Method, device and equipment for processing data of heating and ventilation system based on large language model

Abstract

The present disclosure provides a heating and ventilation system data processing method, device and equipment based on a large language model, relates to the technical field of artificial intelligence, in particular to the field of data processing, and can be applied to scenarios such as buildings, factories and data centers. The specific implementation scheme is: in response to receiving input information for a heating and ventilation system, recalling context information related to the input information from a plurality of information; determining N target tools in a plurality of candidate tools and a calling sequence of the N target tools according to the input information and the context information, N being an integer greater than or equal to 1; and based on the calling sequence, calling the N target tools to process the input information to obtain processing results of the N target tools.

Description

Data processing method, apparatus and equipment for HVAC systems based on large language models Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and more particularly to the field of data processing. More specifically, this disclosure provides a data processing method, apparatus, electronic device, storage medium, and computer program product for HVAC systems. Background Technology

[0002] With the increasing popularity of intelligent control products, their functions are becoming more and more sophisticated. Among them, HVAC control systems, due to their high power consumption in buildings, have become one of the most common intelligent control products. In practical applications, HVAC systems can be maintained and managed by maintenance personnel, but this method is costly and inefficient. Summary of the Invention

[0003] This disclosure provides a data processing method, apparatus, electronic device, storage medium, and computer program product for HVAC systems.

[0004] According to one aspect of this disclosure, a data processing method for a heating, ventilation, and air conditioning (HVAC) system is provided, comprising: in response to receiving input information for the HVAC system, retrieving context information related to the input information from multiple pieces of information; determining N target tools among multiple candidate tools and the calling order of the N target tools based on the input information and the context information, where N is an integer greater than or equal to 1; and calling the N target tools to process the input information based on the calling order, thereby obtaining the processing results of the N target tools.

[0005] According to another aspect of this disclosure, a data processing apparatus for a heating, ventilation, and air conditioning (HVAC) system is provided, comprising: a recall module, a determination module, and a calling module. The recall module is used to recall context information related to the input information from multiple pieces of information in response to receiving input information for the HVAC system. The determination module is used to determine N target tools from a plurality of candidate tools and the calling order of the N target tools based on the input information and the context information, where N is an integer greater than or equal to 1. The calling module is used to call the N target tools to process the input information based on the calling order, and obtain the processing results of the N target tools.

[0006] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods provided in this disclosure.

[0007] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause a computer to perform the methods provided in this disclosure.

[0008] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the methods provided in this disclosure.

[0009] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0010] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0011] Figure 1 is a schematic diagram of an application scenario of a data processing method and apparatus for a heating, ventilation and air conditioning system according to an embodiment of the present disclosure;

[0012] Figure 2 is a schematic flowchart of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure;

[0013] Figure 3 is a schematic diagram of a prompt message template according to an embodiment of the present disclosure;

[0014] Figure 4A is a schematic diagram of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure;

[0015] Figure 4B is a schematic flowchart of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure;

[0016] Figure 5 is a schematic diagram of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure;

[0017] Figure 6 is a schematic diagram of a control system for HVAC according to an embodiment of the present disclosure;

[0018] Figure 7 is a schematic diagram of a control system for HVAC according to another embodiment of the present disclosure;

[0019] Figure 8 is a schematic structural block diagram of a data processing device for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure; and

[0020] Figure 9 is a structural block diagram of an electronic device used to implement the data processing method of a heating, ventilation and air conditioning system according to an embodiment of the present disclosure. Detailed Implementation

[0021] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0022] In the HVAC field, it is necessary to have an operation and maintenance team familiar with HVAC equipment. This team interacts with users through a platform, answers various user questions, provides services, and realizes intelligent control of the HVAC system.

[0023] However, user needs can be relatively complex, and platforms cannot automate their processing. Often, initial judgments based on human experience are required, followed by a series of data analyses before being translated into control commands. This process relies on experienced operations personnel for decision-making, resulting in high manual operating costs, low automation, low efficiency, and marginal costs that are difficult to reduce as project scale increases.

[0024] The present disclosure aims to provide a data processing method for HVAC systems, which can replace the role of operation and maintenance experts, improve operation and maintenance efficiency, accelerate the response to user needs, and reduce the human resource operation costs of HVAC systems.

[0025] This disclosure applies to scenarios such as buildings, factories, and data centers. Buildings may include commercial buildings, residential buildings, and industrial buildings, while data centers may include IDC (Internet Data Center). It is particularly applicable to various HVAC control equipment and other fields.

[0026] The technical solutions provided in this disclosure will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0027] Figure 1 is a schematic diagram of an application scenario of a data processing method and apparatus for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure.

[0028] It should be noted that Figure 1 is only an example of a system architecture that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but it does not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0029] As shown in Figure 1, the system architecture 100 according to this embodiment may include a terminal device 101, a server 102, a database 104, and multiple tools 103. Data transmission between the devices occurs via a network, which may include various connection types, such as wired and / or wireless communication links, etc.

[0030] Users can interact with server 102 using terminal device 101 to receive or send messages, etc. Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0031] Server 102 can be a server that provides various services, such as a backend management server that supports websites browsed by users using terminal device 101 (for example only). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as processing results determined based on user input information) to the terminal device. Server 102 can be deployed with a pre-trained model, which can be a Large Language Model (LLM).

[0032] Database 104 can store some basic industry knowledge of the system, analysis of problems, and solutions. This system can be used for HVAC systems.

[0033] There can be multiple tools 103. Tools 103 can process data in the system, such as querying data, analyzing data, and calling specific algorithms to process data.

[0034] In some embodiments, a user can interact with terminal device 101 to send input information. Terminal device 101 can then retrieve context information from database 104 and input the input information, context, and other information to server 102. The server processes the information using a predefined model it has deployed, for example, determining the order in which target tools are invoked and providing feedback to terminal device 101. Terminal device 101 can then invoke multiple tools 103 to obtain processing results and provide those results back to the user.

[0035] It should be noted that the data processing method for the HVAC system provided in this disclosure embodiment can generally be executed by the terminal device 101 and / or the server 102. Accordingly, the data processing device for the HVAC system provided in this disclosure embodiment can generally be located in the terminal device 101 and / or the server 102.

[0036] It should be understood that the number of terminal devices, networks, and servers shown in Figure 1 is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0037] Figure 2 is a schematic flowchart of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure.

[0038] As shown in Figure 2, the data processing method 200 of the HVAC system may include operations S210 to S230.

[0039] In operation S210, in response to receiving input information for the HVAC system, context information related to the input information is retrieved from multiple pieces of information.

[0040] For example, users can input text through the front-end page and use that text as input information. Alternatively, users can input voice, which can be recognized and used as text as input information.

[0041] For example, the method provided in this embodiment can be applied to HVAC systems or other systems. Taking an HVAC system as an example, the input information can include at least one category of information. For example, the input information can include questions about HVAC knowledge, query information for querying HVAC data operation data, analysis information for analyzing the operation of the HVAC system, control information for regulating the HVAC system, etc.

[0042] For example, a database can be pre-built. Taking HVAC systems as an example, a database can be built based on maintenance manuals and general knowledge of the HVAC industry, using a pre-trained large language model. The database can store basic industry knowledge, problem analysis, and processing methods. Since the database stores a large amount of data, a model can be used to transform the information into a specific storage format. The database stores multiple paragraph texts or vectors of paragraph texts, each paragraph text can have a corresponding index. After obtaining the user's input information, paragraph texts can be retrieved based on the index; these retrieved paragraph texts or vectors serve as contextual information. This embodiment does not limit the data content stored in the database.

[0043] In operation S220, based on the input information and context information, N target tools and the calling order of the N target tools are determined from multiple candidate tools, where N is an integer greater than or equal to 1.

[0044] For example, keywords can be extracted from input and context information. Based on pre-configured correspondences, the target tool corresponding to the keyword can be determined. The invocation order of multiple candidate tools can also be pre-configured, with the target tool invocation order matching the candidate tool invocation order. Alternatively, a pre-defined model can be used to determine N target tools and their invocation order; the determination method will be explained in detail below.

[0045] For example, the tool can be pre-packaged. For instance, a model or service implementing a specific algorithm or function can be encapsulated into a tool that can be invoked by other devices (e.g., electronic devices executing the methods of this embodiment) or large language models. The encapsulated tool needs to provide other devices or large language models with a description of its functionality, interface parameters, and invocation examples. This embodiment does not limit the implementation principle of the tool.

[0046] For example, candidate tools correspond to categories of input information, with at least one candidate tool for each category of input information, which is used to process the input information of that category.

[0047] For example, tools corresponding to the question can answer platform functions, explain HVAC intelligent control professional knowledge, etc. Tools corresponding to query information can query the system's operating data at a single moment or over a period of time, and query system alarm or abnormal situation information. Tools corresponding to analyze information can include generating reports, calculating statistical information of system data over a period of time, and drawing line charts and bar charts based on the data. Tools corresponding to control information can set system operating parameters, control the overall start and stop of the control system, control the start and stop of individual devices in the control system, and adjust operating strategies according to changes in the environment or on-site needs, etc. The number of target tools can be one, two, or more.

[0048] In operation S230, based on the calling order, N target tools are called to process the input information, and the processing results of N target tools are obtained.

[0049] For example, N target tools can be called sequentially in the order of invocation to process the input information and obtain the processing results.

[0050] According to the technical solution provided in this disclosure, multiple tools capable of performing multiple functions can be integrated in advance. During use, based on the user's input problem information, context information is retrieved from the database, and the calling order of the target tools is planned. Each target tool is then called sequentially according to the calling order to solve the problem. Therefore, it can replace the role of maintenance experts, improve maintenance efficiency, accelerate the response to user needs, and reduce the human resource operating costs of HVAC systems.

[0051] According to another embodiment of this disclosure, the process of determining N target tools and their invocation order from a plurality of candidate tools based on input information and context information may include the following operations: determining a first prompt message based on the input information, context information, and a first prompt message template; then inputting the first prompt message into a predetermined model to obtain the output information of the predetermined model, the output information including the identifiers and invocation order of the N target tools.

[0052] For example, a first prompt information template can be pre-configured, which can be a piece of natural language text. The first prompt information template may include functional descriptions of multiple candidate tools, i.e., text describing the function of each tool. The first prompt information template may also include interface parameter descriptions of multiple candidate tools, i.e., text describing the parameters of each tool, so that the predetermined model can extract the tool's input parameters based on the interface parameter descriptions. The first prompt information template may also include task information, which represents the task of the predetermined model. For example, the task information may include first task information, which indicates that the predetermined model needs to select at least one target tool from multiple candidate tools based on the first prompt information and determine the calling order of the target tools. The task information may also include second task information, which represents extracting the input parameters of each target tool from the information input to the predetermined model. For example, the process of determining the prompt information based on the prompt information template may include combining other information with the prompt information template to obtain the prompt information. For example, an example of a prompt information template 301 is shown in Figure 3.

[0053] For example, the predefined model can be a large language model; this embodiment does not limit the predefined model. It is understood that the target tool is selected using the predefined model, and the target tool processes the data, rather than the predefined model replacing the target tool in data processing. By encapsulating the tool, the problem of large language models' poor performance in handling precise calculations, which affects processing efficiency, can be alleviated. This embodiment uses the predefined model and the first prompt information template to determine the target tool and the invocation order, enabling the integration of the predefined model with the HVAC system and improving the intelligence of the HVAC system.

[0054] Figure 4A is a schematic diagram of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure, and Figure 4B is a schematic flowchart of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure.

[0055] As shown in Figure 4A, this embodiment involves an agent 401, a predetermined model 402, and multiple target tools 403. The agent 401, the predetermined model 402, and the target tools 403 can be deployed in the same electronic device or in different electronic devices. As shown in Figure 4B, the data processing method 400 of this HVAC system may include operations S410 to S420 and operations S431 to S439.

[0056] In operation S410, in response to receiving input information for the HVAC system, context information related to the input information is retrieved from multiple pieces of information.

[0057] In operation S420, based on the input information and context information, N target tools and the calling order of the N target tools are determined from multiple candidate tools, where N is an integer greater than or equal to 1.

[0058] For example, input information, context information, and a first prompt information template can be combined into a first prompt information. Then, the first prompt information can be input into a predetermined model. The predetermined model can output multiple target tools and their calling order, and can also output the input parameters of the target tools.

[0059] In operation S431, the current tool among N target tools is determined according to the calling order.

[0060] For example, in the order of invocation, N target tools are identified as the current tool, and each target tool corresponds to a round of processing.

[0061] When operating S432, if the current tool is the first tool, the current tool is called according to the input parameters for the first tool, and the processing result of the current tool is obtained.

[0062] For example, input parameters can be input into the current tool, and the current tool can output the processing result, thereby realizing the invocation of the current tool.

[0063] In operation S433, in response to determining that the current tool is one of the other tools after the first tool, the second prompt message is determined based on the input information, context information, the processing result of the previous tool, and the second prompt message template.

[0064] For example, the second prompt information template may include functional description information for multiple candidate tools, interface parameter description information for multiple candidate tools, and task information, which may include the aforementioned second task information and the aforementioned first task information.

[0065] The second prompt template can be the same as the first prompt template. The predefined model can also output the identifier of the target tool and the calling order of the target tools. It can be seen that the target tools and calling order output by the predefined model in this round will update the target tools and calling order output in the previous round. For example, if the calling order of the target tools was determined to be A, B, C, and D in the previous round, after obtaining the processing result of tool A, the second prompt information is determined based on the processing result of tool A. The predefined model outputs the calling order of C and D based on the second prompt information. That is, the predefined model in the second round determines that there is no need to call tool B again, and tool C can be called directly.

[0066] The second prompt message template can be different from the first prompt message template. The predetermined model may not output the identifier of the target tool and the calling order of the target tool. That is, after the target tool and calling order are determined in the first round, the target tool and calling order will not be updated. Instead, the calling order determined at the beginning will be used to call each tool in sequence.

[0067] In operation S434, the second prompt information is input into the predetermined model to obtain the input parameters of the current tool.

[0068] When operating S435, based on the input parameters of the current tool, the current tool is invoked to obtain the processing result of the current tool.

[0069] As can be seen, the above operation can sequentially call the current tool among multiple target tools and obtain the processing result of the current tool. It should be noted that sometimes the user input information and the recalled context information are incomplete, and the input parameters of the next target tool cannot be determined solely based on the user input information and the recalled context information. This embodiment calls the target tools sequentially and determines the input parameters of the next target tool based on the processing results of the previous target tools, thereby obtaining valid input parameters for the next target tool.

[0070] In some embodiments, after obtaining the processing result of the current tool, the next round of operation can be directly entered. For example, the subsequent tool can be identified as the current tool, and the operation of determining the second prompt information can be returned, thereby calling the next tool and determining the processing result of the next tool, until the processing result of the last target tool is obtained, thereby ensuring that a complete processing result is obtained.

[0071] In other embodiments, after obtaining the processing result of the current tool and before entering the next round of operation, the processing result of the current tool can be evaluated. The evaluation process will be described below in conjunction with operations S436 to S438.

[0072] In operation S436, the third prompt message is determined based on the input information, the current tool's processing result, and the third prompt message template.

[0073] For example, the third prompt message template may include the order in which the target tools are invoked and the attribute information of the expected results of the target tools. For instance, the third prompt message template may include: "First invoke tool A to obtain XX, then invoke tool B to obtain XX." The third prompt message template may also include third task information, which represents a predetermined model used to evaluate whether the input information is consistent with the processing result of the current tool.

[0074] In operation S437, the third prompt information is input into the predetermined model to obtain an evaluation result of the processing result of the current tool.

[0075] In step S438, determine whether the evaluation result is satisfactory.

[0076] If the operation fails, the system can return to the operation of the current tool among the N target tools until the evaluation is passed or a predetermined termination condition is met. For example, the predetermined termination condition may include: the number of times the operation of the current tool is returned is greater than or equal to a threshold number. The threshold number can be pre-configured according to actual needs, and the threshold number can be 5 times. This embodiment does not limit this.

[0077] If operation S439 is successful, the process can proceed to the next round. For example, the current tool can be selected as the next tool, and the above operation can be returned.

[0078] It is understandable that the current tool's processing result may be incorrect for various reasons, such as errors occurring during the tool's execution. Another example is that tool A should theoretically be called, but tool B is actually called instead. Operations S436 to S438 can evaluate whether the current tool call was successful and whether the current tool's processing result deviates from the original input information. If there is no deviation, the next round of operations can proceed, i.e., calling the next tool and determining its processing result. If there is a deviation, a rollback can be performed to promptly correct the accuracy of the processing result, avoiding errors that could lead to subsequent processing anomalies.

[0079] It should be noted that processing input information sometimes requires calling multiple tools. For example, it may be necessary to first query data, then perform data analysis, and then call an algorithm. This process involves multiple target tools, and each target tool will provide feedback. In another embodiment, before calling the current tool, information such as user input, database retrieval context information, and the results of previous rounds of tool processing can be collected together. Then, a summary of this information can be determined, and the summary information and the second prompt information template can be combined into a second prompt information, which is then input into a predetermined model. This embodiment does not limit the method for generating the summary.

[0080] Compared to inputting user input information, database recall context information, and tool processing results from previous rounds into a predefined model, the second prompt information generated based on summary information contains less text, thereby reducing the number of tokens input into the predefined model and preventing the number of tokens from exceeding the model's input limit.

[0081] In some embodiments, the input information, context information, and processing results of the preceding tool can be segmented, for example, by splitting the information into multiple sub-information, generating sub-summary information for each sub-information, and then combining the multiple sub-summary information into a summary information. This approach ensures that the summary information contains content relevant to the input information, context information, and processing results of the preceding tool, avoiding the problem that some information (such as input information) may not be reflected in the summary information due to its small data volume.

[0082] In other embodiments, after obtaining the summary information, a second summary information can be generated based on the first summary information, thereby further compressing the amount of data input to the predetermined model. Furthermore, preprocessing can be performed first, for example, retaining user input information, tool call order, tool call results, etc., while deleting information such as function descriptions and interface parameter descriptions for multiple candidate tools. Then, summary information is generated based on the preprocessed information, thus allowing for targeted summary information generation.

[0083] Figure 5 is a schematic diagram of a data processing method for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure.

[0084] As shown in Figure 5, this embodiment involves an agent 501, a predetermined model 502, a database 503, and multiple target tools 504. The agent 501 can execute the invocation operations of the target tools 504. The agent 501 can call the model's interface and can encapsulate multiple prompt message templates (such as the first, second, and third prompt message templates mentioned above). The agent 501 can also post-process the results output by the predetermined model 502. For example, if the predetermined model 502 outputs input parameters for the target tools 504, the agent 501 can parse these input parameters, such as performing format conversion, and then invoke the target tools 504 based on the processed input parameters. The multiple target tools 504 may include, for example, data analysis tools, algorithm strategy tools, and device control tools.

[0085] During use, the user can input text or voice data into the agent 501 through the front-end page; this text or voice data constitutes the input information. The agent 501 then retrieves context information related to the input information from the database 503. Next, the agent 501 determines prompt information based on the input information, context information, and prompt template, and inputs the prompt information into the predefined model 502. The predefined model 502 selects the target tool 504 from multiple candidate tools and sorts the order in which the target tools 504 are invoked, then outputs this information to the agent 501. The agent 501 can then invoke multiple target tools 504 to process the input information based on the output information from the predefined model 502, thereby obtaining the processing result.

[0086] Figure 6 is a schematic diagram of a control system for HVAC according to an embodiment of the present disclosure.

[0087] As shown in Figure 6, this embodiment involves an agent 601, a predetermined model 602, a database 603, and multiple tools, including, for example, a query tool and a device control tool 604.

[0088] During use, users can input text or voice information, such as "Adjust the meeting room temperature to 1 degree higher than yesterday at 4 PM tomorrow." Through the interaction between agent 601, pre-defined model 602, and database 603, pre-defined model 602 can determine the parameters to be adjusted, which may include information such as time, location, and target temperature. The mapping between location information and devices can also be pre-configured to determine the device 608 to be adjusted corresponding to that location.

[0089] The pre-defined model 602 can identify the query tool and the equipment control tool 604 as target tools. The query tool can query yesterday's conference room temperature, and the equipment control tool 604 can regulate the HVAC system based on the received information to be regulated. For example, the equipment control tool 604 can send instructions to the controller 605, including the aforementioned parameters to be regulated. Based on the disturbance information 607 and the information to be regulated, the controller 605 calculates the target regulation values ​​of control quantities 606 such as pump frequency and outlet water temperature of the HVAC equipment under the condition of meeting the regulation target, and then sends the target regulation values ​​of control quantities 606 to the equipment to be regulated 608 for regulation. After regulating the equipment 608, the controlled objects 609 such as total system power and terminal temperature can be identified, and then these controlled objects 609 are fed back to the agent 601, which in turn feeds back to the user.

[0090] It should be noted that the above-mentioned control targets may include target temperature or energy consumption constraint information. The energy consumption constraint information restricts the maximum energy consumption or restricts the selection of the most energy-efficient control scheme from multiple feasible control schemes.

[0091] As can be seen, this embodiment allows users to describe their own requirements and then automate the process to achieve intelligent, adaptive, and highly efficient temperature and humidity control of the HVAC system.

[0092] Figure 7 is a schematic diagram of a control system for HVAC according to another embodiment of the present disclosure.

[0093] The embodiments disclosed herein are applicable to scenarios such as buildings, factories, and data centers. Taking the building scenario as an example, buildings are usually equipped with heating, ventilation, and air conditioning (HVAC) systems. HVAC systems include multiple devices, such as cooling towers, chilled water pumps, cooling water pumps, evaporators, condensers, expansion valves, and compressors. This embodiment does not limit the structure of the HVAC system.

[0094] This embodiment involves an agent 701, a predetermined model 702, a database 703, and multiple tools, which may include a query tool 704 and a device control tool 705.

[0095] During use, users can input text or voice information through the front-end page or interactive device. For example, the input information could be "Tomorrow at 4 pm, adjust the temperature of meeting room XX to be 1 degree higher than yesterday".

[0096] Next, agent 701 can retrieve some contextual information from database 703 based on the input information. For example, the retrieved contextual information may include historical temperature information related to the conference room, historical operating parameters of various HVAC system devices related to the conference room, etc.

[0097] Agent 701 can determine the first prompt information based on the input information, context information, and the first prompt information template. The first prompt information template can be referred to above, and will not be repeated in this embodiment.

[0098] Agent 701 inputs the first prompt information into the pre-defined model 702. The pre-defined model 702 outputs some information, which may include, for example, the identifiers of N target tools, their calling order, and the input parameters for the query tool 704. For example, the output information of the pre-defined model 702 may represent: the target tools include, for example, the query tool 704 and the equipment control tool 705, and the calling order is: first, the query tool 704 is called to query yesterday's meeting room temperature information, and then the equipment control tool 705 is called to adjust the equipment to change the meeting room temperature.

[0099] Agent 701 first calls query tool 704. For example, it calls the current tool based on the input parameters for query tool 704 and obtains the processing result of query tool 704. For example, if the input parameters for query tool 704 are: "Time: September 1, 2023, Location: Meeting Room No. XX, Item: Temperature", the query result of query tool 704 may include "26℃".

[0100] Next, the processing result of query tool 704 is evaluated. For example, a third prompt message can be determined based on the input information, the current tool's processing result, and the third prompt message template. The third prompt message template can be referred to above and will not be repeated in this embodiment. Then, the third prompt message is input into the predetermined model 702 to obtain the evaluation result of the current tool's processing result.

[0101] For example, if the processing result of query tool 704 is empty, the predefined model 702 can determine that the evaluation has failed. In this case, a rollback can be performed, such as calling query tool 704 again.

[0102] For example, if the processing result of the query tool 704 is "26℃", then the predetermined model 702 judges that the evaluation is passed. At this time, the next round of operation can be carried out, that is, the equipment control tool 705 is called and the processing result of the equipment control tool 705 is evaluated.

[0103] Next, the device control tool 705 can be invoked. For example, summary information can be determined based on the input information, context information, and the processing result of the query tool 704. Then, based on the summary information and the second prompt information template (which can be referred to above, and will not be repeated in this embodiment), the second prompt information is determined. The second prompt information is then input into the predetermined model 702 to obtain the input parameters for the device control tool 705. Based on the input parameters of the device control tool 705, the current tool is invoked to obtain the processing result of the current tool.

[0104] For example, the input parameters for the equipment control tool 705 could include: "Time: September 2, 2023; Related equipment: Equipment XX and Equipment XX (these devices satisfy a mapping relationship with Conference Room XX); Adjustment target: Terminal temperature 27℃". The equipment control tool 705 can then obtain the parameters to be adjusted based on the above input parameters and input them into the HVAC system controller. The controller can calculate the control quantities for the relevant equipment, which may include cooling tower frequency, cooling pump frequency, chilled water pump frequency, evaporator outlet water temperature, etc. Then, based on these control quantities, the various devices in the HVAC system are adjusted to achieve HVAC system control.

[0105] It can be seen that the main differences between calling the first tool and calling other tools are: when calling other tools, the processing results of the previous tools need to be provided to the predefined model 702, while when calling the first tool, the processing results of the previous tools are empty.

[0106] Next, the processing result of the equipment control tool 705 is evaluated. The process of evaluating the processing result of the equipment control tool 705 is similar to the process of evaluating the processing result of the query tool 704, and will not be described again in this embodiment.

[0107] Figure 8 is a schematic structural block diagram of a data processing device for a heating, ventilation, and air conditioning system according to an embodiment of the present disclosure.

[0108] As shown in Figure 8, the data processing device 800 of the HVAC system may include a recall module 810, a determination module 820, and a call module 830.

[0109] The recall module 810 is used to recall context information related to the input information from multiple pieces of information in response to receiving input information for the HVAC system.

[0110] The determination module 820 is used to determine N target tools and the calling order of N target tools from multiple candidate tools based on input information and context information, where N is an integer greater than or equal to 1.

[0111] The calling module 830 is used to call N target tools to process the input information based on the calling order, and obtain the processing results of N target tools.

[0112] In this embodiment, the determining module includes a first prompt information determining submodule and a first input submodule. The first prompt information determining submodule is used to determine the first prompt information based on input information, context information, and a first prompt information template. The first input submodule is used to input the first prompt information into a predetermined model to obtain the output information of the predetermined model. The output information includes the identifiers and calling order of N target tools. The first prompt information template includes: functional description information for multiple candidate tools; and first task information, representing the predetermined model's use in selecting target tools and determining the calling order of the target tools.

[0113] In this embodiment, N is an integer greater than or equal to 2; the calling module includes: a current tool determination submodule, a second prompt information determination submodule, a second input determination submodule, and a first result determination submodule. The current tool determination submodule is used to determine the current tool among N target tools according to the calling order. The second prompt information determination submodule is used to determine the second prompt information based on input information, context information, the processing result of the preceding tool, and the second prompt information template in response to the determination of the current tool as the first tool and other tools. The second input determination submodule is used to input the second prompt information into a predetermined model to obtain the input parameters of the current tool. The first result determination submodule is used to call the current tool based on the input parameters of the current tool to obtain the processing result of the current tool. The second prompt information template includes: functional description information for multiple candidate tools; interface parameter description information for multiple candidate tools; and second task information, representing the input parameters used by the predetermined model to determine the current tool.

[0114] In this embodiment, the calling module further includes: a third prompt information determination submodule, a third input submodule, and a first return submodule. The third prompt information determination submodule, after obtaining the processing result of the current tool, determines the third prompt information based on the input information, the processing result of the current tool, and the third prompt information template. The third input submodule inputs the third prompt information into a predetermined model to obtain an evaluation result for the processing result of the current tool. The first return submodule, in response to detecting that the evaluation result is unsuccessful, returns to the operation of the current tool among N target tools until a predetermined termination condition is met. The third prompt information template includes: the calling order of the target tools; attribute information of the expected result of the target tools; and third task information, characterizing whether the predetermined model evaluates whether the input information is consistent with the processing result of the current tool.

[0115] In this embodiment, the predetermined termination condition includes: returning to determine that the number of operations performed by the current tool is greater than or equal to a threshold number.

[0116] In this embodiment, the second prompt information determination submodule includes a summary determination unit and an information determination unit. The summary determination unit determines summary information based on input information, context information, and the processing results of previous tools. The information determination unit determines the second prompt information based on the summary information and the second prompt information template.

[0117] In this embodiment, the summary determination unit includes: a splitting determination subunit, a sub-summary determination subunit, and a summary determination subunit. The splitting determination subunit is used to split the input information, context information, and the processing results of the preceding tools into multiple sub-information. The sub-summary determination subunit is used to determine multiple sub-summary information from the multiple sub-information. The summary determination subunit is used to determine the summary information based on the multiple sub-summary information.

[0118] In this embodiment, the output information of the predetermined model also includes the input parameters for the first tool; the calling module further includes: a calling submodule, used to call the current tool according to the input parameters for the first tool in response to determining that the current tool is the first tool, and obtain the processing result of the current tool.

[0119] In this embodiment, the calling module further includes a second return submodule, which, after obtaining the processing result of the current tool, determines the subsequent tool as the current tool and returns the operation of determining the second prompt information until the processing result of the last target tool is obtained.

[0120] In this embodiment, the input information is the problem information of the HVAC system, and the context information includes at least one of the problem analysis and processing methods.

[0121] According to embodiments of this disclosure, this disclosure also provides an electronic device, including at least one processor; and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the data processing method of the HVAC system described above.

[0122] According to embodiments of this disclosure, this disclosure also provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the data processing method of the HVAC system described above.

[0123] According to embodiments of this disclosure, this disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described data processing method for a heating, ventilation, and air conditioning system.

[0124] Figure 9 is a structural block diagram of an electronic device used to implement the data processing method of an HVAC system according to embodiments of the present disclosure. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0125] As shown in Figure 9, device 900 includes a computing unit 901, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 902 or a computer program loaded into random access memory (RAM) 903 from storage unit 908. RAM 903 can also store various programs and data required for the operation of device 900. The computing unit 901, ROM 902, and RAM 903 are interconnected via bus 904. Input / output (I / O) interface 905 is also connected to bus 904.

[0126] Multiple components in device 900 are connected to I / O interface 905, including: input unit 906, such as keyboard, mouse, etc.; output unit 907, such as various types of monitors, speakers, etc.; storage unit 908, such as disk, optical disk, etc.; and communication unit 909, such as network card, modem, wireless transceiver, etc. Communication unit 909 allows device 900 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0127] The computing unit 901 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as data processing methods for HVAC systems. For example, in some embodiments, the data processing methods for HVAC systems can be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed on device 900 via ROM 902 and / or communication unit 909. When the computer program is loaded into RAM 903 and executed by the computing unit 901, one or more steps of the data processing methods for HVAC systems described above can be performed. Alternatively, in other embodiments, the computing unit 901 can be configured to perform data processing methods for HVAC systems by any other suitable means (e.g., by means of firmware).

[0128] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0129] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0130] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0131] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0132] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0133] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

[0134] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0135] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0136] In the technical solution disclosed herein, the user's authorization or consent is obtained before acquiring or collecting the user's personal information.

[0137] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A data processing method for a heating, ventilation, and air conditioning system, comprising: In response to receiving input information for the HVAC system, context information related to the input information is retrieved from multiple pieces of information; Determining N target tools from a plurality of candidate tools and their invocation order based on the input information and the context information includes: determining a first prompt based on the input information, the context information, and a first prompt template; inputting the first prompt into a predetermined model to obtain output information of the predetermined model, the output information including the identifiers of the N target tools and the invocation order; wherein the first prompt template includes: functional description information for the plurality of candidate tools and first task information, the first task information representing the predetermined model's use in selecting target tools and determining the invocation order of the target tools; N is an integer greater than or equal to 1; and invoking the N target tools to process the input information based on the invocation order to obtain the processing results of the N target tools.

2. The method according to claim 1, wherein, N is an integer greater than or equal to 2; the step of calling the N target tools to process the input information based on the calling order and obtaining the processing results of the N target tools includes: determining the current tool among the N target tools according to the calling order; in response to determining the current tool as one of the other tools after the first tool, determining a second prompt message based on the input information, the context information, the processing result of the preceding tool, and a second prompt message template; inputting the second prompt message into a predetermined model to obtain the input parameters of the current tool; and calling the current tool based on the input parameters of the current tool to obtain the processing result of the current tool; wherein, the second prompt message template includes: functional description information for the multiple candidate tools; interface parameter description information for the multiple candidate tools; and second task information, representing the input parameters of the predetermined model used to determine the current tool.

3. The method according to claim 2, wherein, The step of calling the N target tools to process the input information based on the calling order and obtaining the processing results of the N target tools further includes: after obtaining the processing result of the current tool, determining a third prompt message based on the input information, the processing result of the current tool, and a third prompt message template; inputting the third prompt message into the predetermined model to obtain an evaluation result for the processing result of the current tool; and in response to detecting that the evaluation result is not passed, returning to the operation of determining the current tool among the N target tools until a predetermined termination condition is met; wherein, the third prompt message template includes: the calling order of the target tools; attribute information of the expected result of the target tools; and third task information, characterizing whether the predetermined model is used to evaluate whether the input information is consistent with the processing result of the current tool.

4. The method according to claim 3, wherein, The predetermined termination condition includes: the number of times the current tool has been operated is greater than or equal to a threshold number.

5. The method according to claim 2, wherein, The step of determining the second prompt information based on the input information, the context information, the processing result of the preceding tool, and the second prompt information template includes: determining summary information based on the input information, the context information, and the processing result of the preceding tool; and determining the second prompt information based on the summary information and the second prompt information template.

6. The method according to claim 5, wherein, The step of determining the summary information based on the input information, the context information, and the processing result of the preceding tool includes: splitting the input information, the context information, and the processing result of the preceding tool into multiple sub-information; determining multiple sub-summary information of the multiple sub-information; and determining the summary information based on the multiple sub-summary information.

7. The method according to claim 2, wherein, The output information of the predetermined model also includes input parameters for the first tool; the step of calling the N target tools to process the input information based on the calling order and obtaining the processing results of the N target tools also includes: in response to determining that the current tool is the first tool, calling the current tool according to the input parameters for the first tool and obtaining the processing result of the current tool.

8. The method according to claim 2, further comprising: After obtaining the processing result of the current tool, the next tool will be identified as the current tool, and the operation of confirming the second prompt information will be returned until the processing result of the last target tool is obtained.

9. The method according to claim 1, wherein, The input information is problem information of the HVAC system, and the context information includes at least one of the problem analysis and processing methods.

10. A data processing device for a heating, ventilation, and air conditioning system, comprising: The recall module is used to recall context information related to the input information from multiple pieces of information in response to receiving input information for the HVAC system. The determination module is used to determine N target tools from a plurality of candidate tools and the calling order of the N target tools based on the input information and the context information, where N is an integer greater than or equal to 1; The system also includes a calling module for invoking the N target tools to process the input information based on the calling order, thereby obtaining the processing results of the N target tools. The determining module includes: a first prompt information determining submodule for determining first prompt information based on the input information, the context information, and a first prompt information template; and a first input submodule for inputting the first prompt information into a predetermined model to obtain the output information of the predetermined model, the output information including the identifiers of the N target tools and the calling order. The first prompt information template includes: functional description information for the multiple candidate tools; and first task information, characterizing the predetermined model's use in selecting target tools and determining the calling order of the target tools.

11. The apparatus according to claim 10, wherein, N is an integer greater than or equal to 2; the calling module includes: a current tool determination submodule, used to determine the current tool among the N target tools according to the calling order; a second prompt information determination submodule, used to determine a second prompt information based on the input information, the context information, the processing result of the preceding tool, and the second prompt information template in response to determining the current tool as the first tool and other tools; a second input determination submodule, used to input the second prompt information into a predetermined model to obtain the input parameters of the current tool; and a first result determination submodule, used to call the current tool based on the input parameters of the current tool to obtain the processing result of the current tool; wherein, the second prompt information template includes: functional description information for the multiple candidate tools; interface parameter description information for the multiple candidate tools; and second task information, representing the input parameters used by the predetermined model to determine the current tool.

12. The apparatus according to claim 11, wherein, The invocation module further includes: a third prompt information determination submodule, used to determine a third prompt information based on the input information, the current tool's processing result, and a third prompt information template after obtaining the current tool's processing result; a third input submodule, used to input the third prompt information into the predetermined model to obtain an evaluation result for the current tool's processing result; and a first return submodule, used to return to the operation of determining the current tool among the N target tools in response to detecting that the evaluation result is not passed, until a predetermined termination condition is met; wherein, the third prompt information template includes: the invocation order of the target tools; attribute information of the expected result of the target tools; and third task information, characterizing whether the predetermined model is used to evaluate whether the input information and the current tool's processing result are consistent.

13. The apparatus according to claim 12, wherein, The predetermined termination condition includes: the number of times the current tool has been operated is greater than or equal to a threshold number.

14. The apparatus according to claim 11, wherein, The second prompt information determination submodule includes: a summary determination unit, used to determine summary information based on the input information, the context information, and the processing result of the previous tool; and an information determination unit, used to determine the second prompt information based on the summary information and the second prompt information template.

15. The apparatus according to claim 14, wherein, The summary determination unit includes: a splitting determination subunit, used to split the input information, the context information, and the processing result of the preceding tool into multiple sub-information; a sub-summary determination subunit, used to determine multiple sub-summary information of the multiple sub-information; and a summary determination subunit, used to determine the summary information based on the multiple sub-summary information.

16. The apparatus according to claim 11, wherein, The output information of the predetermined model also includes the input parameters for the first tool; the calling module further includes: a calling submodule, used to call the current tool according to the input parameters for the first tool in response to determining that the current tool is the first tool, and to obtain the processing result of the current tool.

17. The apparatus according to claim 11, wherein the calling module further comprises: The second return submodule is used to determine the next tool as the current tool after obtaining the processing result of the current tool, and return the operation of determining the second prompt information, until the processing result of the last target tool is obtained.

18. The apparatus according to claim 10, wherein, The input information is problem information of the HVAC system, and the context information includes at least one of the problem analysis and processing methods.

19. An electronic device comprising: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 9.

20. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 9.

21. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images