AI Model Hot-Swap Method, System, Electronic Device and Storage Medium Based on Building Intelligent Controller

Abstract

The present invention discloses an AI model hot plugging method, system, electronic device and storage medium based on a building intelligent controller. The AI model hot plugging system based on a building intelligent controller includes a hardware resource management layer, which is used to manage the underlying port resources; obtain the DI and AI values according to the model logic / configuration strategy and actively report them; control the operation of the terminal device according to the model logic or the control instructions issued by the upper computer / business platform. The AI model hot plugging method, system, electronic device and storage medium based on a building intelligent controller disclosed by the present invention have the beneficial effects that they are applied to the application framework of the main control devices in a building automation system, are based on the physical model, and support the hot plugging, multi-task concurrency and collaboration, active detection and active push of port status of functional module / algorithm block modules of application types.

Description

Technical Field

[0001] The present invention belongs to the field of automatic control, and particularly relates to an AI model hot plugging method, system, electronic device and storage medium based on a building intelligent controller. Background Art

[0002] The utility model patent with the publication number CN204536884 and the theme name of a multi-channel serial interface intelligent building control system has an IPC classification number of G05B19 / 418, and its technical solution discloses "including a serial interface device 5, the serial interface device 5 is connected to a switch 2, the switch 2 is connected to an intelligent management host 1 and a video processing center 3, and the video control center 3 is connected to a display 4; and the serial interface device 5 is composed of a controller 9, a switch 10, a voltage regulator 11, a one-key recovery unit 12, an indicator light 13, a serial interface unit 14 and a network interface unit 15, and the controller 9 is respectively connected to the switch 10, the voltage regulator 11, the one-key recovery unit 12, the indicator light 13, the serial interface unit 14 and the network interface unit 15. The controller 9 is a programmable PLC".

[0003] It can be seen that the above utility model patent has already disclosed a building control system with a programmable PLC as the controller. However, the technical solution disclosed in the above utility model patent does not further solve the problems solved by the present invention and needs to be further improved.

[0004] Specifically, in a building automation control system, the main control devices (DDC and PLC) are all devices with fixed functions. Their existing equipped functions can be programmed, but they cannot increase capabilities beyond the inherent functions, nor do they support the currently booming AI, resulting in a relatively low level of intelligence of the main control devices (DDC and PLC).

[0005] At the same time, the configuration of the main control devices (DDC and PLC) is difficult and the programming requirements have a high threshold, which is not conducive to the rapid application of the current rapidly developing / rapidly applied Internet +.

[0006] At the same time, the current main control devices (DDC and PLC) all work in a task-based scanning manner (that is, the scanning of hardware ports is triggered by tasks to obtain the status of hardware ports), which is not conducive to multi-task collaborative work and the working efficiency of hardware resources is relatively low. Summary of the Invention

[0007] In view of the current situation of the prior art, the present invention overcomes the above defects and provides an AI model hot plugging method, system, electronic device and storage medium based on a building intelligent controller.

[0008] The AI model hot-swap method, system, electronic device and storage medium based on a building intelligent controller disclosed by the present invention aim, among other things, to load and unload functional modules in a hot-swap manner for the intelligent controller / PLC.

[0009] The AI model hot-swap method, system, electronic device and storage medium based on a building intelligent controller disclosed by the present invention aim, among other things, for the intelligent controller / PLC to have the ability to collect data and actively share functional block data.

[0010] The present invention adopts the following technical solutions. The AI model hot-swap system based on a building intelligent controller includes:

[0011] The hardware resource management layer is used to manage the underlying port resources; obtain the values of DI and AI according to the model logic / configuration strategy and report them actively; control the operation of the terminal device according to the model logic or the control instructions issued by the host computer / business platform.

[0012] The service management layer is used to manage the operation mode of the hardware resource management layer, manage algorithm blocks, sub-application blocks, and programming logic blocks; receive the actively reported data from the hardware resource management layer and distribute it to the required algorithm blocks, sub-application blocks, and programming logic blocks; obtain the operation results of the algorithm blocks, sub-application blocks, and programming logic blocks and share them with other required algorithm blocks, sub-application blocks, and programming logic blocks; execute the device control instructions generated by the algorithm blocks, sub-application blocks, and programming logic blocks.

[0013] The algorithm module includes multiple functional blocks. The functional blocks include algorithm blocks, sub-application blocks, and programming logic blocks: the algorithm module generates calculation result data or device control instructions, and any one functional block conducts data interaction with other functional blocks of the algorithm module through the service management layer.

[0014] The sub-application block provides specific functional services to users.

[0015] The programming logic block is used to execute the operation logic of the physical model.

[0016] The present invention adopts the following technical solutions. The AI model hot-swap method based on a building intelligent controller is applied to the AI model hot-swap system based on a building intelligent controller. The AI model hot-swap method based on a building intelligent controller includes an active detection step, and the specific implementation of the active detection step is as follows:

[0017] Step Q1: After the service management layer receives the configuration file of the hardware from the host computer / platform software, it distributes the configuration file to the hardware resource management layer.

[0018] Step Q2: The interface service in the hardware resource management layer submits the interpreted configuration file to the core thread of the hardware resource management layer, and the core thread of the hardware resource management layer executes it;

[0019] Step Q3: The hardware resource management layer performs the action of obtaining port resources according to the configuration file.

[0020] The present invention adopts the following technical solutions. The AI model hot plugging method based on a building intelligent controller is applied to the AI model hot plugging system based on a building intelligent controller. The AI model hot plugging method based on a building intelligent controller includes an information active push step, and the information active push step is specifically implemented as the following steps:

[0021] Step W1: The algorithm block / sub-application block / programming logic block subscribes to the working status of the device model through the interface service bus of the service management layer;

[0022] Step W2: The hardware resource management layer obtains the status of the port resources through the configuration file. If the change range of the port resources meets the preset conditions, the hardware resource management layer reports to the service management layer through the service port;

[0023] Step W3: The service management layer matches the port resources to the working status of the device model;

[0024] Step W4: The service management layer pushes data to the algorithm block / sub-application block / programming logic block that subscribes to the device model through the interface service bus;

[0025] Step W5: The algorithm block / sub-application block / programming logic block processes the received device status data according to its own operation logic and generates an operation result. The operation result is pushed to the service management layer through the interface service bus of the service management layer, and the service management layer processes according to the received operation result.

[0026] The present invention adopts the following technical solutions. The AI model hot plugging method based on a building intelligent controller is applied to the AI model hot plugging system based on a building intelligent controller. The AI model hot plugging method based on a building intelligent controller includes a hot insertion step, and the hot insertion step is specifically implemented as the following steps:

[0027] Step S1: The upper computer / business platform sends a loading notice of the algorithm block / sub-application block / programming logic block to the intelligent controller;

[0028] Step S2: After receiving the loading notice of the algorithm block / sub-application block / programming logic block, the intelligent controller downloads the target function block from the upper computer / business platform through FTP. After successful download, it verifies the integrity of the algorithm block / sub-application block / programming logic block;

[0029] Step S3: The intelligent controller loads algorithm blocks / sub-application blocks / programming logic blocks and obtains the registration information of the target function block;

[0030] Step S4: The intelligent controller loads the configuration parameters of the target function block and configures the target function block;

[0031] Step S5: The intelligent controller allocates resources for the target function block according to the function of the target function block, and subscribes to the operation status data of the corresponding device model or the results of other function blocks according to the configuration parameters of the target function block;

[0032] Step S6: When the service management layer receives the operation status data of the device model or the results of other function blocks and the results change, the service management layer submits the operation status data of the device model or the results of other function blocks to the target function block;

[0033] Step S7: The target function block calculates according to the received data and outputs the results to the service management layer;

[0034] Step S8: After receiving the output results of the target function block, the service management layer performs corresponding processing.

[0035] The present invention adopts the following technical solutions. An AI model hot-plugging method based on a building intelligent controller is applied to an AI model hot-plugging system based on a building intelligent controller. The AI model hot-plugging method based on a building intelligent controller includes a hot-unplugging step, and the specific implementation of the hot-unplugging step is as follows:

[0036] Step K1: The upper computer / business platform sends a target function block unloading instruction to the intelligent controller;

[0037] Step K2: The intelligent controller notifies other function blocks related to the target function block of the message that the target function block stops running;

[0038] Step K3: The intelligent controller stops pushing data to the target function block;

[0039] Step K4: The intelligent controller stops the operation of the target function block;

[0040] Step K5: The intelligent controller unregisters and deletes the target function block;

[0041] Step K6: The intelligent controller releases the resources allocated for the target function block.

[0042] The present invention adopts the following technical solutions. An electronic device includes:

[0043] A processor and a memory, where the memory stores executable instructions of the processor; wherein:

[0044] The processor is configured to execute the steps of any of the above technical solutions of the AI model hot plugging method based on the building intelligent controller by executing executable instructions.

[0045] The present invention adopts the following technical solution: a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above technical solutions of the AI model hot plugging method based on the building intelligent controller are implemented.

[0046] The AI model hot plugging method, system, electronic device and storage medium based on the building intelligent controller disclosed by the present invention have the beneficial effects that they are applied to the application framework of the main control devices (intelligent controller / PLC) in the building automation system. Based on the object model, they support the hot plugging, multi-task concurrency and collaboration of function module / algorithm block modules of application types, and the active detection and active push of port states, so as to expand the intelligent capabilities of the intelligent controller / PLC, reduce its configuration difficulty and programming threshold, and at the same time improve the working efficiency of the hardware resources of the intelligent controller / PLC. Description of the Drawings

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts. In addition, other devices mentioned below specifically refer to the supporting devices outside the intelligent controller.

[0048] Figure 1 It is the overall framework diagram of an intelligent controller provided by the technical solution of the present application.

[0049] Figure 2 It is the active detection workflow diagram of an intelligent controller provided by the technical solution of the present application.

[0050] Figure 3 It is the information active push flowchart of an intelligent controller provided by the technical solution of the present application.

[0051] Figure 4 It is the schematic diagram of the hot plugging of the function module / AI algorithm of an intelligent controller provided by the technical solution of the present application.

[0052] Figure 5 It is the hot insertion flowchart of the function module / AI algorithm of an intelligent controller provided by the technical solution of the present application.

[0053] Figure 6 It is the hot extraction flowchart of the function module / AI algorithm of an intelligent controller provided by the technical solution of the present application. Detailed Implementation Modes

[0054] The present invention discloses an AI model hot plugging method, system, electronic device and storage medium based on a building intelligent controller. The following further describes the specific implementation modes of the present invention in combination with preferred embodiments.

[0055] Those skilled in the art should note that the "DDC" and "intelligent controller" involved in various embodiments of the present invention are the same concept and will not be distinguished anymore.

[0056] Those skilled in the art should note that the "physical model" and "device model" involved in various embodiments of the present invention are the same concept and will not be distinguished anymore.

[0057] Those skilled in the art should note that the "building intelligent controller" involved in various embodiments of the present invention is defined as: the building intelligent controller, as an intelligent controller applied to a specific field (i.e., buildings) in intelligent controllers, highly conforms to the actual requirement of "building automation" and works in cooperation with terminal devices to achieve the expected purpose.

[0058] Embodiment 1.

[0059] Preferably, an AI model hot plugging system based on a building intelligent controller includes:

[0060] A hardware resource management layer, which is used to manage the underlying port resources; obtain the values of DI and AI according to the model logic / configuration strategy and report them actively; control the operation of the terminal device according to the model logic or the control instructions issued by the host computer / business platform;

[0061] A service management layer, which is used to manage the operation mode of the hardware resource management layer, manage the algorithm blocks, sub-application blocks, and programming logic blocks; receive the data actively reported by the hardware resource management layer and distribute it to the required algorithm blocks, sub-application blocks, and programming logic blocks; obtain the operation results of the algorithm blocks, sub-application blocks, and programming logic blocks and share them with other required algorithm blocks, sub-application blocks, and programming logic blocks; execute the device control instructions generated by the algorithm blocks, sub-application blocks, and programming logic blocks;

[0062] An algorithm module, which includes multiple functional blocks. The functional blocks include algorithm blocks, sub-application blocks, and programming logic blocks: the algorithm module generates calculation result data or device control instructions, and any one functional block performs data interaction with other functional blocks of the algorithm module through the service management layer;

[0063] A sub-application block, which provides specific functional services to users;

[0064] A programming logic block, which is used to execute the operation logic of the physical model.

[0065] Preferably, the AI model hot-plugging method based on a building intelligent controller is applied to an AI model hot-plugging system based on a building intelligent controller. The AI model hot-plugging method based on a building intelligent controller includes an active detection step, and the active detection step is specifically implemented as the following steps:

[0066] Step Q1: After the service management layer receives the configuration file of the hardware from the upper computer / platform software, it distributes the configuration file to the hardware resource management layer;

[0067] Step Q2: The hardware resource management layer submits the interpreted configuration file to the hardware resource management layer for execution by the hardware resource management layer;

[0068] Step Q3: The hardware resource management layer performs the action of obtaining port resources according to the configuration file.

[0069] Preferably, the AI model hot-plugging method based on a building intelligent controller is applied to an AI model hot-plugging system based on a building intelligent controller. The AI model hot-plugging method based on a building intelligent controller includes an information active push step, and the information active push step is specifically implemented as the following steps:

[0070] Step W1: The algorithm block / sub-application block / programming logic block subscribes to the working status of the device model through the interface service bus of the service management layer;

[0071] Step W2: The hardware resource management layer obtains the status of the port resources through the configuration file. If the change range of the port resources meets the preset conditions, the hardware resource management layer reports to the service management layer through the service port;

[0072] Step W3: The service management layer matches the port resources to the working status of the device model;

[0073] Step W4: The service management layer pushes data to the algorithm block / sub-application block / programming logic block that subscribes to the device model through the interface service bus;

[0074] Step W5: The algorithm block / sub-application block / programming logic block processes the received device status data according to its own operation logic and generates an operation result. The operation result is pushed to the service management layer through the interface service bus of the service management layer, and the service management layer processes according to the received operation result.

[0075] Preferably, the AI model hot-plugging method based on a building intelligent controller is applied to an AI model hot-plugging system based on a building intelligent controller. The AI model hot-plugging method based on a building intelligent controller includes a hot insertion step, and the hot insertion step is specifically implemented as the following steps:

[0076] Step S1: The host computer / business platform sends a loading notification for the algorithm block / sub-application block / programming logic block to the intelligent controller;

[0077] Step S2: After receiving the loading notification for the algorithm block / sub-application block / programming logic block, the intelligent controller downloads the target function block from the host computer / business platform via FTP. After successful download, it performs integrity verification on the algorithm block / sub-application block / programming logic block;

[0078] Step S3: The intelligent controller loads the algorithm block / sub-application block / programming logic block and obtains the registration information of the target function block;

[0079] Step S4: The intelligent controller loads the configuration parameters of the target function block and configures the target function block;

[0080] Step S5: The intelligent controller allocates resources for the target function block according to the function of the target function block, and subscribes to the operation status data of the corresponding device model or the results of other function blocks according to the configuration parameters of the target function block;

[0081] Step S6: When the service management layer receives the operation status data of the device model or the results of other function blocks and the results change, the service management layer submits the data to the target function block;

[0082] Step S7: The target function block performs calculations based on the received data and outputs the results to the service management layer;

[0083] Step S8: After receiving the output results of the target function block, the service management layer performs corresponding processing.

[0084] Preferably, the AI model hot-swap method based on the building intelligent controller is applied to the AI model hot-swap system based on the building intelligent controller. The AI model hot-swap method based on the building intelligent controller includes a hot-unplugging step. The specific implementation of the hot-unplugging step is as follows:

[0085] Step K1: The host computer / business platform sends a target function block unloading instruction to the intelligent controller;

[0086] Step K2: The intelligent controller notifies other function blocks related to the target function block of the message to stop the operation of the target function block;

[0087] Step K3: The intelligent controller stops pushing data to the target function block;

[0088] Step K4: The intelligent controller stops the operation of the target function block;

[0089] Step K5: The intelligent controller unregisters and deletes the target function block;

[0090] Step K6: The intelligent controller releases the resources allocated to the target function block.

[0091] Preferably, this embodiment discloses an electronic device, including:

[0092] a processor, and a memory storing executable instructions of the processor; wherein:

[0093] The processor is configured to execute the steps of any of the above technical solutions of the AI model hot-swap method based on the building intelligent controller by executing the executable instructions.

[0094] Preferably, this embodiment discloses a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, it implements the steps of any of the above technical solutions of the AI model hot-swap method based on the building intelligent controller.

[0095] The working principles of the AI model hot-swap method, system, electronic device, and storage medium based on the building intelligent controller disclosed in this embodiment are described below.

[0096] 1) As Figure 1 shown, the AI model hot-swap system based on the building intelligent controller includes the following parts:

[0097] (1) Hardware resource management layer (interface service): mainly manages the underlying port resources, including DI, AI, DO, AO; obtains the values of DI and AI according to the model logic / configuration strategy (including high-speed scanning / timed scanning / task-based scanning, etc.) and reports them actively; controls the operation of the terminal devices according to the model logic or control instructions issued by the upper computer / business platform, and uses the first-come-first-served method to solve the problem of control instruction conflicts of terminal devices.

[0098] (2) Service management layer (interface service bus): mainly responsible for managing the operation mode of the hardware resource management layer (issuing configurations / models); manages function blocks such as algorithm blocks, sub-application blocks, and programming logic blocks (including insertion and removal, etc.); receives the actively reported data from the hardware resource management layer and distributes it to the required algorithm blocks, sub-application blocks, programming logic blocks, etc.; obtains the operation results of the algorithm blocks, sub-application blocks, programming logic blocks, etc. and shares them with other required algorithm blocks, sub-application blocks, programming logic blocks, etc.; executes the device control instructions generated by the algorithm blocks, sub-application blocks, programming logic blocks, etc.

[0099] (3) Algorithm module: It contains multiple functional blocks with specific algorithms. The functional blocks include algorithm blocks, sub - application blocks, and programming logic blocks, and generally generate calculation result data or device control instructions (such as predicting the device energy consumption requirements in a building based on the operating data of the device and changes in the spatial environment). Any functional block can perform data interaction with other functional blocks (other algorithm blocks, sub - application blocks, and programming logic blocks) in the algorithm module through the service management layer.

[0100] (4) Sub - application block: A functional block that provides specific functional services to users (such as energy - saving management, etc.).

[0101] (5) Programming logic block: It mainly executes the operation logic of the physical model, including the linkage control between multiple models under the same intelligent controller, such as the fan operation module, adjustable lighting fixture operation module, etc.

[0102] 2) As Figure 2 shown, the AI model hot - plugging method based on the building intelligent controller (of the AI model hot - plugging system based on the building intelligent controller) includes an active detection step, and the specific implementation of the active detection step is as follows:

[0103] Step Q1: After the service management layer receives the configuration file (including high - speed scanning / timed scanning / task - based scanning, etc.) of the hardware from the upper computer / platform software, (the service management layer) sends the configuration file to the interface service in the hardware resource management layer;

[0104] Step Q2: The interface service in the hardware resource management layer submits the interpreted configuration file to the core thread of the hardware resource management layer, and the core thread of the hardware resource management layer executes it;

[0105] Step Q3: The hardware resource management layer performs the action of obtaining port resources (i.e., the working status of the end - device or the detection value of the sensor) according to the configuration file (including high - speed scanning / timed scanning / task - based scanning, etc.).

[0106] 3) As Figure 3 shown, the AI model hot - plugging method based on the building intelligent controller (of the AI model hot - plugging system based on the building intelligent controller) includes an information active push step, and the specific implementation of the information active push step is as follows:

[0107] Step W1: The algorithm block / sub - application block / programming logic block subscribes to the working status of the device model (physical model) through the interface service bus of the service management layer;

[0108] Step W2: The hardware resource management layer obtains the status of port resources through a configuration file (including high-speed scanning / timed scanning / task-based scanning, etc.). If the change range of the port resources (including the ID information, data type, and value of the resources, etc.) meets the preset conditions (when the data change exceeds the specified amplitude or is forced to report as required, etc.), the hardware resource management layer reports to the service management layer through the service port;

[0109] Step W3: The service management layer matches the port resources to the working status of the device model (physical model);

[0110] Step W4: The service management layer pushes data to the algorithm block / sub-application block / programming logic block that subscribes to the device model (physical model) through the interface service bus;

[0111] Step W5: The algorithm block / sub-application block / programming logic block processes the received device status data according to its own operating logic and generates an operation result. The operation result is pushed to the service management layer through the interface service bus of the service management layer, and the service management layer processes according to the received operation result (if it is a data result, it is shared according to the configuration; if it is a control instruction, it is sent to the hardware resource management layer for execution).

[0112] 4) As Figure 4 shown, the AI model hot-swap system based on the building intelligent controller includes the following:

[0113] The service management layer and two parts: the algorithm block / sub-application block / programming logic block. Among them: The service management layer (including the interface service bus) provides management functions for the algorithm block / sub-application block / programming logic block, including hot-swap of the algorithm block / sub-application block / programming logic block, data sharing, and execution of device control instructions; The algorithm block / sub-application block / programming logic block is a functional block with certain business functions (such as device linkage control, the device automatically adjusting the operating state according to the environmental situation, etc.).

[0114] 5) As Figure 5 shown, the AI model hot-swap method based on the building intelligent controller (of the AI model hot-swap system based on the building intelligent controller) includes a hot-insertion step, and the specific implementation of the hot-insertion step is the following steps:

[0115] Step S1: The upper computer / business platform sends a loading notice of the algorithm block / sub-application block / programming logic block to the intelligent controller;

[0116] Step S2: After the intelligent controller receives the loading notification of the algorithm block / sub-application block / programming logic block, it downloads the target function block from the host computer / business platform via FTP. After successful download, it performs integrity verification on the algorithm block / sub-application block / programming logic block (including verifying the resources required for the operation of the target function block: CPU, memory, storage space, etc. If the resources required for the operation of the target function block exceed the current idle resources of the intelligent controller, the intelligent controller replies to the host computer / business platform that the current algorithm block / sub-application block / programming logic block loading fails);

[0117] Step S3: The intelligent controller loads the algorithm block / sub-application block / programming logic block and obtains the registration information of the target function block (such as the name, ID information, main function, operation result information, etc. of the function block);

[0118] Step S4: The intelligent controller loads the configuration parameters of the target function block and configures the target function block (such as the control delay time of the target device, etc.);

[0119] Step S5: The intelligent controller allocates resources for the target function block according to the function of the target function block (including memory operation resources, data storage resources, etc., the memory address of the operation result, etc.), and subscribes to the operation status data of the corresponding device model (physical model) or the results of other function blocks according to the configuration parameters of the target function block;

[0120] Step S6: When the service management layer receives the operation status data of the device model (physical model) or the results of other function blocks, and the results (of the operation status data of the device model (physical model) or other function blocks) change, the service management layer submits the operation status data of the device model (physical model) or the results of other function blocks to the target function block;

[0121] Step S7: The target function block calculates according to the received data and outputs the results (data or device control instructions) to the service management layer;

[0122] Step S8: After the service management layer receives the output results of the target function block, it performs corresponding processing (if it is data, it is stored and then pushed to other function blocks that need the results; if it is a device control instruction, it is sent to the hardware resource management layer for execution).

[0123] 6) As Figure 5 shown, the AI model hot-swap method based on the building intelligent controller (of the AI model hot-swap system based on the building intelligent controller) includes a hot-unplugging step, and the specific implementation of the hot-unplugging step is as follows:

[0124] Step K1: The host computer / business platform sends a target function block unloading instruction to the intelligent controller;

[0125] Step K2: The intelligent controller notifies other function blocks related to the target function block of the message that the target function block stops running;

[0126] Step K3: The intelligent controller stops pushing data to the target function block (including the state change of the device and the calculation results of other function blocks, etc.);

[0127] Step K4: The intelligent controller stops the operation of the target function block;

[0128] Step K5: The intelligent controller de-registers and deletes the target function block (including the configuration);

[0129] Step K6: The intelligent controller releases the resources allocated to the target function block (including memory operation resources and data storage resources).

[0130] The overall concept of the AI model hot-swap method, system, electronic device and storage medium based on the building intelligent controller disclosed in this embodiment is elaborated below.

[0131] Specifically, the intelligent controller / PLC of building automation can add functions in a hot-swap manner (if the demand for new functions exceeds the resources of the hardware device itself, the hardware device does not load the function and prompts the user that the loading fails), and after loading and unloading function blocks (algorithm blocks / sub-application blocks / programming logic blocks), it does not affect the operation of the original functions. At the same time, it can actively scan port values (the working state of the device / sensor values) in multiple ways (high-speed scanning / timed scanning / task-based scanning) and actively report the changed data to function blocks (algorithm blocks / sub-application blocks / programming logic blocks) to prompt the function blocks to perform their functions.

[0132] Among them, compared with similar controllers (DDC / PLC) in the building automation system, the AI model hot-swap method, system, electronic device and storage medium based on the building intelligent controller disclosed in this embodiment are based on the physical model, and through supporting the hot-swap of function modules / algorithm block modules of application types, multi-task concurrency and coordination, and active detection and active pushing of port states, it realizes the expansion of the intelligent capabilities of the intelligent controller / PLC, solves the problems of difficult configuration, low intelligence level and high programming threshold of the current intelligent controller. At the same time, it aims to improve the working efficiency of the hardware resources of the intelligent controller / PLC, accelerate the development of AI in building automation, and bring diverse intelligent management and control for industry applications.

[0133] The industrial background of the AI model hot-swap method, system, electronic device and storage medium based on the building intelligent controller disclosed in this embodiment is elaborated below.

[0134] Specifically, in modern industrial production, PLC (Programmable Logic Controller) has become an indispensable technology. It plays a core role in achieving precise and efficient industrial automation. However, as industrial systems become increasingly complex, traditional PLC programming methods are beginning to show their limitations. These methods are often time-consuming and error-prone, restricting the improvement of production efficiency. Fortunately, in recent years, the rise of LLMs (Large Language Models) like ChatGPT, as an advanced artificial intelligence (AI) technology, is bringing a revolutionary change to the PLC industry:

[0135] 1) The revolutionary role of AI in PLC programming

[0136] (1) Improve efficiency: Engineers no longer need to spend a lot of time writing and debugging code. They can invest more time in creativity and strategic planning rather than being tied up with trivial programming tasks.

[0137] (2) Reduce errors: Human programming is prone to errors, but AI can generate more accurate and reliable code by analyzing a large number of cases and data. It's like having an expert who never makes mistakes to help you.

[0138] (3) Easy to use: Even people without a deep programming background can generate the required PLC code by communicating with an AI assistant. This greatly reduces the technical threshold.

[0139] 2) Applying artificial intelligence to industrial automation control systems constructs a more efficient, accurate, and adaptive manufacturing and processing environment. With the support of artificial intelligence, automated systems can learn from historical and real-time data, predict future trends and potential problems, and thus can make corresponding adjustments in advance, reducing downtime and production delays. The integration of artificial intelligence enables devices to interact with workers more intelligently, enhancing operational safety and flexibility. At the same time, it also significantly improves resource utilization efficiency and productivity, reduces energy consumption and waste. Using artificial intelligence technology can also better improve the performance of industrial automation control systems. Using artificial intelligence technology, the performance of the entire automated control system can be enhanced. For example, in the working process of a fuzzy logic controller, the reaction speed of the controller can be well improved, greatly increasing the work efficiency. The advantages of using artificial intelligence in industrial automation control systems are as follows:

[0140] (1) Improve precision

[0141] Monitor and analyze all aspects of the production process in real time, effectively identify and predict factors that may lead to production deviations. With advanced sensors and computer vision technology, it is possible to finely adjust the operation of mechanical equipment even when minor changes can be detected, ensuring the precise use of materials and the precision machining of products during the production process.

[0142] (2) Improve stability

[0143] Applying artificial intelligence technology to industrial automation control systems can effectively reduce the process links in traditional industrial automation control systems. It only needs to follow the guidance of artificial intelligence technology and input according to the requirements, and the stability of its operation can be ensured. In addition, the computer system in artificial intelligence technology can ensure highly accurate calculations and effectively eliminate many external interference factors, making the industrial automation control system operate more stably, smoothly and safely, thereby improving the overall operation quality and operation efficiency of industrial automation control.

[0144] (3) Reduce human resources

[0145] Through artificial intelligence technology, the automation system can perform complex tasks while reducing the burden of manual monitoring. While liberating human resources, it can also rely on the data processing system in the artificial intelligence system to effectively process a large amount of data. Secondly, the amount of data generated in industrial automation control systems is very large. Relying on artificial intelligence technology can effectively reduce such cumbersome work, liberate human resources and improve the quality and efficiency of work in all aspects of the industry.

[0146] (4) Reduce the error frequency

[0147] Traditional industrial automation control systems rely on manual operations and are prone to human errors caused by individual differences, which will directly affect the speed and quality of industrial operation, and even cause economic losses and safety hazards. Applying artificial intelligence to actual industrial automation control systems can effectively reduce some unnecessary processes and ensure the timeliness of work; on the other hand, it can effectively reduce human errors, thus avoiding the above-mentioned problems such as safety and work errors.

[0148] (5) Obtain valuable experience from data

[0149] The system generates a large amount of valuable data in a day. With the help of the right artificial intelligence, all raw data can be transformed into useful insights to guide designers or engineers to improve and update new methods based on the latest technological discoveries.

[0150] (6) Enhance control technology and process conceptual data through data-driven deep learning and cognitive computing

[0151] Deep learning uses artificial intelligence technology based on artificial neural networks and can extract advanced insights from raw data inputs. Cognitive computing focuses on high-level understanding and reasoning.

[0152] (7) Assist decision-making through reinforcement learning and big data analysis

[0153] Reinforcement learning is a cutting-edge artificial intelligence technology that attempts to train machine learning models for advanced decision-making. The ML model uses trial and error to find appropriate solutions to any complex problems. Big data analysis can discover valuable patterns, trends, correlations, and preferences to help users make better decisions.

[0154] (8) Analyze and predict future trends through artificial intelligence

[0155] Therefore, deep learning models can use unstructured data sets to predict future trends and provide a lead for system control.

[0156] The following elaborates on the experimental verification of the AI model hot-swap method, system, electronic device, and storage medium based on a building intelligent controller disclosed in this embodiment.

[0157] Specifically, the AI model hot-swap method, system, electronic device, and storage medium based on a building intelligent controller disclosed in this embodiment have been coded and verified. The experimental system includes 1 intelligent controller, 1 set of host computer software, 1 switch panel, 1 fixed-brightness lamp, 1 adjustable-brightness lamp, and 1 set of timing dimming modules. The experimental content is as follows:

[0158] 1) Power on the intelligent controller and connect it to the network. Connect the switch panel, fixed-brightness lamp, adjustable-brightness lamp, and timing dimming module to the intelligent controller.

[0159] 2) Connect the host computer to the network and connect to the intelligent controller.

[0160] 3) The host computer writes a set of programs for the intelligent controller to control the dimming module regularly (relatively quickly) to make the intelligent controller run normally. The host computer can normally receive the device status reported by the intelligent controller (the status value of the dimming module and the status value of the adjustable-brightness lamp). At this time, the switch panel and the fixed-brightness lamp are both unavailable.

[0161] 4) The host computer writes another set of programs to enable the switch panel to control the on / off of the fixed-brightness lamp.

[0162] 5) Send the program written by the host computer to the intelligent controller, and observe the changes of the dimmable lights during the sending process (Normal: The brightness change of the dimmable lights maintains the original rule, and the status data of the dimmable lights received by the host computer maintains the original rule; Abnormal: The control of the dimmable lights is abnormal, and the status data of the dimmable lights received by the host computer is abnormal).

[0163] It is worth mentioning that the specific selection of the processor and other technical features involved in this invention patent application should be regarded as the prior art. The specific structure, working principle, possible control methods, and spatial arrangement methods of these technical features can be selected conventionally in the art, and should not be regarded as the invention point of this invention patent. This invention patent will not be further specifically elaborated.

[0164] For those skilled in the art, it is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An AI model hot-swap system based on a building intelligent controller, characterized in that: include: The hardware resource management layer is used to manage the underlying port resources; obtain the values ​​of DI and AI according to the model logic / configuration strategy, and actively report them; control the operation of the terminal equipment according to the control instructions issued by the model logic or the host computer / business platform; The service management layer is used to manage the operation mode of the hardware resource management layer, manage the algorithm blocks, sub-application blocks, and programming logic blocks; receive the data actively reported by the hardware resource management layer and distribute it to the required algorithm blocks, sub-application blocks, and programming logic blocks; Obtain the calculation results of the algorithm block, sub-application block, and programming logic block and share them with other required algorithm blocks, sub-application blocks, and programming logic blocks; execute the device control instructions generated by the algorithm block, sub-application block, and programming logic block; Algorithm module, which includes multiple functional blocks, including algorithm block, sub-application block and programming logic block: the algorithm module generates calculation result data or device control instructions, and any functional block exchanges data with other functional blocks of the algorithm module through the service management layer; Sub-application block: The sub-application block provides corresponding functional services to users; Programming logic block, which is used to execute the operation logic of the object model; The AI ​​model hot-swap method based on a building intelligent controller, which is applied to an AI model hot-swap system based on a building intelligent controller, comprises an information active push step, which is specifically implemented as the following steps: Step W1: the algorithm block / sub-application block / programming logic block subscribes to the working status of the device model through the interface service bus of the service management layer; Step W2: The hardware resource management layer obtains the status of the port resources through the configuration file. If the change range of the port resources meets the preset conditions, the hardware resource management layer reports to the service management layer through the service port; Step W3: The service management layer matches the port resources to the working status of the device model; Step W4: The service management layer pushes data to the algorithm block / sub-application block / programming logic block that subscribes to the device model through the interface service bus; Step W5: The algorithm block / sub-application block / programming logic block processes the received device status data according to its own operation logic and generates an operation result, which is pushed to the service management layer through the interface service bus of the service management layer, and the service management layer processes the received operation result.

2. A hot-swap method for an AI model based on a building intelligent controller, characterized in that: The AI ​​model hot-swap system based on a building intelligent controller and the AI ​​model hot-swap method based on a building intelligent controller applied to claim 1 include an active detection step, which is specifically implemented as the following steps: Step Q1: After receiving the hardware configuration file from the host computer / platform software, the service management layer sends the configuration file to the hardware resource management layer; Step Q2: The interface service in the hardware resource management layer submits the interpreted configuration file to the core thread of the hardware resource management layer, which executes it; Step Q3: The hardware resource management layer executes the port resource acquisition action according to the configuration file.

3. A hot-swap method for an AI model based on a building intelligent controller, characterized in that: The AI ​​model hot-swap system based on a building intelligent controller and the AI ​​model hot-swap method based on a building intelligent controller applied to claim 1 include a hot-swap step, which is specifically implemented as the following steps: Step S1: The host computer / business platform sends a notification of loading the algorithm block / sub-application block / programming logic block to the intelligent controller; Step S2: After receiving the loading notification of the algorithm block / sub-application block / programming logic block, the intelligent controller downloads the target function block from the host computer / business platform via FTP, and performs integrity verification on the algorithm block / sub-application block / programming logic block after successful download; Step S3: the intelligent controller loads the algorithm block / sub-application block / programming logic block and obtains the registration information of the target function block; Step S4: the intelligent controller loads the configuration parameters of the target function block and configures the target function block; Step S5: the intelligent controller allocates resources to the target function block according to the function of the target function block, and subscribes to the operation status data of the corresponding device model or the results of other function blocks according to the configuration parameters of the target function block; Step S6: When the service management layer receives the running status data of the device model or the result of other functional blocks, and the result changes, the service management layer submits the running status data of the device model or the result of other functional blocks to the target functional block; Step S7: The target function block performs calculations based on the received data and outputs the results to the service management layer; Step S8: After receiving the output result of the target function block, the service management layer performs corresponding processing.

4. A hot-swap method for an AI model based on a building intelligent controller, characterized in that: The AI ​​model hot-swap system based on a building intelligent controller and the AI ​​model hot-swap method based on a building intelligent controller applied to claim 1 include a hot-swap step, which is specifically implemented as the following steps: Step K1: The host computer / business platform sends a target function block uninstallation instruction to the intelligent controller; Step K2: the intelligent controller notifies other function blocks related to the target function block of the message that the target function block stops running; Step K3: the intelligent controller stops pushing data to the target function block; Step K4: the intelligent controller stops the operation of the target function block; Step K5: The intelligent controller logs out and deletes the target function block; Step K6: The intelligent controller releases the resources allocated to the target function block.

5. An electronic device, characterized in that: include: A processor, and a memory storing executable instructions of the processor; wherein: The processor is configured to perform the steps of the AI ​​model hot-swap method based on a building intelligent controller of any one of claims 2-4 by executing executable instructions.

6. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the AI ​​model hot-swap method based on a building intelligent controller of any one of claims 2 to 4 are implemented.

Images