http data query interface provided by the platform within the model / algorithm program to obtain the data required for the operation. The address and parameters used to make a query request to the platform are passed to the model / algorithm program when the platform calls the model / algorithm program;
[0077] The solution to the problem of outgoing parameters:
[0078] (1) After the platform initiates an HTTP request to call the model / algorithm program, the model / algorithm program returns the result synchronously in this request after processing is completed, which is also called synchronization;
[0079] (2) After the platform initiates an HTTP request to call the model / algorithm program, there is no need to pay attention to the response result of this request. After the model / algorithm program completes the asynchronous processing, it calls the platform's callback interface to transmit the result, which is also called asynchronous;
[0080] 3.2 Program Call
[0081] The platform deploys the basic environment for model and algorithm operation, configures some specified third-party toolkits, and does not allow modification of the model and algorithm operating environment after configuration. First, the main development language of the platform software is Java. There are two solutions for implementing Java calling model and algorithm programs:
[0082] (1) Package the model and algorithm into a Jar class library and then run it based on the most basic environment support;
[0083] (2) Using the API interface provided by the model and algorithm software / program, the platform makes the call;
[0084] There are two ways of remote calling, namely RMI remote method call and RPC remote procedure call. The design patterns are similar and both require a server and a client. The approach is to first extract the common parts, such as the definition of the calling interface, the implementation of the interface, the client and project definition, the server and project definition, and then implement the specific client and server respectively, which can be adapted and selected according to the further development of the platform software.
[0085] Furthermore, the model and algorithm deployment are as follows:
[0086] The platform has two deployment schemes for models and algorithm programs:
[0087] 4.1 Deployment based on server physical machine:
[0088] Implementation process based on server physical machine deployment:
[0089] The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH remote connection information and label information;
[0090] Administrators upload different types of model / algorithm program scripts on the platform. The platform automatically selects the server to be deployed based on the type of uploaded script. The platform connects to the remote server through SSH and transfers the script to the server. After the transfer is completed, the script is executed.
[0091] The model / algorithm program script must include the dependencies required for installation and complete content for starting services. If an environment conflict occurs, the administrator must manually check the server.
[0092] 4.2 Based on server container deployment:
[0093] Implementation process based on server container deployment:
[0094] The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH remote connection information, label information, and maintaining the Docker API address for the platform to operate through the interface;
[0095] The administrator uploads an image to the platform and can make some simple settings for running the container on the platform, or directly use the docker compose orchestration file to start it. It can run on any server with Docker installed. There is no need to understand environmental issues, only to pay attention to whether there are sufficient resources.
[0096] The beneficial effects of the present invention are:
[0097] (1) Realize the digital twin function of the ship power system;
[0098] (2) Support communication methods such as TCP / IP, UDP, MQTT, and IEC61850;
[0099] (3) Time series data: Query time for one day's data (86,400 data items) of 100 measurement points is ≤5s;
[0100] (4) The platform calls the model and the algorithm result return time (excluding calculation time) is ≤ 2s;
[0101] (5) Realize flexible configuration and call of advanced application algorithms of ship power system;
[0102] (6) Realize the intelligent operation and maintenance function of the ship power system. BRIEF DESCRIPTION OF THE DRAWINGS
[0103] Figure 1 This is a diagram of the digital platform architecture of the ship power system of the present invention;
[0104] Figure 2 This is a functional diagram of the digital platform for the ship power system of the present invention;
[0105] Figure 3 A schematic diagram of the deployment of the algorithm model of the digital platform of the ship power system of the present invention based on a physical machine;
[0106] Figure 4 This is a schematic diagram of server deployment based on the algorithm model of the digital platform of the ship power system of the present invention. DETAILED DESCRIPTION
[0107] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.
[0108] A digital platform architecture for ship power systems. Figure 1 As shown in the figure, the platform is divided into network layer, platform hardware layer, operating environment layer, data management layer, data interface layer, business layer, service layer, unified authentication, front-end interface and display terminal. Figure 2 As shown, the digital platform also has the function of managing data, models, algorithms, communications and other services, provides an operating environment for models, algorithms and other applications, and serves as a carrier to realize the centralized operation of digital functions.
[0109] 1. Platform architecture:
[0110] (1) Network layer
[0111] The network layer mainly indicates the communication method and protocol between the acquisition and control layer and the platform. The two mainly use Ethernet for communication and TCP / IP, UDP, MQTT, Modbus TCP and other protocols for data exchange.
[0112] (2) Platform hardware layer
[0113] The hardware composition of the platform hardware layer is shown in Table 1.
[0114] Table 1 Digital platform hardware composition
[0115]
[0116] (3) Operational environment layer
[0117] The operating environment of the digital platform includes the operating system, language, and model algorithm dependency framework. The operating system uses Kylin as the system foundation.
[0118] For the implementation of models, standardized algorithms and advanced algorithm applications, the digital platform needs to have corresponding software development tools. The digital twin model simulation tools, standardized algorithms and advanced algorithm application development tools are summarized as follows:
[0119] Electrical simulation software: CloudPSS;
[0120] Thermodynamic simulation software: MWorks;
[0121] Vibration dynamics simulation software: MWorks;
[0122] Algorithm development software: Python, etc.
[0123] Platform software development language environment: C / C++, Python, JAVA (Ailbaba Dragonwell).
[0124] (4) Data management layer
[0125] The data management layer is divided into relational database, message queue and time series database according to the different data requirements of the business. The relational database is used to store data with relational models such as device attributes and alarm information; the message queue is used to cache real-time data to meet the business's demand for real-time data; the time series data is used to store sensor data uploaded by the device in real time to improve the business's use of data. The relational database and the in-memory database use the domestic database DAMO; the time series database is planned to use the domestic TDengine, DolphinDB, and CeresDB.
[0126] (5) Data interface layer
[0127] The data interface layer mainly opens up the data channel between the data management layer and the business layer to provide the business layer with relevant data services including real-time data, historical data, etc., and can also realize functions such as data preprocessing.
[0128] (6) Business layer
[0129] The business layer consists of software framework technology, model algorithm interface, twins and applications, and management services.
[0130] Software framework technology includes the software framework and technology stack that are mainly used in the platform development process. Based on the mature technology framework, the stability and convenience of the system can be guaranteed during the software development process.
[0131] The model algorithm interface is used to provide a standard access interface and control method for model algorithms developed by third parties, so as to realize the unified control and management of models and algorithms by the platform.
[0132] Twins and applications realize the twinning of systems / devices in the digital platform and the display of advanced applications.
[0133] Management services enable the platform to manage all information such as devices, protocols, models, algorithms, and users.
[0134] At the same time, during the operation of the business layer, user authority control is implemented and corresponding log records are made for the steps in the operation process.
[0135] (7) Service Layer
[0136] The service layer mainly uses TCP, Http, Https, WebSocket, and Restful API interfaces to represent the relevant data transmission methods that can be provided when the front-end interface accesses the back-end service.
[0137] (8) Unified Authentication
[0138] The platform adopts B / S architecture, and there is a need for multiple front-ends to access the business layer simultaneously. In order to ensure the security and reliability of the front-end interface's access to the business layer, it is necessary to authenticate the front-end access and then assign the corresponding permissions to the business.
[0139] (9) Front-end interface
[0140] The front-end interface mainly includes two modes: WEB and large screen. The WEB is mainly used to realize management and control of all businesses in the digital platform; the large screen is mainly used to display the final display status of the business, so the two use different technical frameworks.
[0141] (10) Display terminal
[0142] The display end mainly refers to the device used to display the front-end interface. Since the platform adopts B / S architecture, the platform can perform differentiated display on PCs, mobile terminals and other devices.
[0143] 2. Platform Functions
[0144] 2.1 Model Management
[0145] (1) Visual model configuration
[0146] The 3D visual model file corresponding to the real device is uploaded to the platform, and data interaction with the real device is achieved by binding with the real device. The 3D model can be decomposed into different important components, each of which can be displayed and analyzed separately, or it can be configured and managed to meet various visual display needs through assembly and splicing. At the same time, combined with advanced application processing such as algorithms and models, relevant parameter evaluations of real projects can be displayed, such as: evaluation of current health status, evaluation of future life, fault evaluation, treatment opinions, etc.
[0147] (2) Model hierarchical management
[0148] The platform can decompose and manage projects according to the hierarchical levels of systems, equipment, and modules, and bind models to real systems / equipment / modules, thereby achieving a one-to-one correspondence between the underlying interactions and implementation logic of models at different levels of digital projects, and completing corresponding operating status management and control.
[0149] 2.2 Algorithm Management
[0150] (1) Algorithm configuration
[0151] According to the needs of digitalization, the algorithm can be implemented as a locally running thread or a remotely called service. The platform can quickly assemble the algorithm in the form of configuration according to the level of the system or device corresponding to the algorithm itself, and in accordance with the functional logic and interaction system of the algorithm itself, thereby achieving the overall management and application of the algorithm.
[0152] (2) Standard Algorithm Tools
[0153] The platform provides standard algorithm tools, forming a typical algorithm library with sample functions. By providing an algorithm framework through application tools, it can quickly realize the classic algorithm solution of the mechanism model and the artificial intelligence algorithm implementation of the data-driven model.
[0154] 2.3 Data Management
[0155] (1) Database management
[0156] The platform manages multi-source heterogeneous data through databases. Relational databases are used to store data with relational models, such as equipment attributes and alarm information; message queues are used to cache real-time data to meet business needs for real-time data; and time series data are used to store sensor data uploaded by devices in real time to improve business efficiency in using data.
[0157] (2) Data query management
[0158] The data query function of the platform is divided into historical data and real-time data query, and data export can be realized, which is convenient for offline analysis or real-time preview of the platform or corresponding project data. At the same time, with mature chart components, graphical display of data and standardized report generation can be realized.
[0159] 2.4 Communication Management
[0160] (1) Protocol Management
[0161] The data message communication protocols transmitted by different equipment manufacturers' hardware connected to the digital platform may be different. The platform needs to support the parsing configuration of different communication protocol messages. At the same time, if the background management configuration of communication protocol parsing messages is not supported, the platform needs to support uploading custom scripts to implement communication message parsing.
[0162] (2) Network management
[0163] The lower-level networks connected to the digital platform include Ethernet, CAN network, RS485, etc. Different devices may have different network configurations when transmitting data with the platform. At the same time, in order not to affect the independence of data transmission, the same device may have different network configurations. Therefore, the platform coordinates network configuration management to achieve access to different network forms.
[0164] 2.5 Other Services
[0165] (1) Safety management
[0166] The platform supports customized addition, deletion, password management and maintenance of basic user information. It divides different permissions for different user roles and unlocks different levels of service content, thereby achieving secure isolation of different user operations.
[0167] (2) Knowledge base management
[0168] The platform provides structured, easy-to-operate, easy-to-use, and comprehensive knowledge cluster services. It stores solutions to general problems or certain faults in advance and provides auxiliary decision-making suggestions, so that users can perform preventive maintenance and predictive regulation on the results predicted by the digital platform.
[0169] In order to realize the functional design content of the digital platform architecture of the ship power system described above, the implementation plan of important functions is further introduced.
[0170] 3. Model Algorithm Interface
[0171] 3.1 Data Interaction Methods
[0172] The data interaction between the platform and the model and algorithm program is mainly planned to use the HTTP interface. The main calling method is for the model / algorithm program to provide HTTP services and provide the corresponding interface address for the platform to call. If the model / algorithm program itself does not have this capability, it is encapsulated in other languages supported by the model to ultimately achieve the ability to provide HTTP services.
[0173] The data transmission of the platform is divided into input parameters and output parameters according to the type, each with two solutions. In the future, we plan to select the appropriate method to solve the data interaction problem based on the further development of the software.
[0174] The solution to the problem is:
[0175] (1) The platform transmits all parameters required by the model / algorithm program through JSON packages;
[0176] (2) The platform transmits the basic parameters for starting the model, and calls the http query data interface provided by the platform within the model / algorithm program to obtain the data required for the operation (the address and parameters used to make query requests to the platform are passed to the model / algorithm program when the platform calls the model / algorithm program).
[0177] The solution to the problem of outgoing parameters:
[0178] (1) After the platform initiates an HTTP request to call the model / algorithm program, the model / algorithm program returns the result synchronously in this request after processing is completed, which is also called synchronization;
[0179] (2) After the platform initiates an HTTP request to call the model / algorithm program, there is no need to pay attention to the response result of this request. After the model / algorithm program completes the asynchronous processing, it calls the platform's callback interface to transmit the result, which is also called asynchronous.
[0180] 3.2 Program Call
[0181] The platform deploys the basic environment for model and algorithm operation, configures some specified third-party toolkits, and does not allow modification of the model and algorithm operating environment after configuration. First, the main development language of the platform software is Java. There are two solutions for implementing Java calling model and algorithm programs:
[0182] (1) Package the model and algorithm into a Jar class library and then run it based on the most basic environment support;
[0183] (2) Use the API interface provided by the model and algorithm software / program to be called by the platform.
[0184] There are two ways to call remotely, namely RMI (Remote Method Invocation) and RPC (Remote Procedure Call). The design patterns are similar and both require a server and a client. The main approach is to first extract the common parts, such as the definition of the calling interface, the implementation of the interface, the client and project definition, the server and project definition, etc., and then implement the specific client and server respectively, which can be selected according to the further development of the platform software.
[0185] 4. Model and Algorithm Deployment
[0186] The platform's deployment schemes for models and algorithm programs are mainly divided into two types:
[0187] 4.1 Deployment based on server physical machine
[0188] The following is an example of deployment based on a physical server: Figure 4 As shown, the main implementation process is:
[0189] (1) The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH (network encryption method) remote connection information, labels and other information;
[0190] (2) The administrator uploads different types of model / algorithm program scripts on the platform. The platform automatically selects the server to be deployed based on the type of uploaded script. The platform obtains a connection with the remote server through SSH and transfers the script to the server. After the transfer is completed, the script is executed.
[0191] Note: The model / algorithm program script must include complete content such as the dependencies required for installation and startup services. If you encounter problems such as environment conflicts, the administrator needs to manually check the server.
[0192] 4.2 Server container deployment
[0193] The main implementation process based on server container deployment:
[0194] (1) The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH remote connection information, labels and other information, and maintaining the Docker API address for the platform to operate through the interface.
[0195] (2) The administrator uploads an image to the platform and can make some simple settings for running the container on the platform, or directly use the docker compose orchestration file to start it. It can run on any server with Docker installed. There is no need to understand environmental issues, only to pay attention to whether there are sufficient resources.
[0196] The above-mentioned embodiment only expresses one implementation mode of the present invention, and its description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached claims.
Claims
1. A digital platform architecture for ship power systems, It is characterized in that The platform as a whole is divided into network layer, platform hardware layer, operating environment layer, data management layer, data interface layer, business layer, service layer, unified authentication, front-end interface and display end; at the same time, the digital platform has the function of managing data, models, algorithms and communication services, providing the operating environment for model and algorithm application, and serving as a carrier to realize the centralized operation of digital functions; Platform architecture: (1) Network layer The network layer indicates the communication method and protocol between the acquisition and control layer and the platform. The two use Ethernet for communication and TCP / IP, UDP, MQTT, and Modbus TCP protocols for data exchange. (2) Platform hardware layer The hardware of the platform hardware layer includes: CPU: Feiteng 2000+, 64 cores, 2.2GHz, Graphics card: Jingjiawei JM7200, Memory: Changxin Storage 128GB DDR4, Hard disk: V-KOOL 4T MSATA, Operating system: Kylin V10; (3) Operational environment layer The operating environment of the digital platform includes the operating system, language, and model algorithm dependency framework; the operating system uses Kylin as the system foundation; For the implementation of models, standardized algorithms and advanced algorithm applications, the digital platform needs to have corresponding software development tools. The digital twin model simulation tools, standardized algorithms and advanced algorithm application development tools are summarized as follows: Electrical simulation software: CloudPSS; Thermodynamic simulation software: MWorks; Vibration dynamics simulation software: MWorks; Algorithm development software: Python; Platform software development language environment: C / C++, Python, JAVA; (4) Data management layer The data management layer is divided into relational database, message queue and time series database according to the different data requirements of the business. The relational database is used to store data with relational models such as device attributes and alarm information; the message queue is used to cache real-time data to meet the business's demand for real-time data; the time series data is used to store sensor data uploaded by the device in real time to improve the business's data utilization efficiency; the relational database and memory database use the domestic database DAMO; the time series database is planned to use the domestic TDengine, DolphinDB, and CeresDB; (5) Data interface layer The data interface layer opens up the data channel between the data management layer and the business layer, providing the business layer with relevant data services including real-time data and historical data, and can also realize data pre-processing functions; (6) Business layer The business layer includes software framework technology, model algorithm interface, twins and applications, and management services; Software framework technology includes the software framework and technology stack used in the platform development process. Based on the mature technology framework, the stability and convenience of the system can be guaranteed during the software development process. The model algorithm interface is used to provide a standard access interface and control method for model algorithms developed by third parties, so as to realize the unified control and management of models and algorithms by the platform; Twins and applications realize twins of systems / devices in digital platforms and display advanced applications; Management services enable the platform to manage all information about devices, protocols, models, algorithms, and users; At the same time, during the operation of the business layer, user authority control is implemented and corresponding log records are made for the steps in the operation process; (7) Service Layer The service layer uses TCP, Http, Https, WebSocket, and Restful API interfaces to represent the relevant data transmission methods that can be provided when the front-end interface accesses the back-end service; (8) Unified Authentication The platform adopts B / S architecture, and there is a need for multiple front-ends to access the business layer at the same time. To ensure the security and reliability of the front-end interface's access to the business layer, it is necessary to perform identity authentication on the front-end access and then assign the corresponding permissions to the business; (9) Front-end interface The front-end interface includes two modes: the WEB end and the large screen. The WEB end is used to implement the management and control operations of all businesses in the digital platform; the large screen is used to display the final display status of the business, so the two use different technical frameworks; (10) Display terminal The display end refers to the device used to display the front-end interface. Since the platform adopts B / S architecture, the platform can be displayed differently on PCs and mobile terminal devices; 2. Platform Functions 2.1 Model Management (1) Visual model configuration The 3D visual model file corresponding to the real device is uploaded to the platform, and data interaction with the real device is achieved by binding with the real device. The 3D model can be decomposed into different important components, each of which can be displayed and analyzed separately, or it can be configured and managed to meet various visual display needs through assembly and splicing. At the same time, combined with the processing of advanced algorithms and models, it can display relevant parameter evaluations of real projects, such as: evaluation of current health status, evaluation of future life, fault evaluation, and treatment opinions. (2) Model hierarchical management The platform can decompose and manage projects according to the hierarchical classification of systems, equipment, and modules, and bind models to real systems / equipment / modules, thereby achieving a one-to-one correspondence between the underlying interactions and implementation logic of models at different levels of digital projects, and completing the corresponding operation status management and control; 2.2 Algorithm Management (1) Algorithm configuration According to the needs of digitalization, the implementation of the algorithm can be either a locally running thread or a remotely called service. The platform can realize the rapid configuration of the algorithm in the form of configuration according to the level of the system or device corresponding to the algorithm itself, and in accordance with the functional logic and interaction system of the algorithm itself, so as to realize the overall management and overall application of the algorithm. (2) Standard Algorithm Tools The platform provides standard algorithm tools, forming a typical algorithm library with sample functions; by providing an algorithm framework through application tools, it can quickly realize the classic algorithm solution of the mechanism model and the artificial intelligence algorithm implementation of the data-driven model; 2.3 Data Management (1) Database management The platform manages multi-source heterogeneous data through a database; the relational database is used to store data with relational models such as device attributes and alarm information; the message queue is used to cache real-time data to meet the business's demand for real-time data; and the time series data is used to store sensor data uploaded by the device in real time to improve the business's efficiency in using data. (2) Data query management The data query function of the platform is divided into historical data and real-time data query, and data export can be realized, which is convenient for offline analysis or real-time preview of the platform or corresponding project data; at the same time, with mature chart components, graphical display of data and standardized report generation can be realized; 2.4 Communication Management (1) Protocol Management The data message communication protocols transmitted by hardware from different equipment manufacturers connected to the digital platform may be different. The platform needs to support the parsing configuration of messages with different communication protocols. At the same time, if the background management configuration of communication protocol parsing messages is not supported, the platform needs to support uploading custom scripts to implement communication message parsing. (2) Network management The lower-layer networks connected to the digital platform include Ethernet, CAN network, and RS485. Different devices may have different network configurations when transmitting data with the platform. At the same time, in order not to affect the independence of data transmission, the same device may also have different network configurations. Therefore, the platform coordinates network configuration management to achieve access to different network forms. 2.5 Other Services (1) Safety management The platform supports the custom addition, deletion, password management and maintenance of basic user information. By assigning different permissions to different user roles and unlocking different levels of service content, it can achieve secure isolation of different user operations. (2) Knowledge base management The platform provides structured, easy-to-operate, easy-to-use, and comprehensive knowledge cluster services. It stores solutions to common problems or certain faults in advance and provides auxiliary decision-making suggestions, so that users can perform preventive maintenance and predictive regulation on the results predicted by the digital platform. In order to realize the functional design content of the digital platform architecture of the ship power system described above, the implementation plan of important functions is further introduced; 3. Model Algorithm Interface 3.1 Data Interaction Methods The data interaction between the platform and the model and algorithm program is planned to adopt the HTTP interface; the calling method is for the model / algorithm program to provide HTTP services and provide the corresponding interface address for the platform to call. If the model / algorithm program itself does not have this capability, it is encapsulated in other languages supported by the model to ultimately achieve the ability to provide HTTP services; The data transmission of the platform is divided into input parameters and output parameters according to the type, each of which has two solutions. In the future, we plan to select the appropriate solution to solve the data interaction problem based on the further development of the software; The solution to the problem is: (1) The platform transmits all parameters required by the model / algorithm program through JSON packages; (2) The platform transmits the basic parameters for starting the model, and calls the http data query interface provided by the platform within the model / algorithm program to obtain the data required for the operation. The address and parameters used to make a query request to the platform are passed to the model / algorithm program when the platform calls the model / algorithm program; The solution to the problem of outgoing parameters: (1) After the platform initiates an HTTP request to call the model / algorithm program, the model / algorithm program returns the result synchronously in this request after processing is completed, which is also called synchronization; (2) After the platform initiates an HTTP request to call the model / algorithm program, there is no need to pay attention to the response result of this request. After the model / algorithm program completes the asynchronous processing, it calls the platform's callback interface to transmit the result, which is also called asynchronous; 3.2 Program Call The platform deploys the basic environment for the model and algorithm operation, and configures some specified third-party toolkits. After the configuration is completed, the operating environment of the model and algorithm is not allowed to be modified. First of all, the main development language of the platform software is Java. There are two solutions for implementing Java to call model and algorithm programs: (1) Package the model and algorithm into a Jar class library and then run it based on the most basic environment support; (2) Using the API interface provided by the model and algorithm software / program, the platform makes the call; There are two ways of remote calling, namely RMI remote method call and RPC remote procedure call. The design patterns are similar and both require a server and a client. The approach is to first extract the common parts, such as the definition of the calling interface, the implementation of the interface, the client and project definition, the server and project definition, and then implement the specific client and server respectively, which can be adapted and selected according to the further development of the platform software.
2. According to claim 1, the digital platform architecture of the ship power system, It is characterized in that The model and algorithm deployment are as follows: The platform has two deployment schemes for models and algorithm programs: 4.1 Deployment based on server physical machine: Implementation process based on server physical machine deployment: The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH remote connection information and label information; Administrators upload different types of model / algorithm program scripts on the platform. The platform automatically selects the server to be deployed based on the type of uploaded script. The platform connects to the remote server through SSH and transfers the script to the server. After the transfer is completed, the script is executed. in, The model / algorithm program script must contain the dependencies required for installation and complete content for starting services. If an environment conflict occurs, the administrator must manually check the server. 4.2 Based on server container deployment: Implementation process based on server container deployment: The platform needs to provide the function of maintaining the server, that is, maintaining the server's SSH remote connection information, label information, and maintaining the Docker API address for the platform to operate through the interface; The administrator uploads an image to the platform and can make some simple settings for running the container on the platform, or directly use the docker compose orchestration file to start it. It can run on any server with Docker installed. There is no need to understand environmental issues, only to pay attention to whether there are sufficient resources.
Citation Information
Cited By
Data processing system of intelligent fusion terminal data center
CN119669328A
Digital base model design system for shipbuilding digital transformation
CN121144284A
Data access method based on digital platform equipment
CN121388046A
