Intelligent manufacturing internet of things platform based on in-memory relational database and implementation method thereof

Abstract

This invention provides an intelligent manufacturing IoT platform and its implementation method based on an in-memory relational database, relating to the field of industrial IoT technology. It includes an industrial gateway, an in-memory relational database, dual-link independent channels, and an application terminal. The industrial gateway has a bus protocol database adapter, one end of which connects to the device, and the other end connects to the in-memory relational database via the bus protocol database adapter. The in-memory relational database is configured with a fusion object model, which integrates data from multiple monitoring points of the same device into a data table based on row attribute correlation features. The in-memory relational database also provides a standard open interface to connect to the application terminal, transmitting device data to the application terminal. The in-memory relational database pushes the changes in the collected device data to a time-series database, achieving real-time data synchronization. This invention realizes real-time data processing, open access, and correlation analysis, improving the effectiveness of decision-making and response.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT) technology, specifically to an intelligent manufacturing IoT platform based on an in-memory relational database and its implementation method. Background Technology

[0002] With the rapid development of Internet of Things (IoT) technology, its applications in smart homes, industrial manufacturing, and other fields are becoming increasingly widespread. In the industrial IoT scenario, the time-series data generated by devices has the core characteristics of "high-frequency acquisition, massive storage, real-time analysis, and low update rate," which places stringent requirements on the performance, cost, and stability of data processing. Currently, mainstream IoT platforms typically adopt a combined architecture of "time-series database + Redis," where Redis serves as an in-memory database for caching real-time data, and the time-series database is used for storing historical data. However, this architecture has significant limitations when facing the millisecond-level response requirements of industrial production lines: First, data storage is incomplete, typically using a minute-level mechanism to refresh Redis data to the time-series database, resulting in non-real-time decision-making and easy data loss during downtime; second, data openness is insufficient, as Redis, based on key-value pair storage, cannot explicitly represent data correlations, requiring customized interface services and increasing development costs; third, data correlation is weak, as traditional point-based models cannot effectively describe the correlation between multiple points on the same device, limiting the accuracy of real-time analysis; finally, protocol compatibility is poor, with industrial fieldbuses primarily using ProfiNet / EtherCat / EtherNet, while traditional database clients mostly use TCP / IP protocols, affecting communication security and latency. Therefore, there is an urgent need for a data processing platform that can meet the "three highs" (high concurrency, high real-time performance, and high reliability) requirements of the Industrial Internet of Things (IIoT) and possesses good openness and correlation. Summary of the Invention

[0003] To address the issues of poor real-time data performance, insufficient openness, and weak data correlation in existing industrial IoT platforms, this application introduces an in-memory relational database to replace the traditional combination of Redis and time-series databases, constructing an intelligent manufacturing IoT platform based on an in-memory relational database, thereby achieving real-time data processing, open access, and correlation analysis.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A smart manufacturing IoT platform based on an in-memory relational database includes an industrial gateway, an in-memory relational database, an application terminal, and a time-series database. The industrial gateway has a built-in bus protocol database adapter; one end of the industrial gateway connects to the device, and the other end connects to the in-memory relational database via the bus protocol database adapter. The in-memory relational database has a fusion object model, which integrates data from multiple monitoring points of the same device into a data table based on row attribute correlation features. The in-memory relational database also provides a standard open interface to connect to the application terminal, transmitting device data to the application terminal. The in-memory relational database pushes changes in the collected device data to the time-series database to achieve real-time data synchronization between the in-memory relational database and the time-series database.

[0005] The above solution introduces an in-memory relational database as a bridge between the industrial gateway and the application end. By leveraging its row attribute correlation characteristics, it integrates data from multiple device locations. This not only meets the real-time requirements of data processing in industrial scenarios but also achieves data openness through standard open interfaces. Furthermore, it ensures data synchronization with the time-series database through a change subscription mechanism, effectively solving the problems of incomplete, closed, and uncorrelated data in traditional architectures.

[0006] Preferably, the bus protocol database adapter supports ProfiNet, EtherCat, and EtherNet industrial bus protocols, and the industrial gateway communicates with the memory relational database through ProfiNet, EtherCat, or EtherNet.

[0007] Preferably, the memory relational database pushes the changes in the collected device data to the time-series database through a built-in change subscription module.

[0008] Preferably, the standard open interface is the SQL standard interface, through which the application queries device data in the memory relational database.

[0009] Preferably, in the data table, each row represents the collected data of a device identifier, and the collected data includes the update time sequence and the update values ​​of all points.

[0010] Preferably, the smart manufacturing IoT platform is equipped with dual independent channels, which include a data channel and a command channel. The data channel is used for the uplink of device data and the downlink of data analysis results from the application end, while the command channel is used to transmit control commands from the application end to the device and configuration commands from the industrial gateway.

[0011] Furthermore, this invention also provides a method for implementing an intelligent manufacturing IoT platform based on an in-memory relational database, comprising the following steps: collecting data from multiple monitoring points of the equipment, and transmitting the collected multi-dimensional equipment time-series data to an in-memory relational database; integrating and processing the multi-dimensional equipment time-series data corresponding to the same equipment identifier based on row attribute correlation features, and storing the integrated data in a structured manner in a preset data table; actively pushing the structured and integrated equipment data to the application terminal for the application terminal to retrieve and use; incrementally identifying the continuously collected equipment data, extracting the changes in the equipment data in real time, and incrementally pushing the data to the time-series database, thereby achieving full real-time synchronization of equipment data between the in-memory relational database and the time-series database.

[0012] As a preferred embodiment, the method of actively pushing the structured and integrated device data to the application terminal is as follows: using the SQL standard interface as the standard open interface, the application terminal performs query operations on the device data in the memory relational database based on the SQL standard interface.

[0013] As a preferred embodiment, the method for incremental identification of continuously collected device data is as follows: the amount of change in device data is identified in real time through the change subscription module built into the memory relational database.

[0014] Beneficial effects: This invention replaces the traditional Redis with an in-memory relational database, leveraging its relational database characteristics to achieve structured data storage and relational management. Specifically, based on row attribute correlation characteristics, data from multiple monitoring points of the same device are integrated into a single data table, making data correlation explicit. This simplifies the complexity of application-layer analysis and decision-making, and improves the effectiveness of decision response. Simultaneously, a built-in change subscription module enables real-time data synchronization from the in-memory relational database to the time-series database, ensuring both fast access to real-time data and complete storage of historical data. Furthermore, by supporting an adapter for industrial bus protocols, direct communication between the industrial gateway and the database is achieved, reducing communication latency and improving system security and stability. Finally, through the decoupling design of data and command channels, uplink and downlink data are separated, effectively avoiding transmission channel congestion issues. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the system architecture of the intelligent manufacturing IoT platform based on an in-memory relational database provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of an IoT platform in existing technology; Figure 3 This is a comparative schematic diagram of the fusion model provided in the embodiments of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terms “first,” “second,” and similar terms used herein do not indicate any order, quantity, or importance, but are merely used to distinguish different components.

[0018] Example 1: like Figure 1 As shown, this embodiment provides an intelligent manufacturing IoT platform based on an in-memory relational database. The platform mainly comprises four core components: an industrial gateway, an in-memory relational database, an application terminal, and a time-series database. These four components together construct a closed-loop system for real-time data flow, processing, and storage, with the in-memory relational database occupying a central hub position, unlike existing technologies where Redis is merely used as a cache (e.g.,...). Figure 2 As shown in the figure, it undertakes the key functions of real-time data processing and correlation analysis.

[0019] The industrial gateway is equipped with a bus protocol database adapter. One end of the industrial gateway connects to the device, and the other end connects to an in-memory relational database through the bus protocol database adapter. Specifically, the industrial gateway acts as a bridge between the physical and digital worlds, responsible for collecting raw data from various industrial devices (such as CNC machine tools, sensors, PLC controllers, etc.). The presence of the bus protocol database adapter allows the industrial gateway to directly "speak" the language of the industrial bus, connecting directly to the database without going through a complex protocol conversion gateway, thereby significantly reducing latency on the communication link. In one specific implementation, the industrial gateway establishes a long-term connection with the in-memory relational database through this adapter, ensuring that data can be written to the database at millisecond-level speeds.

[0020] The in-memory relational database described herein employs a fusion model. This model integrates data from multiple monitoring points of the same device into a data table based on row attribute correlation characteristics, achieving unified management of data from different monitoring points of the same device and minimizing data volume through an update mechanism. This is one of the core improvements of this embodiment. Traditional Redis databases typically store data in key-value pairs, lacking explicit relationships between data points. This leads to multiple requests and secondary reassembly during complex queries by the application. The in-memory relational database in this embodiment not only possesses the high-speed read / write characteristics of in-memory databases but, more importantly, introduces "relational" characteristics. Through the fusion model, it integrates multiple monitoring points of the same device (e.g., temperature, speed, and vibration frequency of a motor) into the same row of the same data table. This row attribute correlation-based storage method gives the data a natural correlation at the physical storage level, greatly simplifying subsequent data analysis logic. It should be understood that although this embodiment focuses on the role of the in-memory relational database as a real-time processing center, it does not exclude persistent storage functionality; its core advantage lies in the rapid integration and response to real-time data.

[0021] The in-memory relational database also provides a standard open interface for connecting to applications, transmitting device data to the application. Specifically, this standard open interface typically refers to the SQL standard interface. Compared to traditional Redis, which requires custom-developed interfaces, the universality of the SQL interface allows applications (such as MES systems, ERP systems, and large visualization dashboards) to directly query data in the in-memory relational database using standard SQL statements. This not only lowers the development threshold but also achieves true data openness, enabling application software from different vendors to seamlessly access the platform and obtain data.

[0022] The in-memory relational database pushes changes in collected device data to the time-series database, enabling real-time data synchronization between the two databases. Within the overall data flow framework, the in-memory relational database primarily handles "hot data" (i.e., real-time data), meeting the needs of real-time monitoring and control; while the time-series database stores "cold data" (i.e., historical data), meeting the needs of long-term storage and trend analysis. This embodiment replaces the outdated "minute-level refresh" mechanism of traditional architectures by having the in-memory relational database actively push changes to the time-series database. This mechanism ensures that even in the event of system failure or downtime, the time-series database retains the latest data state, preventing data loss and achieving data integrity and consistency.

[0023] Through the above architectural design, this embodiment constructs a high-efficiency data flow consisting of "devices, gateways, in-memory relational databases, and application terminals." The in-memory relational database is no longer a simple data transfer station, but a core engine integrating data integration, real-time analysis, and open services, effectively solving the technical pain points of traditional IoT platforms such as data lack of correlation, openness, and real-time updates.

[0024] Example 2: This embodiment, based on Embodiment 1, provides a detailed description of the data organization model and its open interfaces within the in-memory relational database. For example... Figure 3 As shown, Figure 3 The left side shows a traditional point-based object model in existing technology, while the right side shows a fused object model provided by an embodiment of the present invention. The two differ fundamentally in their data storage logic, which directly determines the platform's data processing efficiency and openness.

[0025] In traditional point-based data models, data is typically managed columnarly and stored based on key-value pairs. Specifically, if a device has four monitoring points, such as temperature, pressure, speed, and voltage, the traditional model would split them into four independent data items for management. In this model, the data is discrete, and the correlation between points cannot be explicitly shown. If an application needs to obtain the overall operating status of a device, it must initiate multiple query requests, followed by secondary assembly and association at the application layer. This not only increases network overhead but also significantly limits the efficiency of real-time analysis. Furthermore, because key-value databases like Redis lack standard relational interfaces, applications must rely on custom-developed dedicated interface services to access data, leading to system closure and increasing the cost of subsequent integration and maintenance.

[0026] To address the aforementioned shortcomings, this embodiment employs a fusion model based on row attribute correlation characteristics. In the data table, each row represents the collected data for a single device identifier, including update timing and updated values ​​for all points. Specifically, the in-memory relational database, through the fusion model, integrates multiple monitoring points belonging to the same physical device (such as the aforementioned temperature, pressure, speed, and voltage) into the same row of the same data table. This storage structure leverages the "row" characteristic of relational databases, establishing a natural connection between data at the physical storage level. When device data is updated, the in-memory relational database directly updates the corresponding fields of that row, achieving integrated management of multi-dimensional data. This design makes data correlation explicit, eliminating the need for complex secondary correlation operations on the application side to directly obtain complete device status information, thereby significantly improving the accuracy and response speed of real-time analysis.

[0027] Based on the aforementioned row-based storage structure, this embodiment further enhances data openness. The standard open interface is a standard SQL interface, through which the application queries device data in the in-memory relational database. Since the data is stored in a standardized manner according to the relational table structure, the application can directly access the in-memory relational database using common SQL statements (such as SELECT, WHERE, etc.). For example, the application can construct the statement "SELECT * FROM DeviceTable WHERED deviceID='001'" to query all real-time parameters of the device at once. This process requires no customization of any dedicated API interface; any software or tool supporting the SQL standard can be seamlessly integrated. This open architecture significantly lowers the barrier to system integration, solves the problems of data silos and high degree of interface customization in traditional IoT platforms, and provides a solid technical foundation for interconnectivity in industrial settings.

[0028] Example 3: This embodiment, based on the above embodiments, provides a detailed explanation of the communication protocol adaptation mechanism between the industrial gateway and the in-memory relational database, and the synchronization update mechanism between the in-memory relational database and the time-series database. These improvements collectively address the technical pain points of poor real-time communication, easy congestion of transmission channels, and incomplete data storage of historical data in industrial scenarios.

[0029] Firstly, addressing the complex communication environments of industrial sites, the bus protocol database adapter supports ProfiNet, EtherCat, and EtherNet industrial bus protocols. Industrial gateways communicate with the in-memory relational database via ProfiNet, EtherCat, or EtherNet. In traditional IoT architectures, database clients typically use the common TCP / IP protocol for communication. While acceptable in consumer applications, the processing latency and non-deterministic data transmission mechanisms of the TCP / IP protocol stack often fail to meet millisecond-level control requirements in industrial production lines and are susceptible to electromagnetic interference. Specifically, the bus protocol database adapter in this embodiment incorporates a parsing engine for industrial bus protocols, enabling direct identification and processing of data packets from ProfiNet, EtherCat, or EtherNet protocols. For example, in high-speed motion control scenarios, motor encoder data acquired by the industrial gateway is transmitted via the EtherCat bus; the adapter can directly write the data into the in-memory relational database without complex protocol conversion. This "straight-through" communication method not only significantly reduces latency in the communication link, but also utilizes the deterministic mechanism of the industrial bus protocol itself to ensure the real-time performance and stability of data transmission, thereby meeting the stringent requirements of intelligent manufacturing for high real-time performance.

[0030] Secondly, to resolve the conflict between high-concurrency data transmission and real-time control commands, the intelligent manufacturing IoT platform is equipped with dual-link independent channels, including a data channel and a command channel. The data channel is used for seamless integration of device-side data uplink and application-side data analysis, while the command channel is used to transmit application-side control commands to the device and configuration commands for the industrial gateway. In existing technologies, data transmission and command issuance often share the same physical or logical channel. When the device performs high-frequency data acquisition, the massive amount of time-series data consumes a large amount of bandwidth, causing urgent control commands to queue and wait, resulting in unacceptable delays or even safety incidents. This embodiment decouples the two, constructing a physically or logically isolated dual-channel mechanism. The data channel is dedicated to carrying high-volume uplink transmission of device status data and application-side data analysis; while the command channel is designed as a high-priority, low-latency channel, dedicated to transmitting control commands. It should be understood that this separation design ensures that even when the data channel is running at full capacity due to massive data transmission, the command channel remains unobstructed, enabling control commands issued by the application to reach the industrial gateway or execution device in milliseconds. This effectively avoids the problem of transmission channel congestion and greatly improves the system's security and response speed.

[0031] Finally, to ensure the consistency and integrity of real-time and historical data, the in-memory relational database pushes changes in collected device data to the time-series database through a built-in change subscription module. Traditional IoT platforms typically use scheduled tasks (e.g., once per minute) to batch refresh data from the in-memory database to the time-series database. This mechanism has significant drawbacks: if the system crashes during the interval between two refreshes, the data in memory will be permanently lost; and the minute-level latency cannot meet the accuracy requirements for post-event traceability. In this embodiment, the change subscription module uses an incremental recognition mechanism to monitor changes in data in the in-memory relational database in real time. Specifically, when the industrial gateway updates a row in a data table, the change subscription module immediately captures this change, extracts the amount of changed data, and actively pushes it to the time-series database for storage. This mechanism achieves full real-time synchronization between the in-memory relational database and the time-series database, leveraging the real-time processing advantages of the in-memory relational database while using the time-series database to compensate for the volatility of the in-memory database, ensuring the integrity and traceability of data storage.

[0032] Example 4: This embodiment provides a method for implementing an intelligent manufacturing IoT platform based on an in-memory relational database. This method is based on the platform architecture described in the foregoing embodiment and focuses on describing the entire time sequence of data acquisition, processing and application.

[0033] First, step S100 is executed to collect data from multiple monitoring points on the device and transmit the collected multi-dimensional device time-series data to an in-memory relational database. Specifically, the industrial gateway, as the front-end device for data acquisition, establishes a communication connection with various industrial devices (such as CNC machine tools, sensors, PLCs, etc.) through its built-in bus protocol adapter. The gateway reads the operating parameters of multiple monitoring points on the device in real time, such as the temperature, vibration frequency, and speed of a motor. These parameters constitute multi-dimensional device time-series data. Thanks to the high real-time characteristics of the industrial bus protocol, data can be collected and directly written to the in-memory relational database at millisecond speeds, avoiding the latency caused by traditional TCP / IP protocol conversion and ensuring the real-time performance and integrity of the data source.

[0034] Subsequently, step S200 is executed, where multi-dimensional device time-series data corresponding to the same device identifier are integrated based on row attribute correlation characteristics, and the integrated data is structured and stored in a preset data table. This is the core step of the method in this embodiment. In traditional methods, collected data such as temperature and pressure are often stored discretely as independent key-value pairs, lacking inherent correlation. In this embodiment, however, the in-memory relational database identifies the data based on the device identifier (such as device ID). For multiple monitoring point data belonging to the same device, the system utilizes row attribute correlation characteristics to integrate them into the same row record of the same data table. For example, when motor data with device ID "Device_001" arrives, the system finds the row corresponding to that device in the data table, fills the temperature value into the "Temperature" column, and the speed value into the "Speed" column. This integrated processing method establishes explicit correlations at the data storage level, making the data structure more consistent with business logic and providing a structured foundation for subsequent real-time analysis.

[0035] Next, step S300 is executed, actively pushing the structured and integrated device data to the application for retrieval and use. After the data is integrated and stored, the in-memory relational database exposes its data capabilities to the application through a standard open interface (such as an SQL interface). The application can construct standard SQL queries according to business needs to directly retrieve the complete status information of a device from the in-memory relational database. Since the data has been structured and integrated, the application does not need to perform complex secondary assembly or multiple requests to directly obtain highly correlated datasets. This proactive push or on-demand call mechanism greatly simplifies the development logic of the application layer and improves data response speed.

[0036] Finally, step S400 is executed to incrementally identify the continuously collected device data, extract the changes in the device data in real time, and incrementally push the data to the time-series database, achieving full real-time synchronization of device data between the in-memory relational database and the time-series database. During continuous data acquisition, the built-in change subscription module of the in-memory relational database monitors data changes in the data tables in real time. When a field value in a record is updated, the module immediately identifies the change and extracts the amount of data changed. Unlike traditional timed batch refresh mechanisms, this embodiment uses incremental identification and real-time push; once a change is identified, it is immediately synchronized to the time-series database. This full real-time synchronization method ensures a high degree of consistency between historical data stored in the time-series database and real-time data in memory. Even in the event of system failure or downtime, the latest data is preserved to the greatest extent possible, effectively solving the data loss problem caused by refresh intervals in traditional architectures and ensuring the integrity and traceability of industrial data assets.

[0037] Example 5: Based on Example 4, this embodiment provides a detailed explanation of the data interaction method on the application side and the specific implementation logic of data increment identification, aiming to further demonstrate the executability and superiority of the technical solution of the present invention.

[0038] Firstly, regarding how the application side can efficiently acquire data, the method of actively pushing the structured and integrated device data to the application side is as follows: using the SQL standard interface as the standard open interface, the application side performs query operations on the device data in the in-memory relational database based on the SQL standard interface. Specifically, since the data in the in-memory relational database has been structured and stored according to the fusion object model described in Embodiment 2, that is, multiple points of data of the same device are located in the same row of the same table, when the application side needs to obtain the device status, it does not need to call customized API interfaces to make multiple requests and assemble data as in the traditional architecture. The application side can directly construct a standard SQL query statement, such as "SELECT * FROMDevice_TableWHEREDevice_ID='Target_Device'", and send it to the in-memory relational database through the SQL standard interface. After parsing the statement, the in-memory relational database can quickly locate the target row and return complete device status data. This interaction method based on standard SQL not only greatly reduces the development difficulty of the application side, but also allows any third-party software or analysis tool that supports the SQL standard to seamlessly access the platform, truly realizing open data sharing.

[0039] Secondly, regarding how to achieve efficient data synchronization, the method for incremental identification of continuously collected device data is as follows: The change subscription module built into the in-memory relational database identifies changes in device data in real time. Specifically, the change subscription module uses an incremental identification mechanism to monitor data changes. In a preferred embodiment, the change subscription module continuously monitors the data update log in the in-memory relational database. When new data is written by the industrial gateway, causing an update to a row in a data table, the module automatically captures this update event. For example, by comparing the data frames before and after the update, or by identifying the update timestamp field in the data record, the changed field and its new value are extracted. Subsequently, the module only encapsulates and pushes this change to the time-series database, rather than transferring the entire table or row of data. This logic ensures the real-time nature of data synchronization, avoids the data delay and loss risks associated with traditional timed batch refresh mechanisms, and also significantly reduces network transmission load and the write pressure on the time-series database.

[0040] Example 6: To more intuitively demonstrate the application effect of the technical solution of this invention in a real industrial scenario, this embodiment takes a manufacturing production line as an example for detailed explanation. This production line mainly includes key processing equipment such as CNC machine tools and robotic arms, and has extremely high requirements for the real-time performance of data acquisition, the certainty of command issuance, and the integrity of historical data.

[0041] In this application scenario, the industrial gateway, acting as the front-end device for data acquisition, is deployed next to the CNC machine tool. The industrial gateway has a built-in bus protocol database adapter that supports the EtherCat industrial bus protocol. Temperature and vibration sensors are installed on the CNC machine tool's spindle to monitor its operating status in real time. When the machine tool is running, the industrial gateway collects multi-dimensional timing data such as spindle temperature and vibration frequency at millisecond-level frequencies via the EtherCat bus. Compared to the traditional TCP / IP protocol, the EtherCat protocol guarantees deterministic latency in data transmission, avoiding data delays caused by network congestion and meeting the stringent real-time requirements of industrial environments.

[0042] The collected data is directly written to the in-memory relational database via the bus protocol database adapter. Based on a built-in fusion model, the in-memory relational database utilizes row attribute correlation characteristics to integrate monitoring data from the same CNC machine tool, such as spindle temperature and vibration frequency, into the same row of the same data table. For example, when the machine tool ID is "CNC_001", its corresponding row record contains fields such as "temperature", "vibration frequency", and "speed". This storage method creates a natural correlation between the data. Applications (such as production monitoring MES systems) can directly construct the query statement "SELECT * FROM Machine_Status WHERE ID = 'CNC_001'" through the standard SQL interface to retrieve all status parameters of the machine tool at once. This greatly simplifies the application development logic, eliminating the need for separate querying of key-value pairs and subsequent reassembly as in traditional architectures.

[0043] In the production process, assuming CNC machine tool production line data is scattered, it would take at least several hours to integrate and generate a quality analysis report based on big data analysis, resulting in extremely high rework costs after defective products are discovered. By introducing an in-memory relational database, production line data can be aggregated and analyzed at the millisecond level, reducing the time to detect defective products to minutes. This reduces the number of defective products and rework costs, demonstrating the advantages of millimeter-level latency.

[0044] Meanwhile, the built-in change subscription module of the in-memory relational database monitors data changes in the data tables in real time. When the spindle temperature data is updated, the change subscription module immediately identifies the change and actively pushes it to the time-series database for storage. This process achieves full real-time synchronization between the in-memory relational database and the time-series database. Even in the event of a sudden system failure or power outage, the time-series database still retains the latest temperature anomaly records, providing complete data support for subsequent accident tracing and fault analysis, effectively solving the data loss problem that may be caused by the traditional "minute-level refresh" mechanism.

[0045] Through the implementation of the above application scenarios, this invention verifies the significant advantages of the intelligent manufacturing IoT platform based on in-memory relational database in terms of data correlation, real-time control response, and data integrity, which can effectively meet the complex needs of intelligent manufacturing scenarios.

[0046] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention, such as replacing the memory relational database with other real-time databases with row attribute correlation characteristics, or replacing the industrial bus protocol with other real-time communication protocols suitable for industrial scenarios, should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A smart manufacturing IoT platform based on an in-memory relational database, characterized by: This includes industrial gateways, in-memory relational databases, application-side components, and time-series databases. The industrial gateway (controller) is equipped with a bus protocol database adapter. One end of the industrial gateway is connected to the device, and the other end is connected to the memory relational database through the bus protocol database adapter. The memory relational database is equipped with a fusion model, which integrates data from multiple monitoring points of the same device into a data table based on row attribute correlation features; The aforementioned memory relational database also provides a standard open interface to connect with the application terminal and transmit device data to the application terminal; The memory relational database pushes the changes in the collected device data to the time-series database to achieve real-time data synchronization between the memory relational database and the time-series database.

2. The intelligent manufacturing IoT platform based on an in-memory relational database according to claim 1, characterized in that, The bus protocol database adapter supports ProfiNet, EtherCat, and EtherNet industrial bus protocols. The industrial gateway communicates with the memory relational database via ProfiNet, EtherCat, or EtherNet.

3. The intelligent manufacturing IoT platform based on an in-memory relational database according to claim 1, characterized in that, The memory-based relational database pushes the changes in collected device data to the time-series database through a built-in change subscription module.

4. The intelligent manufacturing IoT platform based on an in-memory relational database according to claim 1, characterized in that, The standard open interface is the SQL standard interface, through which the application queries device data in the memory relational database.

5. The intelligent manufacturing IoT platform based on an in-memory relational database according to claim 1, characterized in that, In the data table, each row represents the collected data of a device identifier. The collected data includes the update time sequence and the update values ​​of all points.

6. The intelligent manufacturing IoT platform based on an in-memory relational database according to claim 1, characterized in that, It is equipped with dual-link independent channels, which include a data channel and a command channel. The data channel is used for the uplink of device data and the downlink of data analysis results from the application end. The command channel is used to transmit control commands from the application end to the device and configuration commands from the industrial gateway.

7. A method for implementing an intelligent manufacturing IoT platform based on an in-memory relational database, based on the intelligent manufacturing IoT platform based on an in-memory relational database as described in any one of claims 1-6, characterized in that, Includes the following steps: Data is collected from multiple monitoring points of the equipment, and the collected multi-dimensional equipment time-series data is transferred to an in-memory relational database. Based on the row attribute correlation features, the multi-dimensional device time series data corresponding to the same device identifier are integrated and processed in an integrated manner, and the integrated data is stored in a pre-set data table in a structured manner. The structured and integrated device data is proactively pushed to the application so that the application can access and use it. Incremental identification is performed on continuously collected device data, and the changes in the device data are extracted in real time and incrementally pushed to the time-series database to achieve full real-time synchronization of device data between the memory relational database and the time-series database.

8. The implementation method of the intelligent manufacturing IoT platform based on an in-memory relational database according to claim 7, characterized in that, The method of proactively pushing structured and integrated device data to the application is as follows: Using the SQL standard interface as the standard open interface, the application performs query operations on device data in the memory relational database based on the SQL standard interface.

9. The implementation method of the intelligent manufacturing IoT platform based on an in-memory relational database according to claim 7, characterized in that, The method for incremental identification of continuously collected device data is as follows: the amount of change in device data is identified in real time through the change subscription module built into the memory relational database.

Images