Central air conditioning wind system control method based on large language model and digital twinning

Abstract

The application discloses a central air-conditioning wind system control method based on a large language model and digital twinning, which comprises the following steps: receiving a control instruction issued by a user, and preprocessing the control instruction matched successfully; judging whether the processing instruction indicates device information and whether the device is available and the type of the device to be controlled after judging the user type; determining the control variable name and the corresponding control variable set value, inquiring the running state of each device according to the inspection level, changing the control variable set value according to the control strategy, simultaneously performing simulation, feeding back the analysis result of the control effect to the user by the large language model, and transmitting the control strategy meeting the user demand to the field edge controller to execute the control strategy. The method of the application marks that the building heating ventilation and air conditioning system management has made an important step towards the direction of more intelligent and information-based, and exhibits great potential and application value of the intelligent control method in the field of building technology.

Description

Technical Field

[0001] This invention belongs to the field of automatic control of building heating, ventilation and air conditioning systems, and particularly relates to a control method for central air conditioning air systems based on large language models and digital twins. Background Technology

[0002] In the field of automatic control of building heating, ventilation, and air conditioning (HVAC) systems, efficient management of HVAC air systems is key to achieving energy savings and improving indoor environmental comfort. In traditional HVAC air system control, end users often cannot directly and accurately express their needs. Whether adjusting temperature and humidity based on specific environmental conditions or personal comfort preferences, existing system control logic and interfaces often appear rigid and imprecise. The lack of dynamic response to changes in outdoor temperature and the provision of predictive comfort service suggestions significantly limits the user experience. Furthermore, user interaction with the system largely relies on traditional physical interfaces or simple remote control devices, lacking convenience and intuitiveness, making remote control and personalized settings more difficult. Traditional equipment control strategies mainly rely on preset logic or algorithms and manual intervention. For maintenance personnel, maintaining and adjusting HVAC system control strategies presents equally significant challenges. Although methods such as PID control are widely used in the air conditioning field, their limitations in real-time correction and anti-interference capabilities often force maintenance personnel to manually intervene in control and debugging, even redesigning control strategies and algorithm parameters. This not only increases the requirements for maintenance skills but also significantly increases the complexity and labor intensity of maintenance. With the increasing informatization and intelligence of the industry, the operation and maintenance management of building systems faces the challenge of transforming from traditional models to intelligent and automated ones.

[0003] Large language models, such as GPT (Generative Pretrained Transformer) and BERT (Bidirectional Encoder-Representation Transformer), are natural language processing tools based on deep learning technology. They are capable of understanding, generating, and translating human natural language. Through pre-training on large amounts of text data, these models learn the grammatical structure, contextual meaning, and complexity of language, enabling them to understand user commands and generate corresponding responses. Applying these models to HVAC digital twin systems allows users to communicate with the system using their everyday language. The system understands the specific content and underlying intent of the commands, achieving precise control and adjustments. For operations and maintenance personnel, this means they can interact with the system more intuitively, adjusting and verifying control strategies, algorithm parameters, and troubleshooting without needing in-depth knowledge of complex programming languages ​​or control protocols, greatly improving work efficiency and system maintenance convenience. Summary of the Invention

[0004] The purpose of this invention is to solve the problems existing in the prior art and to provide a central air conditioning air system control method based on large language model and digital twin.

[0005] To achieve the above-mentioned objectives, the present invention specifically adopts the following technical solution:

[0006] A control method for a central air conditioning system based on a large language model and digital twin, comprising the following steps:

[0007] S1. The input module receives the control command issued by the user, matches the control command with the pre-built keyword library, and determines whether the control command contains the keywords in the keyword library: if so, the control command is preprocessed to obtain a new control command; otherwise, the user is prompted that the control command entered does not belong to the building energy field and needs to be re-entered.

[0008] S2. Determine the user type based on the permission level of the control command: If the user is not an operations and maintenance personnel, automatically add the thermal partition where the user is located to the new control command based on the address where the control command was sent, and use the new control command with the added thermal partition as the processing command; otherwise, there is no need to add the thermal partition, and the new control command is used as the processing command.

[0009] After determining the user type, the processing command is input to the large language model. The system then checks if the necessary information is specified in the processing command. If not, the user is prompted on the interactive interface to supplement the device information and re-enter the control command. If the necessary information is specified, the system retrieves the equipment component list for the thermal zone from the building air conditioning system in the digital twin model. Based on the necessary information in the processing command, the system matches the proposed control device from the digital twin model in the equipment component list. If the proposed control device is unusable, the user is prompted on the interactive interface to reconfirm the proposed control device and re-enter the control command. The necessary information includes the proposed control device type and device number, or the proposed control device type and thermal zone.

[0010] S3. Determine the type of device the user intends to control based on the processing command: If the type is the entire device, then execute S4; if the type is a device component and the user type is not an operations and maintenance personnel, then prompt the user on the interactive interface that the user's permissions are insufficient and that the control command needs to be re-entered; if the type is a device component and the user type is an operations and maintenance personnel, then call the building air conditioning system in the digital twin model system to obtain the list of control components of the device to be controlled. If the processing command and the list of control components match, then determine the name of the control component from the list of control components. If the processing command and the list of control components do not match, then prompt the user on the interactive interface that the control component of the device to be controlled does not exist and that the control command needs to be re-entered.

[0011] S4. The large language model determines whether the processing instruction matches the control command type of the device to be controlled: if they do not match, the user is prompted to reconfirm the control method and re-enter the control instruction on the interactive interface; if they match, the automatic control variable list of each component of the device to be controlled is obtained by calling the building air conditioning system; if the processing instruction and the automatic control variable list match, the control variable name is determined from the automatic control variable list, and the control variable setting value corresponding to the control command type is output; if the processing instruction and the automatic control variable list do not match, the user is prompted to reconfirm the control variables of the device to be controlled and re-enter the control instruction on the interactive interface.

[0012] S5. Obtain the name of the device to be controlled from the building air conditioning system, and obtain the topology connection relationship of the components to be controlled; sort the components in the topology connection relationship according to the pre-built inspection level of each component, and obtain the sorted component list; query the sorted component list in turn. For the current component in the sorted component list, call the building air conditioning system to first obtain the current component's operating status. If the current component's operating status meets the requirements for executing processing instructions and setting control variables, then query the next component until all components to be controlled are traversed. Otherwise, call the building air conditioning system to adjust the current component's operating status until the current component's operating status meets the requirements for executing processing instructions and setting control variables.

[0013] S6. Obtain the control strategy of the building air conditioning system corresponding to the processing command, input the control strategy into the building air conditioning system, change the set value of the control variable, obtain the performance data of each device of the building air conditioning system before the change, and perform digital twin model simulation on the building air conditioning system after the change to obtain the performance data of each device of the building air conditioning system after the change. Input the performance data before the change and the performance data after the change into the large language model to obtain the analysis results of the control effect and feed them back to the user.

[0014] S7. After receiving feedback from the user regarding the analysis results, input the user's feedback results into the large language model. The large language model determines whether the control strategy meets the user's needs. If it does, the modified control variable settings are saved to the historical information database, and the control strategy that meets the user's needs is passed to the field edge controller of the building air conditioning system. The field edge controller then executes the control strategy that meets the user's needs. Otherwise, the control command is re-entered.

[0015] Based on the above scheme, each step can be implemented in the following preferred manner.

[0016] Preferably, in step S1, the input module is capable of receiving various forms of control command input and has the ability to convert speech into text; the preprocessing includes correcting grammatical errors in the control commands, extracting key information points from the control commands, and eliminating ambiguity in control commands with multiple interpretations by using contextual information or requesting more information from the user.

[0017] Preferably, in step S4, if the control command type is an "on" command, the control variable setting value is 1; if the control command type is a "off" command, the control variable setting value is 0; if the control command type is a parameter adjustment command, the large language model uses the setting value in the processing instruction as the control variable setting value; if the processing instruction does not have a setting value, the retrieval model searches for historical setting values ​​of the automatic control variable in the historical information database based on the current operating status of the building's air conditioning system, and uses the historical setting value closest to the value as the control variable setting value; if the control command type is a control mode adjustment command, the corresponding control variable setting value is the air handling unit operating mode specified in the processing instruction; if not specified, the user is prompted on the interactive interface to reconfirm the air handling unit operating mode and re-enter the control instruction.

[0018] Preferably, in step S5, if the control command type is an open command, a control mode adjustment command, or a parameter adjustment command, the equipment order corresponding to the inspection level is: all types of duct valves, all types of equipment air valves, all types of equipment fans, and all types of cooling / heating coils; if the control command type is a close command, the equipment order corresponding to the inspection level is: all types of cooling / heating coils, all types of equipment fans, all types of equipment air valves, and all types of duct valves.

[0019] Preferably, the building air conditioning system also includes equipment modules that can dynamically update and maintain equipment information.

[0020] As a preferred option, the historical information database is used to collect and store historical control commands, analysis results corresponding to control strategies, and user feedback results.

[0021] Preferably, the building air conditioning system supports remote control operation, allowing users or maintenance personnel to remotely monitor and control the system via network connection.

[0022] As a preferred option, when acquiring control strategies, the control strategies can be adjusted according to the real-time operating status of each device in the building's air conditioning system, or users can select control strategies based on the control effect, and the control strategies can be automatically assessed for risk.

[0023] As a preferred option, a self-learning mechanism allows the building air conditioning system to automatically optimize control strategies and adjust the understanding and generation logic of the large language model based on data in the historical information database.

[0024] Preferably, the edge controller is deployed on the building site, responsible for executing control strategies and directly managing and controlling the building's air conditioning system equipment on site. The edge controller supports real-time data acquisition and analysis, and can instantly feed back the operating status of each piece of equipment in the building's air conditioning system to the digital twin platform, providing a basis for the generation and optimization of control strategies. Through a secure communication protocol established between the digital twin platform and the edge controller, the secure transmission and execution of processing instructions are ensured, preventing data tampering and unauthorized access.

[0025] Compared with the prior art, the present invention brings significant beneficial effects and technological progress:

[0026] First, by introducing large language models and digital twin technology, this invention significantly improves the intelligence and efficiency of building air conditioning system control. Users and maintenance personnel do not need in-depth professional knowledge to interact directly with the system through natural language, greatly simplifying the operation process and lowering the professional threshold for user operation. This method not only improves the user experience but also reduces reliance on professional technicians.

[0027] Secondly, this invention ensures accurate execution of control commands and safe system operation by precisely parsing user instructions and monitoring the building's ventilation system status in real time. This advanced control logic and status monitoring mechanism overcomes the limitations of traditional control systems in equipment identification, status judgment, and control strategy execution, providing more accurate and reliable control results.

[0028] Furthermore, through simulation analysis and optimization using digital twin models, this invention can predict the implementation effects of control strategies and methods, providing a scientific basis for their adjustment and optimization. This not only improves the system's control accuracy but also enhances its adaptability to complex environmental changes. Finally, this invention provides an innovative intelligent control solution for central air conditioning systems in the field of modern smart buildings. It not only meets current needs for intelligent air conditioning control and operation and maintenance but also has broad application prospects and promotional value, opening up new paths for the intelligent control and management of building energy systems. These features make this invention a significant technological advancement in improving energy efficiency, reducing operation and maintenance costs, and enhancing user experience. Attached Figure Description

[0029] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0030] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Many specific details are set forth in the following description to provide a thorough understanding of the present invention. However, the present invention can be practiced in many other ways different from those described herein, and those skilled in the art can make similar modifications without departing from the spirit of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below. Technical features in the various embodiments of the present invention can be combined accordingly without mutual conflict.

[0031] This invention aims to propose an innovative intelligent control method for building HVAC systems. This method combines the advanced understanding capabilities of large language models with the simulation and prediction capabilities of digital twin technology. It can not only accurately analyze the control needs of room users or maintenance personnel, and automatically identify relevant equipment, their control variables, and algorithm parameters, but also pre-verify the effectiveness and security of control strategies or algorithms in a virtual environment. Furthermore, by supporting multimodal data input, such as text and voice, this method further optimizes the interaction process between users and maintenance personnel and the system, improving the flexibility and accuracy of the control logic. Finally, through the collaborative work of a cloud platform and edge controllers, efficient execution of control strategies is achieved, overcoming traditional challenges in user interaction and maintenance, significantly reducing operational risks and the requirements for maintenance personnel, improving system operating efficiency and security, and promoting the development of HVAC system control towards intelligence and automation, thus opening up new avenues for technological innovation and application.

[0032] This invention aims to achieve intelligent control of central air conditioning systems by integrating large language models and digital twin technology, thereby improving the system's automation level and operational efficiency while ensuring the accuracy and security of the control strategy. Figure 1 As shown, in a preferred embodiment of the present invention, the above-mentioned central air conditioning air system control method based on large language model and digital twin includes the following steps S1 to S7. The specific implementation process of each step will be described in detail below.

[0033] S1. The input module receives the control command issued by the user, matches the control command with the pre-built keyword library, and determines whether the control command contains keywords from the keyword library: if so, the control command is preprocessed to obtain a new control command; otherwise, the user is prompted that the control command entered does not belong to the building energy field and needs to be re-entered.

[0034] It should be noted that in step S1 of the present invention, the input module is capable of receiving various forms of control command input and has the ability to convert speech into text.

[0035] In this embodiment, the input module receives control commands issued by the user in natural language. These commands can relate to various aspects of the corresponding building energy system, including building control, air handling unit control, operation, and maintenance. These commands can take various forms, such as text input, voice input, and other input methods. Specifically, the user can submit commands via text or voice. For text input, the user can directly input text using input devices such as a keyboard. For users who wish to ask questions via voice, the input module will use speech-to-text functionality to convert the voice information into text. The module is also designed to process multimodal information and has corresponding conversion capabilities to handle other forms of input besides text and voice, further ensuring that different forms of input can be accurately understood and processed, such as gestures or image input (if applicable), and converting this information into a format that can be understood and processed.

[0036] In this embodiment, the speech-to-text module of the input module uses the Wav2Vec 2.0 model, which learns representations directly from the original speech signal using unsupervised pre-training. This process can be represented as follows:

[0037] Encoded Representation=Wav2Vec2.0(Raw Audio)

[0038] Wherein, Encoded Representation represents the encoded representation, Raw Audio represents the original speech signal, and Wav2Vec2.0(·) represents the Wav2Vec 2.0 model.

[0039] In this embodiment, if the user input includes images and videos, the input module's processing model uses the ViLT model, which understands multimodal content by fusing images (or video frames) and text. The formula for the ViLT model to process images and text can be simplified as follows:

[0040] Unified Representation=ViLT(Image Embeddings,Text Embeddings)

[0041] Here, Unified Representation refers to synthetic representation, Image Embeddings refers to image embedding, and Text Embeddings refers to text embedding.

[0042] It should be noted that in step S1 of this invention, after receiving the user's control command, the input module matches the control command with a pre-built keyword library, using a keyword-based retrieval method to analyze the control command issued by the user through natural language. The input module sets a series of keywords related to building energy systems and forms a keyword library. When receiving the user's input control command, it first performs a rapid scan to identify these keywords. If the user's question does not contain these preset keywords, the question will be marked as currently unsolvable because it does not meet the query parameters of the building energy system, and the user will be prompted that the input control command does not belong to the building energy field. For example, when a user asks "1+1=?", it will be considered that the question does not belong to the building energy field. This keyword matching mechanism ensures that the response is focused on highly relevant questions, thereby improving the overall processing efficiency and the relevance of the response. In this way, the method of this invention can more effectively manage user inquiries and ensure that queries to the building energy information system receive timely and accurate responses. In this process, a keyword matching model is used, that is, selection and retrieval based on key variables and variables in the system, such as text-embedding-3-small. For example, if a user needs to issue the control command "X1", the keyword matching model can load and read the text, segment and vectorize the text, match the n most similar words to the question vector in the text vector, and add the matched text as context and question together to X1.

[0043] It should be noted that in step S1 of the present invention, the above preprocessing includes correcting syntax errors in the control instructions, extracting key information points from the control instructions, and eliminating ambiguity in the control instructions for those with multiple interpretations by using contextual information or requesting more information from the user.

[0044] In this embodiment, after receiving the user's control command, the input module performs a series of preprocessing operations to improve the efficiency and accuracy of control command processing. These operations include, but are not limited to, grammar correction, keyword extraction, and disambiguation, to enhance the efficiency and accuracy of question parsing. Specifically, grammar correction automatically corrects grammatical errors in the control command to reduce misunderstandings caused by inaccurate language expression. Keyword extraction extracts key information points from the control command to help the system more quickly locate the user's query intent. Disambiguation involves the input module eliminating ambiguity for expressions that may have multiple interpretations, either through contextual information or by requesting more information from the user, to ensure the control command is correctly understood.

[0045] In this embodiment, the grammar correction and context understanding sections utilize language models and the context-aware model BERT to correct grammatical errors in the text and understand the context of the user's query. These models can infer the most appropriate word replacements or rearrangements based on the context to correct grammatical errors. Furthermore, these models learn deep-level language features on large amounts of text data through pre-training, enabling them to extract contextual information relevant to the query, thereby improving the accuracy of query parsing.

[0046] Through the detailed implementation steps described above, the input module can effectively receive and process control commands issued by users in natural language. Regardless of the form in which the user issues the control commands, the input module can accurately understand and respond, thereby providing users with accurate and efficient services.

[0047] S2. Determine the user type based on the permission level of the control command: If the user is not an operations and maintenance personnel, automatically add the thermal partition where the user is located to the new control command based on the address where the control command was sent, and use the new control command with the added thermal partition as the processing command; otherwise, there is no need to add the thermal partition, and the new control command is used as the processing command.

[0048] After determining the user type, the processing command is input to the large language model. The system then checks if the necessary information is specified in the processing command. If not, the user is prompted on the interactive interface to supplement the device information and re-enter the control command. If the necessary information is specified, the system retrieves the device component list for the thermal zone from the building air conditioning system in the digital twin model. Based on the necessary information in the processing command, the system matches the proposed control device from the digital twin model in the device component list. If the proposed control device is unusable, the user is prompted on the interactive interface to reconfirm the proposed control device and re-enter the control command. The necessary information includes the proposed control device type and device number, or the proposed control device type and thermal zone.

[0049] It should be noted that in step S2 and subsequent steps of using a large language model in this invention, the method of this invention is not limited to using a specific large language model. It can adapt to and integrate multiple large language models, including but not limited to open-source and commercial models, to improve the adaptability and accuracy of the method. In step S2 and subsequent steps of using a large language model in this embodiment, the CHATGLM model is used as the large language model, and the transmission process of questions and responses is realized by calling the API of the corresponding model. CHATGLM (Chat General Language Model) is a general language model that, through training on a large amount of text data, can understand and generate natural language, primarily in Chinese and English usage environments. The CHATGLM model can perform deep analysis of user questions, identifying key variables and information needs in the questions. Specifically, by analyzing the semantic and grammatical structure of user questions, the CHATGLM model can determine the topic and intent of the questions, and thus identify key variables related to the questions. At the same time, the CHATGLM model can also understand the user's information needs based on the context of the questions, thereby providing more accurate and relevant answers.

[0050] It should be noted that in step S2 of this invention, it is necessary to determine the intended control device for the control command. After receiving the control command request from the user via natural language, the system first determines whether the user's thermal zone or the processing command explicitly specifies the type and number of the intended control device, and performs initial parsing to determine the user's intended control device. For end-user control commands, the device can be omitted, and the default intended control device is the corresponding unit responsible for that zone. For operation and maintenance personnel, the control command must explicitly specify the type and number of the intended control device. If the user is an operation and maintenance personnel and the command processing lacks the necessary information such as the type or number of the intended control device, or if there are situations such as the device being unidentifiable or the command being ambiguous, the system interacts with the user and guides them to provide a complete control command again, requiring them to supplement the missing information and provide the control command again, before returning to S1. Next, the system calls the API interface connected to the digital twin model system to obtain a list of all usable equipment components in the current central air conditioning system, including the air handling units corresponding to the current thermal zones and hot zones of the building. Using this list, the system matches the equipment in the user's zone or the equipment information mentioned in the command to select the specific intended control device. Specifically, the large language model combines the relevant components or numbers of the air handling unit specified in the user's instructions to select the equipment components to be controlled in the twin system. It checks whether the components exist in the current thermal zone's HVAC air system, whether the instructions are ambiguous, and whether the selected equipment is allowed to be used. If the component does not exist or is unusable, it interacts with the user, indicating that the intended control equipment or component does not exist or is unusable, requesting reconfirmation of the intended control equipment and a new control instruction.

[0051] It should be noted that in step S2 of this invention, before executing the control command, the system performs a step to determine the user's identity and permission level. This process includes not only verifying whether the user has the permission to execute a specific task, but also identifying the thermal partition where the user is located, and whether they are an end user or a system maintenance personnel. This determination is crucial for ensuring system security and preventing unauthorized access and operation.

[0052] It should be noted that in step S2 of this invention, after the system obtains the thermal zone of the end user, it will query the specific air system equipment corresponding to that zone. The program calls the API interface connected to the digital twin model system to obtain the central air conditioning air system equipment corresponding to the current hot zone. In specific implementation, this is added and executed in the specific project script. For example, `result = platform.get_zone_model("Hot Zone 1")` will return information such as the name of the specific main equipment, such as "Air Handling Unit_1". When there are multiple main equipment in a zone, such as both air handling units and fan coil units, the user needs to specify the specific equipment. The system will interact with the user again, clearly point out the problem, and guide the user to provide control commands again.

[0053] It should be noted that in step S2 of this invention, before executing the control command, the system verifies the real-time status of the device to be controlled, including its existence and availability. This step ensures the effective execution of the control command, avoids control failure due to the unavailability of the target device, and thus improves the safety and reliability of the operation. Specifically, the program uniformly calls the API interface connected to the digital twin model system to obtain the list of equipment components of the current building air conditioning system. For example, by adding and executing the script in a specific project, result = platform.get_all_model() will return information: a set of names, such as ["Air Handling Unit_1", "Air Handling Unit_2"] etc. The equipment deployment list is the list of all existing and usable equipment in the current building air conditioning system, which the large language model uses to select the device to be controlled in the digital twin model system by combining the equipment component list with the equipment information (type and number of the device to be controlled) specified in the processing command. During this process, the system considers the potential ambiguity in the device names provided in the user's control commands. For example, if the device name in the system is "Single Duct Air Handling Unit_1," a user in the corresponding hot zone could use "Air Conditioner No. 1" or even "Air Conditioner" in the command. Maintenance personnel need to specify the device to be controlled, down to the specific device or room. When determining the device to be controlled, the system also checks the availability of the selected device to ensure it exists in the current air conditioning system and is usable. If the device to be controlled does not exist in the current building's air conditioning system or is unusable, the system interacts with the user again, clearly stating the problem: the device to be controlled does not exist or is unusable. The user is asked to reconfirm the device and provide the control command again, returning to S1.

[0054] In this embodiment, when the control commands provided by the user are incomplete or contain errors, the system will guide the user to supplement or correct the information through interactive dialogue. This mechanism not only optimizes the user experience but also improves the success rate of control command execution. For example, if the control device specified in the user's control command does not exist, the system will provide feedback to the user: "The device you want to control does not exist or is unusable. Please reconfirm the device you wish to control and provide the control command again." By automatically adding and supplementing the corresponding explanatory text in each interaction round, the system assists the large language model in more accurately judging the user's intent and needs, improving the accuracy of command parsing and the quality of response. For example, in this step, the example of a "promote" passed to the large language model is:

[0055] 1. "Can you analyze from the following text which specific device the user wants to control (e.g., Air Handling Unit 1)? If a corresponding device and number exist, answer 'YES'; otherwise, answer 'NO'. For example, if the user's question is: 'User's question: Turn on Air Handling Unit 1', then answer 'YSE'. The user's question is: {Question raised by the user}"

[0056] 2. "Please analyze the following text to determine which device the user wants to control. Please follow the format: 'Controlled Device:', e.g., 'Controlled Device: Air Handling Unit 1'. The user's question is: {User's Question}"

[0057] 3. "This is the system's equipment list" + Equipment List + "These are the devices the user needs to control:" + Device to be controlled + "Please select the devices the user needs to control from the system's equipment list. If no matching device is found, answer 'NO'. The numbers _1, 2, 3 after the device name are the device number, such as Air Handling Unit _1 referring to Air Handling Unit No. 1. Note! Do not change the device names in the following operations! The devices must be selected from the equipment list, not fabricated! Please select the relevant devices according to the user's needs and output them directly, such as: 'Air Handling Unit _1', or 'NO'.

[0058] It should also be noted that the equipment scope of the aforementioned building air conditioning system includes, but is not limited to, various types of air handling units and fan coil units, such as single-duct and dual-duct air handling units, as well as variable air volume and constant temperature and humidity air handling units. When determining the equipment to be controlled, the system will also check the availability of the selected equipment to ensure that it exists in the current air conditioning system and is usable. If the selected equipment does not exist or is unusable, the system will interact with the user again, clearly pointing out the problem and guiding the user to provide control commands again. The specific equipment used depends mainly on the composition of the central air conditioning system equipment in the actual project and is not limited in this invention.

[0059] As can be seen, by calling and reading the digital twin system, a list of currently available devices can be dynamically generated and provided. Based on real-time data from the digital twin model, accurate device options are provided for user commands. When a device is unavailable or does not exist, the system will proactively interact with the user to ensure the accuracy of commands and the effectiveness of their execution.

[0060] S3. Determine the type of device the user intends to control based on the processing command: If the type is the entire device, proceed to S4; if the type is a device component and the user is not an operations and maintenance personnel, prompt the user on the interactive interface that their permissions are insufficient and they need to re-enter the control command; if the type is a device component and the user is an operations and maintenance personnel, retrieve the control component list of the device to be controlled from the building air conditioning system in the digital twin model system. If the processing command and the control component list match, determine the control component name from the control component list; if the processing command and the control component list do not match, prompt the user on the interactive interface that the control component of the device to be controlled does not exist and they need to re-enter the control command.

[0061] It should be noted that step S3 of this invention involves determining and identifying the component to be controlled by the control command. Upon receiving a control request from a user via natural language, the system's primary task is to determine the intent of the control command and whether it explicitly specifies the corresponding component of the device to be controlled. If the user's request pertains to the entire device, such as turning an air handling unit on or off, then the component to be controlled will be defined as the entire air handling unit. Conversely, if the user's request points to a specific component within the air handling unit, such as adjusting the cooling coil or the blower, then the system will operate on that specific component. For maintenance personnel, their commands need to be more detailed and specific, including the corresponding component of the device to be controlled. If this crucial information is missing, the system will communicate with the user through an interactive dialogue, requesting the missing information and guiding the user to provide a complete control command again, thus returning to the beginning of the process, i.e., returning to S1.

[0062] Next, the system retrieves a list of all available control components corresponding to the device to be controlled in the current building by calling the API interface connected to the digital twin model system. This step is crucial because it provides the system with the necessary information to determine whether the control components mentioned in the user's command exist and whether they can be controlled. In specific implementation, this is done by adding and executing the script in the specific project, for example: `result = platform.get_zone_model("Air Handling Unit_1")` will return information: a list of components, such as `[Air Handling Unit_1, Supply Fan, Supply Air Valve, Return Air Fan, Return Air Valve, Fresh Air Valve, Heating Coil, Cooling Coil, ...]`, etc.

[0063] Finally, the system will generate and print a list of all existing and usable control components in the device to be controlled. This list will be used by the large language model to select the necessary control components in the twin system based on the user's instructions regarding the devices to be controlled. This provides a basis for subsequent operations. The system will check key factors such as whether the selected component exists in the device to be controlled, whether the instruction is ambiguous, and whether the selected component is allowed to be controlled, based on the user's instructions. If the selected component is found to be missing or unusable, the system will interact with the user again, clearly pointing out the problem and guiding the user to confirm the device to be controlled and provide the control instructions again.

[0064] It should be noted that the system allows for a degree of ambiguity in the naming of equipment components. For example, if the equipment component name in the digital twin system is "Single Duct Air Handling Unit_1_Supply Air Fan," then the user can simply refer to this component as "Supply Air Fan." This flexibility greatly improves the user-friendliness and usability of the system. The digital twin system can identify and include various air handling unit components commonly found in buildings, such as cooling / heating coils, various valves, fans, filters, etc. The specific form and quantity of these components depend on the actual system configuration. For example, if the equipment component name in the digital twin system is "Single Duct Air Handling Unit_1_Supply Air Fan," the user in the corresponding hot zone can use "Supply Air Fan" in the component section of the command; similarly, if the equipment component name in the system is "Room_1 Supply Air Terminal Valve," the user in the corresponding hot zone can use "Terminal Valve" in the command.

[0065] It should be noted that when the control commands provided by the user are incomplete or contain errors, the system will guide the user to supplement or correct the information through interactive dialogue. This mechanism not only optimizes the user experience but also improves the success rate of control command execution. For example, if the control device specified in the user's control command does not exist, the system will provide feedback to the user: "The device component you want to control does not exist or is unusable. Please reconfirm the device component you wish to control and provide the control command again." For example, the example of the `promot` passed to the large language model in this step is:

[0066] 1. "Can you analyze from the following text which specific component of the equipment the user wants to control, including the entire system (e.g., the air supply valve)? If a corresponding component exists, answer 'YES'; otherwise, answer 'NO'. For example, 'User question: Open the air supply valve of air handling unit 1', then answer 'YSE'. The user question is: {Question raised by the user}"

[0067] 2. "Please analyze from the following text which component of the device the user wants to control, including the entire system? Please follow the format: 'Controlled device component:' e.g., 'Controlled device component: air supply valve'. The user's question is: {User's Question}"

[0068] 3. "This is the system's equipment component list" + Component List + "These are the equipment components the user needs to control:" + Device Component to be Controlled + "Please select the equipment components the user needs to control from the system's equipment list. If no matching equipment is found, answer 'NO'. The numbers _1, 2, 3 after the equipment component name are the equipment number.\nNote! Do not change the equipment component names in the following operations! Equipment components must be selected from the equipment component list, not fabricated! Please select the relevant equipment components according to the user's needs and output them directly, such as: 'Air Handling Unit_1_Air Supply Valve', or 'NO'.

[0069] It should be noted that, for the digital twin system of the present invention, the equipment components in the building may include relevant components commonly found in various types of air handling units, such as cooling / heating coils, various valves, fans, filters, etc., the specific form and quantity of which depend on the actual system.

[0070] S4. The large language model determines whether the processing instruction matches the control command type of the device to be controlled: If they do not match, the user is prompted to reconfirm the control method and re-enter the control instruction on the interactive interface; if they match, the system calls the building air conditioning system to obtain the list of automatic control variables for each component of the device to be controlled; if the processing instruction and the list of automatic control variables match, the control variable name is determined from the list of automatic control variables, and the control variable set value corresponding to the control command type is output; if the processing instruction and the list of automatic control variables do not match, the user is prompted to reconfirm the control variables of the device to be controlled and re-enter the control instruction on the interactive interface.

[0071] It should be noted that in step S4 of this invention, it is necessary to confirm the name of the control variable corresponding to the processing instruction and its set value. First, based on the component to be controlled, it is determined whether the control command of the user control instruction is clear, unambiguous, and executable. Second, based on the self-control variables of the component to be controlled, the name of the control variable of the user control instruction is determined. Then, the set value of the control variable is determined. After receiving the user's control instruction, the method program immediately performs a preliminary analysis. The purpose of this stage is to determine whether the instruction contains a clear, unambiguous, and executable control command, and at the same time, based on the characteristics of the target control device (such as a building air conditioning system), to confirm its acceptable command type. These types usually include on / off, control mode adjustment, and parameter adjustment commands. The large language model analyzes the user instruction and the target control device. If the instruction does not belong to any corresponding control method, the system interacts with the user, indicating that the target control device component does not have a corresponding control method, requesting reconfirmation of the control method and re-providing the control instruction, and returning to S1 to ensure that the control command type of the target control device component is clear and unambiguous. Specifically, when interacting with the user, if the processing instruction fails to clearly match any of the above control command types, the system will consider the instruction to be ambiguous. In this situation, the system will proactively interact with the user, guiding the user to reconfirm the control method and return to S1 to provide new control commands. This may involve returning to the initial stage of receiving control commands to ensure that an accurate and clear control command is obtained.

[0072] Once the control command type of the device to be controlled by the processing instruction is successfully matched, the name of the control variable in the processing instruction is determined by combining the list of automatic control variables of the device to be controlled, and then the set value of the control variable is judged and determined. Specifically, the program prints out the list of automatic control variables of the current target controlled device component to help confirm the control variables involved in the user instruction. In this process, the large language model intelligently identifies and determines the most matching control variable name based on the ambiguous description provided by the user and the list of automatic control variables of the device. The program calls the API interface connected to the digital twin model system to obtain the list of all changeable automatic control variables in the current device to be controlled. This is the corresponding functional module of the digital twin platform. In specific implementation, for example, adding and executing `result = models.get_variables_by_type("Air Handling Unit_1_Air Supply Valve","Automatic")` in a specific project script will return information: a set of names, such as ["Remote / Local", "On / Off", "Air Supply Valve Opening"], etc. At the same time, the large language model can understand the abstract intent in the user instruction, allowing users' questions to have a certain degree of ambiguity.

[0073] Finally, the method determines the setpoint of the control variable, based on the type of control command. In this invention, the setpoint is determined according to the type of control command. Specifically, if the control command type is an "on" command, the setpoint of the control variable is 1; if the control command type is a "off" command, the setpoint of the control variable is 0. If the control command type is a parameter adjustment command, the large language model uses the setpoint value in the processing instruction as the setpoint of the control variable. If the processing instruction does not contain a setpoint value, the retrieval model searches for historical setpoints of the automatic control variable in the historical information database based on the current operating status of the building's air conditioning system, and uses the closest historical setpoint as the setpoint of the control variable. If the control command type is a control mode adjustment command, the corresponding setpoint of the control variable is the air handling unit operating mode specified in the processing instruction. If not specified, the user is prompted on the interactive interface to reconfirm the air handling unit operating mode and re-enter the control command.

[0074] In this embodiment, for the control mode adjustment command, the aforementioned air handling unit operating modes, such as summer cooling, winter heating, and transitional season energy-saving control mode, will automatically set the control variable setpoints to a pre-built set of values ​​when switching between these operating modes. If the user does not explicitly specify which operating mode to change to, interaction with the user is required to clarify which mode the air handling unit needs to operate in.

[0075] In this embodiment, for parameter adjustment commands, the system will further analyze the user's instructions to determine the specific setting value. If the user does not explicitly provide a setting value, the system will use a retrieval model to match similar historical operating conditions in the historical information database, taking into account the common range of the self-controlled variables of the components to be controlled, the current air handling unit, and weather load information, to infer the most suitable setting value. Furthermore, parameter adjustment commands are not limited to the physical parameters of the equipment, such as the fan speed, but also include the algorithm parameters used by the control equipment, such as PID control parameters. The system will also evaluate the impact of parameter changes based on the simulation results of the digital twin system before and after the parameter changes.

[0076] It should be noted that in this embodiment, the retrieval model described above is the BGE-M3 model, which is the first text retrieval model with multi-functionality, multi-language, and multi-granularity characteristics. It efficiently supports multi-language, long text, and mixed retrieval, ensuring the accuracy and relevance of the answers. To improve the performance of a single retrieval mode, it includes a self-knowledge distillation function, selects an optimized batch processing strategy, and supports large batch sizes, which can be easily used when fine-tuning vectors for long text or large language models.

[0077] It should be noted that this invention uses an advanced large language model with Chinese as the primary language, and implements the transmission of questions and responses by calling the corresponding model's API. By automatically adding and supplementing relevant explanatory text in each interaction round, it assists the large language model in more accurately judging the user's intent and needs, thereby improving the accuracy of command parsing and the quality of response. For example, the `promot` passed to the large language model in this step is as follows:

[0078] 1. "Can you analyze the user's command type from the following text? Is it a device component activation, deactivation, control mode adjustment, or parameter adjustment? If it's a device activation command, answer '1'; if it's a control mode adjustment command, answer '3'; if it's a device parameter adjustment command, answer '4'; if it's none of these, answer 'NO'. For example: 'User question: Open the air supply valve of air handling unit 1', then answer '1'. User question: {User question}"

[0079] 2. "Device Component: {Device Component to be Controlled}, Device's Automatic Control Variables: {All Automatic Control Variables of the Device}. Can you analyze the user's command type from the following text and determine which automatic control variable of the device needs adjustment? Please do not change the device name in the following operations! The automatic control variables of the device component must be selected from the device list, not fabricated! Please select the relevant device according to the user's needs and output it directly.\nUser's Question: {User's Question}"

[0080] 3. "Equipment component: {The device component to be controlled}, the automatic control variables that need to be set for the device are: {The automatic control variables that need to be set for the device component}. Can you analyze from the following text what value the user's command needs to set for the automatic control variables? If a corresponding value exists, please answer directly with the corresponding number, such as "1"; if not, answer "NO".\nUser's question: {User's question}."

[0081] It should be noted that when the control commands provided by the user are incomplete or contain errors, the system will guide the user to supplement or correct the information through interactive dialogue. This mechanism not only optimizes the user experience but also improves the success rate of control command execution. For example, if the control device specified in the user's control command does not exist, the system will provide feedback to the user: "The device component you want to control does not have the corresponding automatic control variable. Please reconfirm the control method of the device component you wish to control and provide the control command again."

[0082] S5. Obtain the name of the device to be controlled from the building air conditioning system, and obtain the topological connection relationship of the components of the device to be controlled; sort the components in the topological connection relationship according to the pre-built inspection level of each component, and obtain a sorted component list; query the sorted component list in turn. For the current component in the sorted component list, call the building air conditioning system to first obtain the operating status of the current component. If the operating status of the current component meets the requirements of executing processing instructions and setting control variables, then query the next component until all components to be controlled have been traversed. Otherwise, call the building air conditioning system to adjust the operating status of the current component until the operating status of the current component meets the requirements of executing processing instructions and setting control variables.

[0083] It should be noted that step S5 of this invention is to check and confirm that the current status of the building's air conditioning system (air system) meets the control conditions. The specific implementation process of step S5 is briefly described below.

[0084] First, to ensure the effectiveness of control operations, the system must first check and confirm whether the current air system status meets the control conditions. This process begins by checking the status of the air valves in the user's room and its heated area. This step is crucial because it directly relates to the effectiveness of the airflow path and the operating status of the air system. Confirming that the air valves are open properly ensures that air can flow smoothly to the designated area, and also verifies the operating status of the air system, ensuring that the temperature control system can work normally according to the set parameters.

[0085] Secondly, it is necessary to confirm the topological connections between the device component to be controlled and its related air system components and valves. This includes all equipment and valves involved in the airflow and circulation process, ensuring the relevance and effectiveness of control commands. To this end, the system is designed with a program input mechanism that allows maintenance personnel to input the name of the device to be controlled, automatically querying and returning a list of related equipment components and pipe valves in the current air system that are topologically connected to that device, providing necessary information for subsequent status checks. Specifically, the program calls the API interface connected to the digital twin model system. The topological connections between the device to be controlled and its related air system components and pipe valves are the corresponding functional modules of the digital twin platform. In practical implementation, the topological connections of the component to be controlled consider all components involved in the air handling circulation process, as well as air valves, hot and cold water coils, etc. The program is designed to input the name of the component to be controlled to query the related components, air valves, hot and cold water coils in the current air system that are topologically connected to the component, and return a list. For example, adding and executing `result = models.get_related_equipments("Air Handling Unit_1_Supply Air Valve")` in a specific project script will return a collection of equipment names, including all equipment components and pipe valves involved in airflow and circulation.

[0086] Next, the inspection levels of equipment components and piping valves need to be prioritized. Based on the specific requirements of the control commands, the system prioritizes the equipment to be controlled and its related air system equipment components and piping valves according to their inspection levels. Depending on the control method, the operating status of connected components, air valves, and hot / cold water coils with higher priority than the component to be controlled is inspected sequentially using different standards. When an error is found in the operating status of a component, air valve, or hot / cold water coil, the program will actively control the component, air valve, or hot / cold water coil to open / close according to the pre-set control commands, thereby ensuring that the current air system status meets the control conditions.

[0087] Specifically, the inspection level ranking of equipment components and piping valves varies depending on the type of control command. Components and valves in the ventilation system are ranked according to the control method, based on the necessity of equipment and piping valves in the airflow and circulation circuit. For example, before starting a fan, the corresponding air valves must be ensured to be open to guarantee proper airflow. Then, the operating status of the hot and cold water coils is checked to ensure effective temperature regulation. Therefore, the control level ranking for open commands, control mode adjustment commands, and parameter adjustment commands is generally: all types of duct valves, all types of equipment valves, all types of equipment fans, and all types of cooling / heating coils. The control level ranking for close commands is generally: all types of cooling / heating coils, all types of equipment fans, all types of equipment valves, and all types of duct valves. For example, before executing an open or parameter adjustment command, it is necessary to ensure that the corresponding series of valves are open to allow the pump to start or its frequency to be adjusted.

[0088] In addition, for parameter adjustment commands, it is also necessary to ensure that the device to be controlled is turned on. Therefore, during the level sorting process, the device to be controlled will be added to the list of control devices to be checked in advance before the check is performed.

[0089] Finally, the control conditions need to be verified. Following the verification priority order of equipment components and piping valves, the system will sequentially verify the operating status of connected equipment components and piping valves with higher priority than the equipment component to be controlled. This process uses different standards depending on the control method to ensure that each step meets the conditions for executing control commands. If any problems are found in the operating status of equipment components or piping valves during the verification process, the system will proactively open / close the equipment or piping valves according to preset control commands and operate the corresponding parameter values ​​set according to the control mode, thereby ensuring that the entire ventilation system meets the control conditions.

[0090] During the inspection process, the system automatically executes a pre-defined program to check the status of devices in the device list. This mainly involves calling the corresponding interfaces of the digital twin platform. First, it obtains the current on / off status and specific settings of the device components. If any status does not meet the control conditions, it will automatically adjust. For example, for the "on" command to check the valve opening of Air Handling Unit_1_Supply Air Valve_1, the command is: result = models.get_var_value_by_modelname_and_varname("Air Handling Unit_1_Supply Air Valve_1","On / Off Status"). If the valve opening status of Supply Air Valve_1 is found to be 0, then Valve_1 should be opened to the pre-set default opening of 40%. This will be done using the command models.set_parameter_value("Air Handling Unit_1_Supply Air Valve_1","Valve Opening","0.4") to open the valve.

[0091] S6. Obtain the control strategy of the building air conditioning system corresponding to the processing command, input the control strategy into the building air conditioning system, change the set value of the control variable, obtain the performance data of each device of the building air conditioning system before the change, and perform digital twin model simulation on the building air conditioning system after the change to obtain the performance data of each device of the building air conditioning system after the change. Input the performance data before the change and the performance data after the change into the large language model to obtain the analysis results of the control effect and feed them back to the user.

[0092] It should be noted that in step S6 of this invention, the digital twin system automatically performs a risk assessment before executing control commands. This assessment evaluates the potential risks and impacts of command execution, including their effects on system stability and equipment safety, ensuring the security of the control commands. Furthermore, the digital twin system supports multi-level control strategy verification. By simulating the control effects under different operating conditions, it automatically selects the optimal control strategy to maximize the control effect.

[0093] It should be noted that step S6 of the present invention includes the automatic generation and execution of the control strategy, the simulation of the building air conditioning system, the acquisition of performance data before and after the execution of the control strategy, and the analysis of the control effect and feedback of the analysis results to the user. The specific implementation process of step S6 is briefly described below.

[0094] It should be noted that step S6 of this invention first involves the automatic generation and execution of the control strategy. After successfully collecting information on the device to be controlled and its corresponding control variables and setpoints, the system enters the automatic generation stage of the control strategy. At this point, the system utilizes advanced algorithms integrated with the digital twin model to automatically construct a control strategy that matches the user's control requirements based on the acquired input information (the determined device to be controlled, the determined control variables and setpoints, the thermal zone where the user is located, etc.) and executes the twin system modification. This process not only considers the operating parameters of the device but also integrates the current operating status of the system to ensure that the generated control strategy is both effective and practical.

[0095] Subsequently, the control strategy is directly applied to the digital twin model, enabling real-time modification and control of the twin system. The core of this step lies in translating the theoretical control strategy into practically executable operations. Through precise mapping of the digital twin model, it ensures that the designed control strategy can be pre-verified and executed in the virtual environment. This not only significantly reduces the risks that may be encountered during actual execution but also improves the efficiency and accuracy of the overall control process. While the control strategy is being executed, the system monitors the execution process in real time, ensuring that each operation is completed accurately and without error according to the predetermined strategy.

[0096] It should be noted that this invention supports real-time data synchronization and can adjust the control strategy according to the real-time status information of various devices such as air handling units and fan coil units in the building air conditioning system. It also supports multi-level control strategy verification. When acquiring the control strategy, the control strategy can be adjusted according to the real-time operating status of each device in the building air conditioning system. It can also simulate the control effect under different operating conditions, allowing users to select the control strategy based on the control effect and automatically conduct risk assessment of the control strategy to ensure the maximization of the control effect.

[0097] As can be seen, the automatic generation and execution process of the above control strategy fully demonstrates the application value of digital twin technology in intelligent control systems. Through the close integration of virtual and reality, the control process becomes more flexible and efficient, and can be dynamically adjusted based on real-time feedback. The main approach involves calling the corresponding interfaces of the digital twin platform and using the command `models.set_parameter_value("device", "automatic variable", "set value")` to control various devices and valves within the digital twin platform.

[0098] It should be noted that step S6 of this invention involves performing a digital twin model simulation of the building's air conditioning system. Once the control strategy is applied in the digital twin model, the method of this invention immediately initiates the simulation program. This process utilizes high-precision computational simulation to predict and analyze the actual impact of the control strategy on the air system under control operations, including equipment response, system status, room terminal temperature and humidity, and various possible scenarios, and waits for the simulation calculation to complete. Through this simulation, the system can predict the effect of the control strategy without risk, and identify and resolve potential problems in advance.

[0099] The performance data collected during the simulation provides empirical evidence for evaluating the effectiveness of the control strategy. Based on this data, the system will predict the performance of the air system and its control equipment components after the implementation of control operations, including energy efficiency, stability, and potential risks, as well as the temperature, humidity, and comfort at the room's terminals.

[0100] It is evident that the simulation execution steps of the digital twin model are a crucial link in intelligent control systems. They make the verification and optimization process of control strategies more scientific and precise, greatly improving the success rate of control scheme implementation and the overall performance of the system.

[0101] It should be noted that the digital twin system integrates a fault detection and diagnosis module, which can automatically detect and diagnose potential faults during the execution of control strategies, and provide timely fault recovery suggestions to reduce system downtime. This ensures that user-generated commands do not cause actual system malfunctions or even system paralysis.

[0102] It should be noted that in step S6 of this invention, the next step is to obtain key performance variable information (performance data) of the air system and control equipment before and after the control strategy is implemented. After the control strategy is implemented and the simulation is completed, the performance data acquisition stage begins. During this process, the interface of the digital twin model is called to record in detail the key performance variables of the central air conditioning air system and related control equipment components and room terminals before and after the implementation of the control strategy. These variables include, but are not limited to, parameters such as temperature, pressure, and flow rate, which are directly related to the system performance and the effectiveness evaluation of the control strategy or algorithm. Specifically, the program needs to call the values ​​of various variables in the digital twin model before and after the simulation. The specific variables are fixed as a corresponding list. Using the developed interface models.get_var_value_by_modelname_and_varname("equipment","variable"), the corresponding values ​​of the variables can be obtained. The result returned during the recording process is "equipment + variable: value".

[0103] By analyzing the changing trends of these key variables, the system can accurately assess the actual impact of control strategies on the performance of the ventilation system and the status of room terminals. This analysis process not only helps to understand the effect of each control operation but also provides data support for subsequent strategy or algorithm adjustments. Furthermore, the acquisition and analysis of this key variable information also provides valuable data resources for the long-term operation and maintenance and performance optimization of the system.

[0104] It is evident that the process of acquiring and analyzing key variable information demonstrates the powerful capabilities of digital twin models in real-time data processing and analysis, providing solid data support for intelligent control and ensuring the scientific nature and effectiveness of control decisions.

[0105] It should be noted that in step S6 of this invention, the large language model, combined with the user's control commands and the performance data of the terminal units and air handling units and components before and after the execution of the control strategy, analyzes the control effect and feeds back the analysis results to the user. Specifically, after collecting performance data (key variable information), the system performs an in-depth analysis of the control effect. This step utilizes the large language model to comprehensively consider the user's control commands and the acquired key variable information to fully evaluate the execution effect of the control strategy. During the analysis, the system will evaluate whether the control strategy or algorithm has achieved the expected goal, identify the causes of any deviations, and propose possible improvement measures. In this embodiment, the large language model will generate feedback information that is easy for the user to understand, clearly indicating the execution effect of the control strategy, existing problems, and suggested optimization directions. This timely and detailed feedback not only enhances the user's confidence in operating the system but also provides guidance and assistance for the user in actual operation. The analysis results will be displayed on the control interface and played back in voice format.

[0106] Furthermore, when using the large language model to obtain analysis results, the example of the prompt passed to the large language model in this process is: "This is the various system state information before control {various system state information before control}, this is the various system state information after control {various system state information after control}. Please summarize the changes to the user by combining the information before and after the changes."

[0107] As can be seen, the process of analyzing the control effect highlights the intelligent control system's ability to drive decision-making through data. Through precise analysis and feedback generation, even end users or maintenance personnel without a strong background in HVAC and automatic control can easily understand the control effects generated by control commands. The method of this invention further enhances the transparency of control strategy or algorithm performance and user satisfaction, ensuring the efficiency and effectiveness of the intelligent control process.

[0108] S7. After receiving feedback from the user regarding the analysis results, input the user's feedback results into the large language model. The large language model determines whether the control strategy meets the user's needs. If it does, the modified control variable settings are saved to the historical information database, and the control strategy that meets the user's needs is passed to the field edge controller of the building air conditioning system. The field edge controller then executes the control strategy that meets the user's needs. Otherwise, the control command is re-entered.

[0109] It should be noted that step S7 of the present invention is the process of receiving, analyzing and processing user feedback, as well as the process of executing control strategies on-site. The specific implementation process of step S7 will be briefly described below.

[0110] It should be noted that in step S7 of this invention, the user feedback is first received and analyzed, and then processed. If the user provides positive feedback such as satisfaction or permission to execute, this control operation (processing instructions and the corresponding system operation status and control strategy before and after control) is recorded and stored in the historical information database as a valuable reference for future decisions and operations. If the user provides negative feedback such as dissatisfaction or disallowance to execute, the system interacts with the user, provides interactive guidance, helps the user make necessary adjustments based on the control effect, asks the user to reconfirm the control method based on the previous instructions and corresponding control effects, and re-provides control instructions, returning to S1 to optimize the control effect.

[0111] Additionally, when inputting user feedback into the large language model, the example of the prompt passed to the large language model is: "Please judge the user's feedback information: {user feedback}. Does the feedback indicate that the user's needs have been met and can be executed? If it has been met, answer "YES", otherwise answer "NO". For example: 'User question: It's okay, execute is allowed', then answer "YES".

[0112] This process demonstrates not only the system's high flexibility in user interaction and feedback processing, but also its commitment to continuous improvement and optimization of control strategies. By effectively handling user feedback, the method of this invention enhances user engagement and promotes iterative updates and performance improvements in the intelligent control system.

[0113] It should be noted that when inputting user feedback into the large language model, the user feedback can be in any modality, either text or speech. That is, the large language model can accept various forms of user feedback input. If it is necessary to convert the speech input into text form during this process, the Wav2Vec 2.0 model is still used.

[0114] It should be noted that in step S7 of this invention, the cloud-based digital twin platform finally transmits the corresponding control strategy to the edge controllers at the building's central air conditioning system site, and these controllers execute the control strategy on-site. Specifically, the optimized and user-confirmed control strategy is transmitted to the edge controllers located at the building site via the cloud-based digital twin platform. The edge controllers then execute corresponding control operations in the actual building air conditioning system according to the received control strategy, achieving real-time control and adjustment of the physical environment.

[0115] Furthermore, the on-site execution of the control strategy fully utilizes edge computing technology, ensuring the efficiency and timeliness of control operations. By tightly integrating the cloud platform and edge controller, the method of this invention achieves technological innovation and application breakthroughs in the field of intelligent building control, providing a new, efficient, accurate, and responsive model for building automation system management.

[0116] It should be noted that in step S7 of this invention, the cloud-based digital twin platform transmits the corresponding control strategy to the edge controllers at the building's air handling unit site. These controllers then execute the control strategy on-site. The edge controllers, deployed at the building site, are responsible for executing the control strategy generated by the digital twin platform and directly managing and controlling the HVAC system equipment on-site. After execution, the edge controllers feed back the execution results and environmental change data to the cloud, providing feedback to the user. The user can then adjust and optimize the strategy accordingly to achieve continuous performance improvement. Furthermore, through the integration of edge controllers, the method of this invention can directly execute control strategies at the building site, supporting real-time data acquisition and feedback, and providing immediate operational status information for control decisions. Moreover, by establishing a secure communication protocol, the method ensures the secure transmission of control commands between the digital twin platform and the edge controllers, effectively preventing data tampering and unauthorized access, and guaranteeing system security.

[0117] It is evident that the field execution steps of the edge controller mark a crucial transition from virtual verification to practical application of the control strategy. This not only ensures that the implementation effect of the control strategy is consistent with expectations, but also significantly improves the actual operational effectiveness and user satisfaction of the system.

[0118] Additionally, it should be noted that in step S7 of this invention, the historical information database is used to collect and store historical control commands, analysis results corresponding to control strategies, and user feedback results. This data is used to guide the generation of future control strategies, improving the accuracy of control strategies and user satisfaction. Furthermore, this invention, through a self-learning mechanism, allows the building air conditioning system to automatically optimize control strategies and adjust the understanding and generation logic of the large language model based on data in the historical information database, thereby continuously improving the system's performance and accuracy.

[0119] It should be noted that the building air conditioning system of the present invention can support remote control operation, enabling users or maintenance personnel to connect to the digital twin platform via the network, thereby realizing remote monitoring and control of the on-site HVAC system.

[0120] It should be noted that the building air handling units and building air conditioning systems (the entire HVAC system) also include equipment component modules that can dynamically update and maintain equipment information, including adding, deleting, replacing, and changing the status of equipment, to ensure the consistency and real-time performance of the equipment information in the digital twin model system with the actual building air conditioning system (HVAC system).

[0121] It should be noted that the method of the present invention includes an interactive interface through which users can directly input multimodal control commands, view the generation and execution of control strategies, and receive analysis results on the control effects.

[0122] It's important to note that digital twins are simulation processes that fully utilize physical models, sensor measurements, and operational history data to integrate multiple disciplines, physical quantities, scales, and probabilities. They map the actual system in a virtual space, reflecting the entire lifecycle of the corresponding physical entity. Digital twins possess excellent characteristics such as real-time synchronization, accurate mapping, and high fidelity. They have already been implemented and achieved significant success in several related fields, for example, in industry, significantly driving changes in product design, production, operation, and maintenance. Wind systems are large-scale and have complex topologies, far exceeding human real-time simulation and computational capabilities. Digital twin technology can achieve "transparency" of the entire system through simulation calculations, providing a virtual sandbox for energy-efficient system operation. Unlike traditional simulation, digital twins establish a precise two-way connection between the physical and digital worlds. While reflecting the dynamic process of the physical entity's entire lifecycle, they also support precise simulation and control of interconnected networks from multiple perspectives. For practical engineering scenarios, digital twins significantly expand the scope and meaning of simulation, forming a solution that fully leverages the value of data in real-world applications. A digital twin system is a virtual device system built in the cloud based on actual field equipment; it is a cloud-based mapping of the physical world. In this invention, the digital twin system refers to the software platform that provides this mapping function. The wind system is a subsystem of the digital twin system, and because of the different types of mapped equipment, it can be divided into wind systems and others.

[0123] To better demonstrate the specific implementation and technical effects of the present invention, the central air conditioning system control method based on large language model and digital twin shown in steps S1 to S7 of the above preferred implementation is applied to a specific example.

[0124] Example

[0125] The air system of a factory's HVAC equipment consists of 8 air handling units and 24 fan coil units working together, distributed in the hot areas of the workshop.

[0126] The maintenance personnel gave the voice control command: "Change the outlet air temperature of air handling unit No. 1 to 26℃".

[0127] Step 1. Accept user commands and convert speech to text.

[0128] Step 2. The system determines that the device to be controlled in the operation and maintenance personnel's instruction is Air Handling Unit 1. After printing out the system equipment list, the system variable name of the device to be controlled selected by the large language model is: Air Handling Unit_1.

[0129] Step 3. The equipment components of the air handling unit_1 obtained by the system (including itself) are: air handling unit_1, air handling unit_1_fresh air valve, air handling unit_1_supply air valve, air handling unit_1_return air valve, air handling unit_1_supply air fan, air handling unit_1_fresh air fan, air handling unit_1_cooling coil, air handling unit_1_heating coil, and air handling unit_1_filter.

[0130] The system variable name of the component to be controlled, selected based on the user instruction large language model, is: Air Handling Unit_1.

[0131] Step 4. The automatic control variables of the air handling unit_1 obtained from the system are: ['Start / Stop Control Command', 'Supply Air Temperature Setpoint', 'Supply Air Static Pressure Setpoint', 'Operating Mode', 'Remote / Local']. Analysis shows that the control type is a parameter adjustment command, and the automatic control variable is: supply air temperature setpoint, with a setpoint of 20℃.

[0132] Step 5. Ensure that the system conditions for controlling the air handling unit 1 are met by checking the status of various equipment components, duct valves, and water system related to the air handling unit 1.

[0133] Step 6. Generate control commands and set the supply air temperature setpoint of air handling unit 1 to 26°C in the system.

[0134] Step 7. Perform a digital twin model simulation of the building's central air conditioning system.

[0135] Step 8. Use the digital twin model of the building's central air conditioning system to obtain key variable information about the air system, control equipment, and hot zone terminals before and after the control strategy is executed.

[0136] Step 9. The large language model combines the user's control commands and the key variable information of the wind system, control equipment, and hot zone terminal before and after strategy execution obtained from S7 to analyze the control effect of the control strategy execution and provide feedback to the user, obtaining a response.

[0137] User feedback: "As expected, execution is permitted."

[0138] Step 10. Receive and analyze user feedback. If the feedback is positive, allow the instruction to be executed.

[0139] Step 11. The cloud-based digital twin platform transmits the corresponding control strategies to the edge controllers at the site of the building's central air conditioning system, and these controllers execute the control strategies on-site to complete the control.

[0140] In summary, this invention provides a central air conditioning system control method based on a large language model and digital twin, comprising:

[0141] 1) Receiving user questions is the starting point for building an intelligent control system. It aims to capture and parse user needs expressed in natural language through an efficient input module. This step is designed based on understanding the complexity and diversity of user needs, including multimodal information such as direct text and voice input, ensuring the system can accurately understand and process input in different formats. The input module's preprocessing functions, such as syntax correction, keyword extraction, and deambiguation, further improve the efficiency and accuracy of question parsing, providing a clear and accurate description of user needs for subsequent steps.

[0142] 2) Determining the target device for the control command involves accurately identifying the target device type and device number in the user's intent. This step first ensures the clarity and executability of the command by determining the user type, the thermal zone the user is located in, and whether the command explicitly specifies the device type and number. If information is insufficient, the system proactively interacts with the user, requesting additional information, re-issuing the command, and returning to the first step, demonstrating the system's adaptability and user-friendliness. Secondly, the system invokes the digital twin model system to obtain a list of permissible controllable devices, such as air handling units, corresponding to the current building's thermal zones and hot zones. Then, by combining the device the user intends to control with its device number, the system selects the controllable device in the twin system, showcasing the powerful capabilities of digital twin technology in simulating and matching physical world devices.

[0143] 3) Confirming the target component of the control command: This step involves accurately identifying the specific component of the target device in the user's intent. First, it determines the user's control intent and whether the command explicitly specifies the target device component, ensuring the diversity of users and control commands, as well as the clarity and executability of the commands. If information is insufficient, the system proactively interacts with the user, requesting additional information, re-issuing the command, and returning to step one. Next, it calls the digital twin model system to obtain the component list corresponding to the target device in the building. Then, based on the device component intended to be controlled by the user's command, it selects the device component to be controlled from the twin system. If the device component does not exist or is unusable, it interacts with the user, indicating that the target device component is missing or unusable, asking the user to re-issue the command, and returning to step one, demonstrating the system's adaptability and user-friendliness.

[0144] 4) Confirming the control variables and setpoints of the command is the core of intelligent control, aiming to transform the user's abstract instructions into specific control commands. This step first, in conjunction with the components of the device to be controlled, determines and ensures that the user's control command is clear, unambiguous, and executable. If information is insufficient, the system proactively interacts with the user, requesting additional information, a revised command, and returns to step one. Secondly, by analyzing the self-control variables of the device components the user intends to control, the system can accurately identify the names of the self-control variables in the user's intent and determine the setpoints of the control variables based on the type of control command (e.g., on, off, control mode, or parameter adjustment). If the command does not belong to any corresponding control method, the system interacts with the user, indicating that the device component to be controlled does not have a corresponding control method, requesting the user to re-issue the command, and returns to step one. This process considers the diversity and ambiguity of user commands. For example, in parameter adjustment control, if the user does not explicitly specify the setpoints of the control variables, the system can match specific setpoints by combining historical information. This process demonstrates the system's intelligence and flexibility in handling fuzzy information.

[0145] 5) Verify and confirm that the current air system status meets the control conditions. This step ensures the safety and effectiveness of control command execution. This step first checks the status of the air valves in the user's room and hot zone to confirm the airflow path and the operation of the air system. Secondly, it confirms the topological connections of the components to be controlled and the various terminal rooms, including all components and valves such as air valves, fans, hot and cold water coils, etc., and sorts them according to their inspection level. Before executing any control operation, the system verifies whether the current status of the air system meets the preset control conditions. This step reflects the emphasis on system stability and operational safety, ensuring that control commands are executed under appropriate system conditions.

[0146] 6) Automatically generate and execute control strategies. This step demonstrates the automation capabilities of the intelligent control system. Based on the control devices, components, control variables, and specific parameter settings determined in the previous steps, the system can automatically generate the corresponding specific control strategy for the digital twin building wind system and execute the corresponding system changes. This step showcases the powerful capabilities of digital twin technology in achieving real-time synchronization between virtual and real environments and the automated deployment of control strategies.

[0147] 7) Perform digital twin model simulation. This step is crucial for optimizing control strategies or algorithm parameters. High-precision computational simulations predict and analyze changes in various parameters of the building's wind system and room terminals under control operations. This step allows the system to assess the potential impact and effects of control strategies before actual application, thereby optimizing control commands and strategies to ensure optimal control performance.

[0148] 8) Obtain key variable information for the execution of the control strategy or algorithm. This step, by calling the digital twin model, allows the system to obtain key variable information such as temperature, humidity, and comfort parameters of the air system and control equipment before and after the execution of the control strategy or algorithm, as well as the thermal zones or room terminals they are responsible for. This provides data support for subsequent control effect analysis. This step demonstrates the application of digital twin technology in monitoring and recording changes in system state, providing an accurate data foundation for the evaluation and adjustment of control strategies.

[0149] 9) Control Effect Analysis: The large language model comprehensively considers user control commands and acquired key variable information to analyze the effect of control strategies or algorithms, and feeds the analysis results back to the user. This step demonstrates the system's ability to evaluate and optimize the execution effect after completing the control adjustment command, ensuring that the user can obtain the desired control effect and make adjustments when necessary.

[0150] 10) Receive and analyze user feedback. The large language model completes the final evaluation of the control strategy or algorithm by analyzing user feedback on the control effect. Positive feedback will lead to the saving of control records, providing experience for future control operations; while negative feedback will trigger further interaction with the user, requiring reconfirmation of the control method and returning to the first step, demonstrating the system's adaptability and user-centricity in continuously optimizing the control effect.

[0151] 11) The interaction between the cloud-based digital twin platform and the edge controller marks the transition of control strategies or algorithms from virtual simulation to real-world implementation. By transmitting optimized control strategy or algorithm parameters to the on-site edge controller via the cloud platform, this step demonstrates the ability of digital twin technology to integrate cloud and edge computing resources to achieve efficient and reliable deployment of control strategies or algorithms.

[0152] In summary, the core innovation of this invention lies in the integrated application of large language models and digital twin technology, providing a new, efficient, accurate, and low-risk method for the intelligent control of building air conditioning systems. This method cleverly combines large language models and digital twin technology to address the problems of insufficient informatization in the building industry and excessive reliance on manual operation for control and maintenance. In the management of building heating, ventilation, and air conditioning (HVAC) systems, achieving efficient energy utilization and improving indoor environmental comfort has always been a core objective. However, traditional HVAC system control methods often fail to meet the personalized needs of end users due to a lack of flexibility and accuracy, while also posing challenges to maintenance personnel in maintenance and adjustment. These systems often cannot directly respond to specific user needs, have rigid control logic, unintuitive user interfaces, and lack dynamic response capabilities to changes in the outdoor environment. Furthermore, maintenance personnel frequently need to manually intervene to correct or redesign control strategies, which not only increases labor intensity but also raises skill requirements. By introducing natural language processing technology, this method can directly respond to user control requests expressed in natural language, utilizing the advanced understanding and cognitive capabilities of large language models, combined with digital twin technology, to achieve precise control and optimization of various air system equipment, thus realizing intelligent management. This method not only accurately parses user control commands and automatically identifies relevant equipment, control variables, and setpoints, but also performs pre-verification within a digital twin model to ensure the accuracy and integrity of the control scheme. By first verifying the control scheme or algorithm in a virtual environment and then adjusting edge controller commands via a cloud platform for implementation, this invention significantly reduces operational risks and avoids potential errors, thereby improving system efficiency and security. Furthermore, this method supports multimodal data input, including text and voice, enhancing user interaction and optimizing the understanding of control logic and user needs. The implementation of this method not only reduces costs potentially caused by misoperation but also promotes the transformation of system operation and maintenance management towards intelligence and automation, opening new avenues for technological innovation and application. Overall, the core innovation of this invention lies in the integrated application of large language models and digital twin technology, providing a new, efficient, accurate, and low-risk method for the intelligent control of building central air conditioning systems.

[0153] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A central air conditioning system control method based on large language models and digital twins, characterized in that, Includes the following steps: S1. The input module receives the control command issued by the user, matches the control command with the pre-built keyword library, and determines whether the control command contains the keywords in the keyword library: if so, the control command is preprocessed to obtain a new control command; otherwise, the user is prompted that the control command entered does not belong to the building energy field and needs to be re-entered. S2. Determine the user type based on the permission level of the control command: If the user is not an operations and maintenance personnel, automatically add the thermal partition where the user is located to the new control command based on the address where the control command was sent, and use the new control command with the added thermal partition as the processing command; otherwise, there is no need to add the thermal partition, and the new control command is used as the processing command. After determining the user type, the processing command is input to the large language model to determine whether the necessary information is indicated in the processing command. If not, the user is prompted on the interactive interface to supplement the device information and re-enter the control command. If both are specified, the system will retrieve the equipment component list for the thermal zone from the building air conditioning system in the digital twin model system. Based on the necessary information in the processing instructions, the system will match the proposed control device from the digital twin model system in the equipment component list. If the proposed control device cannot be used, the user will be prompted on the interactive interface that the proposed control device cannot be used. Please reconfirm the proposed control device and re-enter the control instructions. The necessary information is the type and number of the proposed control device, or the necessary information is the type and thermal zone of the proposed control device. S3. Determine the type of device the user intends to control based on the processing command: If the type is the entire device, then execute S4; if the type is a device component and the user type is not an operations and maintenance personnel, then prompt the user on the interactive interface that the user's permissions are insufficient and that the control command needs to be re-entered; if the type is a device component and the user type is an operations and maintenance personnel, then call the building air conditioning system in the digital twin model system to obtain the list of control components of the device to be controlled. If the processing command and the list of control components match, then determine the name of the control component from the list of control components. If the processing command and the list of control components do not match, then prompt the user on the interactive interface that the control component of the device to be controlled does not exist and that the control command needs to be re-entered. S4. The large language model determines whether the processing instruction matches the control command type of the device to be controlled: if they do not match, the user is prompted to reconfirm the control command type and re-enter the control instruction on the interactive interface; if they match, the system calls the building air conditioning system to obtain the list of automatic control variables for each component of the device to be controlled; if the processing instruction and the list of automatic control variables match, the control variable name is determined from the list of automatic control variables, and the control variable setting value corresponding to the control command type is output; if the processing instruction and the list of automatic control variables do not match, the user is prompted to reconfirm the control variables of the device to be controlled and re-enter the control instruction on the interactive interface. S5. Obtain the name of the device to be controlled from the building air conditioning system, and obtain the topology connection relationship of the components of the device to be controlled; sort the components in the topology connection relationship according to the pre-built inspection level of each component, and obtain the sorted component list; query the sorted component list in turn. For the current component in the sorted component list, call the building air conditioning system to first obtain the operating status of the current component. If the operating status of the current component meets the executable processing instructions and control variable setting values, then query the next component until all components to be controlled are traversed. Otherwise, call the building air conditioning system to adjust the operating status of the current component until the operating status of the current component meets the executable processing instructions and control variable setting values. S6. Obtain the control strategy of the building air conditioning system corresponding to the processing command, input the control strategy into the building air conditioning system, change the set value of the control variable, obtain the performance data of each device of the building air conditioning system before the change, and perform digital twin model simulation on the building air conditioning system after the change to obtain the performance data of each device of the building air conditioning system after the change. Input the performance data before the change and the performance data after the change into the large language model to obtain the analysis results of the control effect and feed them back to the user. S7. After receiving feedback from the user regarding the analysis results, input the user's feedback results into the large language model. The large language model determines whether the control strategy meets the user's needs. If it does, the modified control variable settings are saved to the historical information database, and the control strategy that meets the user's needs is passed to the field edge controller of the building air conditioning system. The field edge controller then executes the control strategy that meets the user's needs. Otherwise, the control command is re-entered.

2. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, In step S1, the input module is capable of receiving various forms of control command input and has the ability to convert speech into text; the preprocessing includes correcting grammatical errors in the control commands, extracting key information points from the control commands, and eliminating ambiguity in control commands with multiple interpretations by using contextual information or requesting more information from the user.

3. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, In step S4, if the control command type is an "on" command, the control variable setting value is 1; if the control command type is a "off" command, the control variable setting value is 0; if the control command type is a parameter adjustment command, the large language model uses the setting value in the processing instruction as the control variable setting value; if the processing instruction does not have a setting value, the retrieval model searches for historical setting values ​​of the automatic control variables in the historical information database based on the current operating status of the building's air conditioning system, and uses the historical setting value closest to the value as the control variable setting value; if the control command type is a control mode adjustment command, the corresponding control variable setting value is the air handling unit operating mode specified in the processing instruction; if not specified, the user is prompted on the interactive interface to reconfirm the air handling unit operating mode and re-enter the control instruction.

4. The central air conditioning air system control method based on large language model and digital twin as described in claim 3, characterized in that, In step S5, if the control command type is an open command, a control mode adjustment command, or a parameter adjustment command, the equipment order corresponding to the inspection level is: all types of duct valves, all types of equipment air valves, all types of equipment fans, and all types of cooling / heating coils; if the control command type is a close command, the equipment order corresponding to the inspection level is: all types of cooling / heating coils, all types of equipment fans, all types of equipment air valves, and all types of duct valves.

5. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, The building air conditioning system also includes equipment modules that can dynamically update and maintain equipment information.

6. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, The historical information database is used to collect and store historical control commands, analysis results of control strategies, and user feedback.

7. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, The building air conditioning system supports remote control operation, and users or maintenance personnel can remotely monitor and control the building air conditioning system through network connection.

8. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, When acquiring control strategies, the control strategies can be adjusted according to the real-time operating status of each device in the building's air conditioning system, or users can select control strategies based on the control effect, and the system can automatically conduct risk assessments on the control strategies.

9. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, Through a self-learning mechanism, the building air conditioning system is allowed to automatically optimize control strategies and adjust the understanding and generation logic of the large language model based on data in the historical information database.

10. The central air conditioning air system control method based on large language model and digital twin as described in claim 1, characterized in that, The edge controller is deployed on the building site and is responsible for executing control strategies and directly managing and controlling the building's air conditioning system equipment on site. The edge controller supports real-time data acquisition and analysis, and can instantly feed back the operating status of each device in the building's air conditioning system to the digital twin platform, providing a basis for the generation and optimization of control strategies. Through the secure communication protocol established between the digital twin platform and the edge controller, the secure transmission and execution of processing instructions are ensured, preventing data tampering and unauthorized access.

Images