Control method, device and electronic equipment of air conditioner

Abstract

The application provides a control method and device of an air conditioner, a computer readable storage medium and an electronic device. The method comprises: obtaining user interaction data, the user interaction data comprising at least voice data; analyzing the user interaction data through a first language model to obtain expected data, the expected data being data representing the user's adjustment expectation and comprising at least a target temperature, the first language model being obtained by training a plurality of sets of first data, each set of first data comprising: historical user interaction data and historical expected data; generating an initial control program according to the expected data, simulating the initial control program to obtain a simulation result, adjusting the initial control program in the case that the simulation result represents an abnormality of the air conditioner, until the simulation result represents normality of the air conditioner, obtaining a target control program, and executing the target control program to control the air conditioner to operate. Through the application, the problem of poor user experience in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of air conditioner control technology, and more specifically, to an air conditioner control method, apparatus, computer-readable storage medium, and electronic device. Background Technology

[0002] In the current HVAC field, especially in central air conditioning systems, the improvement of intelligence level mostly relies on traditional control algorithms and rudimentary machine learning models. While these technologies have optimized energy use and improved equipment operating efficiency to some extent, they mainly focus on processing structured sensor data, such as environmental parameters like temperature and humidity. In other words, they only operate based on parameters currently set by the user or pre-set parameters, ignoring the potential value of unstructured data. At the same time, they are difficult to handle when the air conditioner malfunctions, resulting in a poor user experience with traditional control methods. Summary of the Invention

[0003] The main objective of this application is to provide a control method, apparatus, computer-readable storage medium, and electronic device for an air conditioner, so as to at least solve the problem of poor user experience in the prior art.

[0004] To achieve the above objectives, according to one aspect of this application, a method for controlling an air conditioner is provided, comprising: acquiring user interaction data, wherein the user interaction data includes at least voice data; analyzing the user interaction data using a first language model to obtain expected data, wherein the expected data is data characterizing the user's adjustment expectations and includes at least a target temperature, the first language model being trained using multiple sets of first data, each set of first data including: historical user interaction data and historical expected data; generating an initial control program based on the expected data; simulating the operation of the initial control program to obtain a simulation result; adjusting the initial control program until the simulation result indicates that the air conditioner is abnormal, if the simulation result indicates that the air conditioner is normal, obtaining a target control program; and executing the target control program to control the operation of the air conditioner.

[0005] Optionally, simulating the initial control program includes: acquiring structural parameters of a predetermined space and location parameters of the air conditioner, wherein the predetermined space represents the space where the air conditioner is located, and the structural parameters include at least area; constructing a virtual operation model based at least on the structural parameters and the location parameters, and simulating the initial control program on the virtual operation model, wherein the virtual operation model is used to simulate the operation of the air conditioner in the predetermined space.

[0006] Optionally, constructing a virtual operating model based at least on the structural parameters and the positional parameters includes: constructing a physical model of the predetermined space based on the structural parameters and the positional parameters of the predetermined space; determining a spatial transformation matrix; converting the physical model into a virtual platform using the spatial transformation matrix, wherein the spatial transformation matrix represents the positional transformation relationship between the physical model and the virtual platform, and the virtual platform includes a data transmission interface; and inputting at least the desired data into the virtual platform through the data transmission interface to obtain the virtual operating model.

[0007] Optionally, generating an initial control program based on the desired data includes: acquiring the current temperature; if the current temperature differs from the target temperature, calculating the absolute value of the difference between the current temperature and the target temperature; generating an initial control strategy based at least on the absolute value of the difference; and generating the initial control program based on the initial control strategy, wherein the initial control strategy includes at least a strategy for adjusting from the current temperature to the target temperature.

[0008] Optionally, after executing the target control program to control the operation of the air conditioner, the method further includes: acquiring user feedback data, wherein the user feedback data represents user feedback on the adjustment results of the air conditioner; analyzing the user feedback data through a second language model to obtain fine-tuning parameters, wherein the fine-tuning parameters represent parameters for adjusting the operation of the air conditioner, the second language model being trained using multiple sets of second data, each set of second data including: historical user feedback data and historical fine-tuning parameters; and controlling the operation of the air conditioner according to the fine-tuning parameters to meet the user's needs.

[0009] Optionally, the method further includes: acquiring operating data and equipment status data of the air conditioner, wherein the operating data includes at least the historical operating frequency of the air conditioner, and the equipment status data represents the operating status of components in the air conditioner; analyzing the operating data and the equipment status data through a first neural network model to obtain a reliability parameter, wherein the reliability parameter represents the reliability of components in the air conditioner, and the first neural network model is trained using multiple sets of third data, each set of third data including: historical operating data, historical equipment status data, and corresponding historical reliability parameters; and generating an alternative control program and executing the alternative control program to improve the reliability of the air conditioner when the reliability parameter is less than a preset threshold.

[0010] Optionally, generating an alternative control program includes: outputting multiple suggested options, allowing a user to select one of the suggested options, wherein the suggested options include at least a replacement option and a retention option, the replacement option indicating the replacement of the air conditioner component whose reliability parameter is less than the preset threshold, and the retention option indicating the retention of the air conditioner component whose reliability parameter is less than the preset threshold; and generating the alternative control program when the user selects the retention option.

[0011] Optionally, the method includes: analyzing the operating data and the device status data using a second neural network model to obtain a predicted energy consumption value, wherein the second neural network model is trained using multiple sets of fourth data, each set of fourth data including: historical operating data, historical device status data, and a corresponding historical predicted energy consumption value; and generating an energy consumption warning message when the predicted energy consumption value is greater than a preset energy consumption value, to alert the air conditioner to high energy consumption.

[0012] According to another aspect of this application, a control device for an air conditioner is provided, comprising: a first acquisition unit for acquiring user interaction data, wherein the user interaction data includes at least voice data; a first analysis unit for analyzing the user interaction data using a first language model to obtain expected data, wherein the expected data is data characterizing the user's adjustment expectations and includes at least a target temperature, the first language model being trained using multiple sets of first data, each set of first data including: historical user interaction data and historical expected data; and an execution unit for generating an initial control program based on the expected data, simulating the operation of the initial control program to obtain a simulation result, adjusting the initial control program until the simulation result indicates that the air conditioner is abnormal, obtaining a target control program, and executing the target control program to control the operation of the air conditioner.

[0013] According to another aspect of this application, a computer-readable storage medium is provided, the computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform any of the control methods of the air conditioner described above.

[0014] According to another aspect of this application, an electronic device is provided, comprising: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including a control method for performing any of the aforementioned air conditioners.

[0015] By applying the technical solution of this application, user interaction data is acquired, analyzed using a first language model to obtain expected data, and an initial control program is generated based on the expected data. The initial control program is then simulated to obtain simulation results. If the simulation results indicate an air conditioner malfunction, the initial control program is adjusted until the simulation results indicate the air conditioner is functioning normally, resulting in a target control program. This target control program is then executed to control the air conditioner's operation. Compared to existing technologies where the air conditioner is directly adjusted based on manually set parameters, leading to a poor user experience, the air conditioner control method of this application can control based on user interaction data and simultaneously simulate the operation results to verify whether the initial control program generated from the expected data will cause air conditioner malfunctions. This avoids air conditioner malfunctions caused by adjustments based on user interaction data, preventing a negative user experience, and eliminates the need for manual adjustments by the user, thus improving the user experience. Therefore, it can solve the problem of poor user experience in existing technologies and achieve the effect of improving user experience. Attached Figure Description

[0016] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0017] Figure 1 A schematic flowchart of a control method for an air conditioner provided in an embodiment of this application is shown;

[0018] Figure 2 A control architecture diagram of a specific air conditioner provided by an embodiment of this application is shown;

[0019] Figure 3 A flowchart illustrating a multimodal alignment training process provided by an embodiment of this application is shown.

[0020] Figure 4 A structural block diagram of a control device for an air conditioner provided in an embodiment of this application is shown. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this application described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0024] As described in the background section, in the prior art, air conditioners adjust directly based on parameters manually set by the user, resulting in a poor user experience. To address this issue, embodiments of this application provide a control method, apparatus, computer-readable storage medium, and electronic device for an air conditioner.

[0025] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0026] This embodiment provides a control method for an air conditioner that runs on a mobile terminal, computer terminal, or similar computing device. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Also, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0027] Figure 1 This is a flowchart of an air conditioner control method according to an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0028] Step S201: Obtain user interaction data, wherein the user interaction data includes at least voice data;

[0029] Specifically, this embodiment no longer relies on manual user settings to control the air conditioner. Instead, it achieves accurate control by combining with a language model, such as a multimodal large language model. Therefore, it is first necessary to acquire user interaction data, including but not limited to voice data and text data. For example, a user might issue a voice command: "I have an important meeting this afternoon, please cool down the temperature in advance but keep it quiet." Besides acquiring user interaction data on-site through the air conditioner itself, it can also be acquired remotely through furniture devices connected to the Internet of Things (IoT) along with the air conditioner. For instance, a user in a different room from the air conditioner could directly say to their mobile phone: "I have an important meeting this afternoon, please cool down the temperature in advance but keep it quiet." Alternatively, other interconnected devices can be used to acquire the data.

[0030] The input layer of the perception network, composed of a multimodal large language model, can accept various types of data, such as text data, voice data, and sensor data. Sensor data is generally structured, including: a ring microphone array supporting 360° voice command capture; thermal imaging sensors for human position and core body temperature detection; UWB (Ultra Wide Band) positioning base stations for accurate personnel positioning; environmental sensor boxes for acquiring temperature, humidity, CO2, etc.; and air vents supporting 0.1° angle adjustment. Unstructured data includes: user voice commands (converted to text via ASR), maintenance work order text (fault descriptions, repair records), and Building Information Modeling (BIM) models (spatial topology and thermal parameters), etc. Multiple data input interfaces are provided. Multimodal alignment training is conducted. Multimodal data alignment refers to establishing and discovering correspondences between different data modalities, thereby achieving effective integration and utilization of information. A cross-modal contrastive learning loss function is designed to align text descriptions with sensor data. For example, if the text is "insufficient cooling", the corresponding sensor data could be a decrease in evaporator pressure or a decrease in current.

[0031] Step S202: Analyze the user interaction data using a first language model to obtain expected data, wherein the expected data is data representing the user's adjustment expectations and includes at least the target temperature. The first language model is trained using multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data.

[0032] Specifically, the language model can be a multimodal large language model integrating multiple functions. This multimodal large language model includes a first language model part that realizes user interaction data analysis. After analyzing the user interaction data through the first language model, the user's expected data can be obtained. For example, analyzing the user's voice command "I have an important meeting this afternoon, cool down the temperature in advance but keep quiet," the expected data obtained are: target temperature 24℃, target time period 14:00-16:00, and target operating mode silent mode, etc. The first language model is trained by machine learning on the basis of a large amount of historical data, including past user interaction records (such as user commands issued by voice) and corresponding expected data (such as the specific temperature or operating mode that the user wants to set). Such a training dataset allows the model to understand and learn the correspondence between various user commands and expected results, so that when new user data is received in the future, it can accurately parse the user's intent and convert it into control parameters that the system can execute. In this way, through a deep understanding of the user's natural language, the system can respond to the user's needs in a more human and intuitive way, without requiring the user to understand the specific technical details or control parameters.

[0033] Step S203: Generate an initial control program based on the expected data, simulate the operation of the initial control program, obtain the simulation operation results, and adjust the initial control program until the simulation operation results indicate that the air conditioner is abnormal, so as to obtain a target control program and execute the target control program to control the operation of the air conditioner.

[0034] Specifically, after obtaining the desired data through analysis, it is converted into an initial control program. The energy efficiency and reliability of the initial control program's control strategy are then simulated. First, its feasibility is determined, and it is assessed whether the air conditioner will be damaged during operation. That is, if the simulation results indicate an air conditioner malfunction, the initial control program is adjusted. For example, if a user issues the instruction: "A major client is coming for a meeting next Wednesday at 2 PM. Cool down the first meeting room beforehand, but don't let the air conditioner hum," the program first checks the schedule to confirm an important meeting in the meeting room next Wednesday; then it checks the weather forecast and finds a sunny day with a temperature of 35°C; simultaneously, based on peak and off-peak electricity pricing, it automatically selects the cheaper midnight hours for early cooling; then, the initial control program is simulated. If it is found that "early cooling will cause pipes to freeze," indicating an air conditioner malfunction, the plan is immediately adjusted; this process continues until a method that saves energy without damaging the air conditioner is found, resulting in the target control program. On the day of the meeting, the meeting room temperature automatically drops to 24°C; before the meeting, the fan speed is quietly lowered to ensure quiet operation.

[0035] This embodiment acquires user interaction data, analyzes it using a first language model to obtain expected data, generates an initial control program based on the expected data, simulates the operation of the initial control program, and obtains the simulation results. If the simulation results indicate that the air conditioner is malfunctioning, the initial control program is adjusted until the simulation results indicate that the air conditioner is functioning normally, resulting in a target control program. The target control program is then executed to control the operation of the air conditioner. Compared to existing technologies where the air conditioner is directly adjusted based on manually set parameters by the user, leading to a poor user experience, the air conditioner control method of this application can control based on user interaction data and simulate the operation results to verify whether the initial control program generated from the expected data will cause the air conditioner to malfunction. This avoids causing air conditioner malfunctions and negative user experiences during adjustments based on user interaction data, and eliminates the need for manual adjustments by the user, thus improving the user experience. Therefore, it can solve the problem of poor user experience in existing technologies and achieve the effect of improving user experience.

[0036] In the specific implementation process, the above step S203, simulating the initial control program, can be achieved through the following steps: Step S2031: Obtain the structural parameters of the predetermined space and the location parameters of the air conditioner, wherein the predetermined space represents the space where the air conditioner is located, and the structural parameters include at least the area; Step S2032: Construct a virtual operation model based at least on the structural parameters and the location parameters, and simulate the initial control program on the virtual operation model, wherein the virtual operation model is used to simulate the operation of the air conditioner in the predetermined space. This method, through the above steps, constructs an accurate virtual operation model and simulates its operation, enabling pre-evaluation of the control program's performance in the actual environment, ensuring that it meets user needs without damaging the air conditioning equipment.

[0037] Specifically, before simulation, it's necessary to collect detailed information about the environment in which the air conditioner is located. Structural parameters encompass basic spatial attributes such as room area, volume, ceiling height, and window location and size—essential for understanding the space's thermodynamic characteristics. Location parameters accurately record the specific layout of the air conditioning equipment within the space, including its location, orientation, and relative distances to walls, doors, and windows. This detailed data ensures the virtual model accurately reflects the physical characteristics of the real environment, thereby improving the reliability of the simulation results. Based on the collected structural and location parameters, a virtual operating model is constructed. This "virtual operating model" is essentially part of a building digital twin, capable of simulating the air conditioner's operation in a specific environment, including temperature distribution, humidity changes, airflow patterns, and energy consumption. Once the model is built, an initial control program derived from user commands can be simulated and run on it. In this way, the effects of the control program can be predicted and its potential for equipment malfunction or poor performance can be assessed without interfering with the actual equipment operation.

[0038] In some optional implementations, step S2032 can be achieved through the following steps: Step S2033: Construct a physical model of the predetermined space based on the structural parameters and positional parameters of the predetermined space; Step S2034: Determine a spatial transformation matrix, and convert the physical model into a virtual platform using the spatial transformation matrix, wherein the spatial transformation matrix represents the positional transformation relationship between the physical model and the virtual platform, and the virtual platform includes a data transmission interface; Step S2035: Input at least the desired data into the virtual platform through the data transmission interface to obtain the virtual operating model. This method constructs a virtual operating model in a digital twin environment through the above steps and maps the user's desired data into this model to achieve accurate simulation and strategy verification.

[0039] Specifically, the system needs to establish a physical model that accurately reflects the actual thermodynamic characteristics of the space based on the structural parameters (including but not limited to area, volume, wall location, window size, etc.) of the predetermined space (i.e., the area served by the air conditioner, such as offices, meeting rooms, etc.) and the specific location information of the air conditioning equipment (referring to the exact layout of the air conditioner in the space, including but not limited to the relative distance between the air conditioner and the walls and windows, the orientation of the air conditioner, etc.). Determining the spatial transformation matrix: Next, in order to use this physical model in the virtual platform (digital twin environment), a spatial transformation matrix must be determined. The specific steps for mathematical modeling the spatial transformation matrix are as follows: Step 1: Define the coordinate system. UWB coordinate system: with the center of the UWB base station network as the origin, usually a three-dimensional rectangular coordinate system (x, y, z); BIM coordinate system: with the global origin of the building BIM model as the reference (such as the southwest corner of the building). Step 2: Collect calibration points. Select at least 4 non-coplanar calibration points within the building (it is recommended to select equipment installation locations, such as air conditioning units and fan coil units); simultaneously record the coordinates of these points in the UWB coordinate system and the BIM coordinate system. Step 3: Calculate the rotation matrix R and translation vector T. The optimal rigid body transformation can be solved using the least squares method: First, center the coordinates, calculate the centroids of the calibration points in the UWB and BIM coordinate systems, then construct the covariance matrix and perform singular value decomposition, calculate the translation vector and transformation matrix, and finally obtain the mapping formula from any UWB coordinate point (x, y, z) to BIM coordinates (X, Y, Z). This matrix defines the positional transformation relationship between the physical model and the virtual platform, ensuring that every point in the physical model (including the location of the air conditioner and other feature points in the space) can be accurately mapped to the virtual platform. Establishing the spatial transformation matrix is ​​a key technical point for achieving data synchronization and strategy pre-simulation between the digital twin model and the real environment. Input desired data using the data transmission interface: Once the virtual platform is ready, the system imports the user's desired data (such as target temperature, noise requirements, energy-saving targets, etc.) into the virtual operating model through the data transmission interface in the virtual platform. In this way, the user's needs can be accurately simulated and pre-simulated in the digital twin environment, thereby verifying the effectiveness and rationality of the control strategy.

[0040] In some alternative implementations, step S203, which generates an initial control program based on the desired data, can be achieved through the following steps: Step S2036: Obtain the current temperature; if the current temperature differs from the target temperature, calculate the absolute value of the difference between the current temperature and the target temperature; Step S2037: Generate an initial control strategy based at least on the absolute value of the difference; generate the initial control program based on the initial control strategy, wherein the initial control strategy at least includes a strategy for adjusting the current temperature to the target temperature. This method achieves end-to-end mapping from user natural language commands to specific air conditioner control programs through the above steps, increasing the diversity of air conditioner control, eliminating the need for manual user settings, and improving the user experience.

[0041] In practice, the system acquires the current ambient temperature in real time using sensors, providing the foundational data for responding to user temperature adjustment requests. Subsequently, it calculates the difference between the user's desired target temperature and the current temperature, taking its absolute value. This step quantifies the temperature difference, providing a numerical basis for subsequent strategy generation. Based on this absolute temperature difference, the system generates an initial control strategy—how to gradually adjust from the current temperature to the target temperature. Strategy generation may consider the air conditioner's heating or cooling capabilities to ensure rapid and smooth temperature adjustment. Furthermore, there is usually multiple strategies to ensure switching between them if one fails. Finally, based on the established initial control strategy, the system translates it into a series of specific instructions executable by the air conditioner, forming the initial control program. These instructions may include adjusting air conditioner settings, such as changing cooling or heating power, adjusting fan speed, or changing operating modes, to achieve precise temperature control. For example, a user instruction like "I have an important meeting this afternoon; cool down the temperature beforehand but keep it quiet" can be translated into the following control instructions:

[0042] set_temp(22);

[0043] if time.between("14:00","16:00");

[0044] fan_mode("silent");

[0045] compressor_freq(40Hz).

[0046] In some optional implementations, the method further includes step S204: after executing the target control program to control the operation of the air conditioner, acquiring user feedback data, wherein the user feedback data represents the user's feedback on the adjustment results of the air conditioner; step S205: analyzing the user feedback data through a second language model to obtain fine-tuning parameters, wherein the fine-tuning parameters represent parameters for adjusting the operation of the air conditioner, the second language model being trained through multiple sets of second data, each set of second data including: historical user feedback data and historical fine-tuning parameters; step S206: controlling the operation of the air conditioner according to the fine-tuning parameters to meet the user's needs. This method, through a closed-loop feedback mechanism, can dynamically adjust the air conditioning control strategy based on real-time user feedback, making the user experience more personalized and comfortable.

[0047] Specifically, during the operation of the air conditioner, users may provide real-time feedback on their current feelings. For example: (1) Taking the fault diagnosis and self-repair scenario as an example, if a user reports "there is a sour smell on the east side of the office" via voice, the visual sensor will be called to scan the east side air vent; the knowledge graph will be searched to find that "sour smell + condensate pH < 5" matches the corrosion of the copper pipe of the condenser, generating fine-tuning parameters and generating the corresponding repair code: raising the pH of the chilled water to 8.5, and finally a work order can be dispatched to remind the user to check the corrosion inhibitor. (2) Taking the user's real-time feedback as an example, if a user says, "the air conditioner is always blowing on my head", the second language model part of the multimodal large language model will automatically learn and analyze the location of the person's workstation. It will also find out whether the air vent is crooked by checking the maintenance records and issue a warning, notifying the maintenance personnel to come and fix it precisely the next day with tools. The second language model is trained based on a large amount of historical user feedback data and corresponding fine-tuning parameters. It can understand the user's natural language feedback and convert it into specific technical parameter adjustments, i.e., fine-tuning parameters. These parameters may include temperature setting, fan speed, operating mode, etc. In addition to user feedback via voice, historical user feedback data can also include fault repair work orders and fault repair records, so that the second language model can learn the solutions to historical problems, more accurately parse user feedback data, and generate corresponding fine-tuning parameters.

[0048] In some alternative embodiments, the method further includes step S207: acquiring the operating data and equipment status data of the air conditioner, wherein the operating data includes at least the historical operating frequency of the air conditioner, and the equipment status data represents the operating status of the components in the air conditioner; step S208: analyzing the operating data and the equipment status data through a first neural network model to obtain a reliability parameter, wherein the reliability parameter represents the reliability of the components in the air conditioner, and the first neural network model is trained using multiple sets of third data, each set of third data including: historical operating data, historical equipment status data, and corresponding historical reliability parameters; step S209: if the reliability parameter is less than a preset threshold, generating an alternative control program and executing the alternative control program to improve the reliability of the air conditioner. This method predicts the reliability of the equipment by analyzing the equipment status in real time through the above steps, ensuring that the air conditioner can operate stably.

[0049] Specifically, the above steps aim to analyze the reliability of the air conditioner through operating data and equipment status data, which can also be understood as the remaining lifespan. Historical equipment status data may include: (1) Control parameters: set temperature, fan speed, valve opening; data source can be the historical log of the air conditioning control system. (2) Environmental parameters: indoor and outdoor temperature and humidity, personnel density, solar radiation; data source is the building BIM model and meteorological database. (3) Equipment status: compressor frequency, refrigerant pressure, current; data source is real-time monitoring by sensors. Historical operating data may include: time series characteristics: energy consumption trend and load fluctuation in the past 1 hour; data source can be a time series database. Corresponding historical reliability parameters: (1) Direct observation values: real-time energy consumption (kW), outlet air temperature (°C), used for supervised learning of data fitting terms; (2) Physical field distribution: evaporator surface temperature field, duct velocity field, used for physical residual term constraints. (3) Equipment health indicators: compressor bearing wear (0-1), refrigerant charge, used for reliability prediction and lifespan modeling. When the reliability parameter is less than a preset threshold, an alternative control program is generated and executed to improve the reliability of the air conditioner. For example, when the predicted lifespan of a component is lower than a safety threshold, the control strategy needs to be automatically adjusted, i.e., the predicted equipment lifespan loss is greater than the threshold, and the following alternative solutions are automatically generated: 1. Alternative solution for dynamic adjustment of operating parameters: that is, extending the remaining lifespan by reducing the component load or switching the operating mode. For example, the alternative solution for compressor lifespan warning (remaining lifespan < 30 days) is to reduce the number of compressor start-stop cycles during peak daytime loads, which is expected to extend the lifespan by 50%. For example, the alternative solution for fan bearing wear (vibration value > 5 mm / s) is to reduce the fan speed from 1200 rpm to 900 rpm and link the adjacent area fans to increase the speed to compensate for the air volume. 2. Human-machine collaboration solution: convert the technical solution into natural language suggestions and negotiate with the user for implementation. For example, if the filter lifespan is exhausted (usage time > 500 hours), the system will automatically reduce the fresh air ratio from 30% to 15% (to reduce dust entry), and then push a message to the user: "The filter in the East Zone has expired and is recommended to be replaced immediately. Temporary solution: reduce the fresh air volume, the indoor CO2 concentration will rise to 800ppm (still in compliance with national standards). Do you agree? Let the user choose

Replace immediately

Delay to 18:00

[0050] In some alternative implementations, step S209 can be achieved through the following steps: Step S2091: Output multiple suggested options, allowing the user to select one of the suggested options, wherein the suggested options include at least a replacement option and a retention option, the replacement option indicating replacement of the air conditioner component whose reliability parameter is less than the preset threshold, and the retention option indicating retention of the air conditioner component whose reliability parameter is less than the preset threshold; Step S2092: If the user selects the retention option, generate the alternative control program. This method generates alternative control programs through user participation in decision-making via the above steps, thereby balancing equipment maintenance and user experience.

[0051] Specifically, the system outputs multiple suggested options for the user to choose from: When the first neural network model analyzes that the reliability parameters of one or more components are below a preset threshold, the system generates and displays multiple suggested options to the user. These options typically include replacement and retention options, allowing the user to make a decision based on the current situation and preferences. The system suggests replacing the unreliable component to ensure system stability and extend equipment lifespan. Alternative control strategies are provided for situations where immediate component replacement is not necessary, such as reducing the component's load, adjusting the operating mode to reduce wear, or activating a backup component. When the user selects the retention option, an alternative control program is generated: If the user selects the retention option, an alternative control program is generated based on this decision, aiming to reduce the risk of component failure by changing the control strategy, while minimizing the impact on user comfort and system performance.

[0052] In some optional implementations, the method further includes step S210: analyzing the operating data and the device status data using a second neural network model to obtain a predicted energy consumption value, wherein the second neural network model is trained using multiple sets of fourth data, each set of fourth data including: historical operating data, historical device status data, and a corresponding historical predicted energy consumption value; step S211: generating an energy consumption warning message when the predicted energy consumption value is greater than a preset energy consumption value, to alert the air conditioner to high energy consumption. This method, by predicting energy consumption values, can balance energy consumption with user comfort.

[0053] Specifically, using the aforementioned historical operating data and historical equipment status data, energy consumption can be predicted through a second neural network model. For example, it can output the predicted energy consumption for the next hour, recommend optimal energy efficiency control parameters (such as chilled water temperature setpoint), and generate energy consumption warning information to alert the air conditioner to high energy consumption. Furthermore, a causal model can be constructed to analyze factors that significantly impact energy consumption, such as the influence intensity of "filter cleaning" on energy consumption (Δ = 0.35).

[0054] To enable those skilled in the art to better understand the technical solution of this application, the implementation process of the air conditioner control method of this application will be described in detail below with reference to specific embodiments.

[0055] This embodiment relates to a specific control architecture diagram of an air conditioner, such as... Figure 2 As shown, it includes the following steps:

[0056] Step S1: Input voice, sensor, text, positioning, BIM and image data through the multimodal input layer; sensor data is stored as Kafka data and transmitted to the edge computing layer through real-time data stream;

[0057] Step S2: Perform local speech recognition, sensor data alignment, and lightweight model inference at the edge computing layer to obtain a draft control instruction (initial control program);

[0058] Step S3: Further perform multimodal alignment training, policy generation and optimization, and causal inference through the causal inference engine in the cloud-based large model inference layer;

[0059] Step S4: Perform digital twin verification through the digital twin verification layer, including virtual air conditioning system simulation, energy efficiency / lifetime prediction, and policy security verification;

[0060] Step S5: The final execution command is executed through the physical execution layer to realize air outlet control, compressor frequency regulation and energy storage system management.

[0061] This embodiment relates to a multimodal alignment training flowchart, such as... Figure 3 As shown, it includes the following steps:

[0062] Step S6: Encode the text data using a text encoder, extract speech features from the speech data, and encode the sensor data using a sensor encoder.

[0063] Step S7: Cross-modal contrastive learning;

[0064] Step S8: Joint optimization;

[0065] Step S9: Comparative evaluation;

[0066] Step S10: Model output.

[0067] This application also provides a control device for an air conditioner. It should be noted that the control device for the air conditioner in this application can be used to execute the control method for an air conditioner provided in this application. This device is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0068] The control device for the air conditioner provided in the embodiments of this application will be described below.

[0069] Figure 4 A schematic diagram of the control device for an air conditioner according to an embodiment of this application. (See diagram below.) Figure 4 The device includes:

[0070] The first acquisition unit 10 is used to acquire user interaction data, wherein the user interaction data includes at least voice data;

[0071] Specifically, this embodiment no longer relies on manual user settings to control the air conditioner. Instead, it achieves accurate control by combining with a language model, such as a multimodal large language model. Therefore, it is first necessary to acquire user interaction data, including but not limited to voice data and text data. For example, a user might issue a voice command: "I have an important meeting this afternoon, please cool down the temperature in advance but keep it quiet." Besides acquiring user interaction data on-site through the air conditioner itself, it can also be acquired remotely through furniture devices connected to the Internet of Things (IoT) along with the air conditioner. For instance, a user in a different room from the air conditioner could directly say to their mobile phone: "I have an important meeting this afternoon, please cool down the temperature in advance but keep it quiet." Alternatively, other interconnected devices can be used to acquire the data.

[0072] The input layer of the perception network, composed of a multimodal large language model, can accept various types of data, such as text data, voice data, and sensor data. Sensor data is generally structured, including: a ring microphone array supporting 360° voice command capture; thermal imaging sensors for human position and core body temperature detection; UWB positioning base stations for accurate personnel positioning; environmental sensor boxes for acquiring temperature, humidity, CO2, etc.; and air vents supporting 0.1° angle adjustment. Unstructured data includes user voice commands (converted to text via ASR), maintenance work order texts (fault descriptions, maintenance records), and building BIM models (spatial topology and thermal parameters). Multiple data input interfaces are provided. Multimodal alignment training is performed. Multimodal data alignment refers to establishing and discovering correspondences between different data modalities, thereby achieving effective information integration and utilization. A cross-modal contrastive learning loss function is designed to align text descriptions with sensor data. For example, the text "insufficient cooling" could correspond to sensor data such as decreased evaporator pressure and decreased current.

[0073] The first analysis unit 20 is used to analyze the user interaction data through a first language model to obtain expected data, wherein the expected data is data that characterizes the user's adjustment expectations and includes at least the target temperature. The first language model is trained by multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data.

[0074] Specifically, the language model can be a multimodal large language model integrating multiple functions. This multimodal large language model includes a first language model part that realizes user interaction data analysis. After analyzing the user interaction data through the first language model, the user's expected data can be obtained. For example, analyzing the user's voice command "I have an important meeting this afternoon, cool down the temperature in advance but keep quiet," the expected data obtained are: target temperature 24℃, target time period 14:00-16:00, and target operating mode silent mode, etc. The first language model is trained by machine learning on the basis of a large amount of historical data, including past user interaction records (such as user commands issued by voice) and corresponding expected data (such as the specific temperature or operating mode that the user wants to set). Such a training dataset allows the model to understand and learn the correspondence between various user commands and expected results, so that when new user data is received in the future, it can accurately parse the user's intent and convert it into control parameters that the system can execute. In this way, through a deep understanding of the user's natural language, the system can respond to the user's needs in a more human and intuitive way, without requiring the user to understand the specific technical details or control parameters.

[0075] The execution unit 30 is configured to generate an initial control program based on the expected data, simulate the operation of the initial control program, obtain simulation results, and adjust the initial control program if the simulation results indicate that the air conditioner is abnormal, until the simulation results indicate that the air conditioner is normal, thereby obtaining a target control program, and executing the target control program to control the operation of the air conditioner.

[0076] Specifically, after parsing the desired data, it is converted into an initial control program. The energy efficiency and reliability of the initial control program's control strategy are then simulated. First, its feasibility is determined, and it is assessed whether the air conditioner will be damaged during operation. That is, if the simulation results indicate an air conditioner malfunction, the initial control program is adjusted. For example, if a user issues the instruction: "A major client is coming for a meeting next Wednesday at 2 PM. Cool down the first meeting room beforehand, but don't let the air conditioner hum." The program first checks the schedule to confirm an important meeting next Wednesday; then it checks the weather forecast, finding a sunny day with a 35°C. Simultaneously, based on peak-valley electricity pricing, it automatically selects the cheaper midnight hours for early cooling. The initial control program is then simulated. If "early cooling will cause pipes to freeze," indicating an air conditioner malfunction, the program is immediately adjusted. This process continues until a device that saves energy without damaging the air conditioner is found, resulting in the target control program. On the day of the meeting, the meeting room temperature automatically drops to 24°C; before the meeting, the fan speed is quietly lowered to ensure quiet operation.

[0077] This embodiment acquires user interaction data, analyzes it using a first language model to obtain expected data, generates an initial control program based on the expected data, simulates the operation of the initial control program, and obtains simulation results. If the simulation results indicate an air conditioner malfunction, the initial control program is adjusted until the simulation results indicate the air conditioner is functioning normally, resulting in a target control program. The target control program is then executed to control the air conditioner's operation. Compared to existing technologies where the air conditioner is adjusted directly based on manually set parameters, leading to a poor user experience, the air conditioner control device of this application can control based on user interaction data and simulate the operation results to verify whether the initial control program generated from the expected data will cause the air conditioner to malfunction. This avoids causing air conditioner malfunctions and a poor user experience during adjustments based on user interaction data, and eliminates the need for manual adjustments by the user, thus improving the user experience. Therefore, it solves the problem of poor user experience in existing technologies and achieves the effect of improving user experience.

[0078] In its specific implementation, the aforementioned execution unit includes a first acquisition module and a first operation module. The first acquisition module is used to acquire the structural parameters of a predetermined space and the position parameters of the air conditioner, wherein the predetermined space represents the space where the air conditioner is located, and the structural parameters include at least area. The first operation module is used to construct a virtual operation model based at least on the structural parameters and the position parameters, and to simulate the initial control program on the virtual operation model, wherein the virtual operation model is used to simulate the operation of the air conditioner in the predetermined space. By constructing an accurate virtual operation model and simulating its operation, this device can pre-evaluate the performance of the control program in the actual environment, ensuring that it meets user needs without damaging the air conditioning equipment.

[0079] Specifically, before simulation, it's necessary to collect detailed information about the environment in which the air conditioner is located. Structural parameters encompass basic spatial attributes such as room area, volume, ceiling height, and window location and size—essential for understanding the space's thermodynamic characteristics. Location parameters accurately record the specific layout of the air conditioning equipment within the space, including its location, orientation, and relative distances to walls, doors, and windows. This detailed data ensures the virtual model accurately reflects the physical characteristics of the real environment, thereby improving the reliability of the simulation results. Based on the collected structural and location parameters, a virtual operating model is constructed. This "virtual operating model" is essentially part of a building digital twin, capable of simulating the air conditioner's operation in a specific environment, including temperature distribution, humidity changes, airflow patterns, and energy consumption. Once the model is built, an initial control program derived from user commands can be simulated and run on it. In this way, the effects of the control program can be predicted and its potential for equipment malfunction or poor performance can be assessed without interfering with the actual equipment operation.

[0080] In some optional implementations, the first operation module includes a construction submodule, a transformation submodule, and an input submodule. The construction submodule is used to construct a physical model of the predetermined space based on the structural parameters and positional parameters of the predetermined space. The transformation submodule is used to determine a spatial transformation matrix and convert the physical model into a virtual platform through the spatial transformation matrix, wherein the spatial transformation matrix represents the positional transformation relationship between the physical model and the virtual platform, and the virtual platform includes a data transmission interface. The input submodule is used to input at least the desired data into the virtual platform through the data transmission interface to obtain the virtual operation model. This device constructs a virtual operation model in a digital twin environment through the above steps and maps the user's desired data into this model to achieve accurate simulation and strategy verification.

[0081] Specifically, the system needs to establish a physical model that accurately reflects the actual thermodynamic characteristics of the space based on the structural parameters (including but not limited to area, volume, wall location, window size, etc.) of the predetermined space (i.e., the area served by the air conditioner, such as offices, meeting rooms, etc.) and the specific location information of the air conditioning equipment (referring to the exact layout of the air conditioner in the space, including but not limited to the relative distance between the air conditioner and the walls and windows, the orientation of the air conditioner, etc.). Determining the spatial transformation matrix: Next, in order to use this physical model in the virtual platform (digital twin environment), a spatial transformation matrix must be determined. The specific steps for mathematical modeling the spatial transformation matrix are as follows: Step 1: Define the coordinate system. UWB coordinate system: with the center of the UWB base station network as the origin, usually a three-dimensional rectangular coordinate system (x, y, z); BIM coordinate system: with the global origin of the building BIM model as the reference (such as the southwest corner of the building). Step 2: Collect calibration points. Select at least 4 non-coplanar calibration points within the building (it is recommended to select equipment installation locations, such as air conditioning units and fan coil units); simultaneously record the coordinates of these points in the UWB coordinate system and the BIM coordinate system. Step 3: Calculate the rotation matrix R and translation vector T. The optimal rigid body transformation can be solved using the least squares method: First, center the coordinates, calculate the centroids of the calibration points in the UWB and BIM coordinate systems, then construct the covariance matrix and perform singular value decomposition, calculate the translation vector and transformation matrix, and finally obtain the mapping formula from any UWB coordinate point (x, y, z) to BIM coordinates (X, Y, Z). This matrix defines the positional transformation relationship between the physical model and the virtual platform, ensuring that every point in the physical model (including the location of the air conditioner and other feature points in the space) can be accurately mapped to the virtual platform. Establishing the spatial transformation matrix is ​​a key technical point for achieving data synchronization and strategy pre-simulation between the digital twin model and the real environment. Input desired data using the data transmission interface: Once the virtual platform is ready, the system imports the user's desired data (such as target temperature, noise requirements, energy-saving targets, etc.) into the virtual operating model through the data transmission interface in the virtual platform. In this way, the user's needs can be accurately simulated and pre-simulated in the digital twin environment, thereby verifying the effectiveness and rationality of the control strategy.

[0082] In some alternative implementations, the execution unit further includes a calculation module and a generation module. The calculation module is used to acquire the current temperature and, if the current temperature differs from the target temperature, calculate the absolute value of the difference between the current temperature and the target temperature. The generation module is used to generate an initial control strategy based at least on the absolute value of the difference, and to generate an initial control program based on the initial control strategy. The initial control strategy at least includes a strategy for adjusting the current temperature to the target temperature. This device achieves end-to-end mapping from user natural language commands to specific air conditioner control programs through the above steps, increasing the diversity of air conditioner control, eliminating the need for manual user settings, and improving the user experience.

[0083] In practice, the system acquires the current ambient temperature in real time using sensors, providing the foundational data for responding to user temperature adjustment requests. Subsequently, it calculates the difference between the user's desired target temperature and the current temperature, taking its absolute value. This step quantifies the temperature difference, providing a numerical basis for subsequent strategy generation. Based on this absolute temperature difference, the system generates an initial control strategy—how to gradually adjust from the current temperature to the target temperature. Strategy generation may consider the air conditioner's heating or cooling capabilities to ensure rapid and smooth temperature adjustment. Furthermore, there is usually multiple strategies to ensure switching between them if one fails. Finally, based on the established initial control strategy, the system translates it into a series of specific instructions executable by the air conditioner, forming the initial control program. These instructions may include adjusting air conditioner settings, such as changing cooling or heating power, adjusting fan speed, or changing operating modes, to achieve precise temperature control. For example, a user instruction like "I have an important meeting this afternoon; cool down the temperature beforehand but keep it quiet" can be translated into the following control instructions:

[0084] set_temp(22);

[0085] if time.between("14:00","16:00");

[0086] fan_mode("silent");

[0087] compressor_freq(40Hz).

[0088] In some optional embodiments, the device further includes a second acquisition unit, a second analysis unit, and a first operation unit. The second acquisition unit is used to acquire user feedback data after executing the target control program to control the operation of the air conditioner. The user feedback data represents the user's feedback on the adjustment results of the air conditioner. The second analysis unit is used to analyze the user feedback data using a second language model to obtain fine-tuning parameters. The fine-tuning parameters represent parameters for adjusting the operation of the air conditioner. The second language model is trained using multiple sets of second data, each set including historical user feedback data and historical fine-tuning parameters. The first operation unit is used to control the operation of the air conditioner according to the fine-tuning parameters to meet the user's needs. This device, through a closed-loop feedback mechanism, can dynamically adjust the air conditioning control strategy based on real-time user feedback, making the user experience more personalized and comfortable.

[0089] Specifically, during the operation of the air conditioner, users may provide real-time feedback on their current feelings. For example: (1) Taking the fault diagnosis and self-repair scenario as an example, if a user reports "there is a sour smell on the east side of the office" via voice, the visual sensor will be called to scan the east side air vent; the knowledge graph will be searched to find that "sour smell + condensate pH < 5" matches the corrosion of the copper pipe of the condenser, generating fine-tuning parameters and generating the corresponding repair code: raising the pH of the chilled water to 8.5, and finally a work order can be dispatched to remind the user to check the corrosion inhibitor. (2) Taking the user's real-time feedback as an example, if a user says, "the air conditioner is always blowing on my head", the second language model part of the multimodal large language model will automatically learn and analyze the location of the person's workstation. It will also find out whether the air vent is crooked by checking the maintenance records and issue a warning, notifying the maintenance personnel to come and fix it precisely the next day with tools. The second language model is trained based on a large amount of historical user feedback data and corresponding fine-tuning parameters. It can understand the user's natural language feedback and convert it into specific technical parameter adjustments, i.e., fine-tuning parameters. These parameters may include temperature setting, fan speed, operating mode, etc. In addition to user feedback via voice, historical user feedback data can also include fault repair work orders and fault repair records, so that the second language model can learn the historical problem-solving methods to more accurately parse user feedback data and generate corresponding fine-tuning parameters.

[0090] In some alternative embodiments, the device further includes a second acquisition unit, a third analysis unit, and a second operation unit. The second acquisition unit acquires the air conditioner's operating data and equipment status data, wherein the operating data includes at least the air conditioner's historical operating frequency, and the equipment status data represents the operating status of components in the air conditioner. The third analysis unit analyzes the operating data and equipment status data using a first neural network model to obtain reliability parameters, wherein the reliability parameters represent the reliability of components in the air conditioner. The first neural network model is trained using multiple sets of third data, each set of third data including: historical operating data, historical equipment status data, and corresponding historical reliability parameters. The second operation unit generates an alternative control program and executes the alternative control program to improve the air conditioner's reliability when the reliability parameters are less than a preset threshold. This device predicts equipment reliability by analyzing equipment status in real time through the above steps, ensuring stable operation of the air conditioner.

[0091] Specifically, the above steps aim to analyze the reliability of the air conditioner through operating data and equipment status data, which can also be understood as the remaining lifespan. Historical equipment status data may include: (1) Control parameters: set temperature, fan speed, valve opening; data source can be the historical log of the air conditioning control system. (2) Environmental parameters: indoor and outdoor temperature and humidity, personnel density, solar radiation; data source is the building BIM model and meteorological database. (3) Equipment status: compressor frequency, refrigerant pressure, current; data source is real-time monitoring by sensors. Historical operating data may include: time series characteristics: energy consumption trend and load fluctuation in the past 1 hour; data source can be a time series database. Corresponding historical reliability parameters: (1) Direct observation values: real-time energy consumption (kW), outlet air temperature (°C), used for supervised learning of data fitting terms; (2) Physical field distribution: evaporator surface temperature field, duct velocity field, used for physical residual term constraints. (3) Equipment health indicators: compressor bearing wear (0-1), refrigerant charge, used for reliability prediction and lifespan modeling. When the reliability parameter is less than a preset threshold, an alternative control program is generated and executed to improve the reliability of the air conditioner. For example, when the predicted lifespan of a component is lower than a safety threshold, the control strategy needs to be automatically adjusted, i.e., the predicted equipment lifespan loss is greater than the threshold, and the following alternative solutions are automatically generated: 1. Alternative solution for dynamic adjustment of operating parameters: that is, extending the remaining lifespan by reducing the component load or switching the operating mode. For example, the alternative solution for compressor lifespan warning (remaining lifespan < 30 days) is to reduce the number of compressor start-stop cycles during peak daytime loads, which is expected to extend the lifespan by 50%. For example, the alternative solution for fan bearing wear (vibration value > 5 mm / s) is to reduce the fan speed from 1200 rpm to 900 rpm and link the adjacent area fans to increase the speed to compensate for the air volume. 2. Human-machine collaboration solution: convert the technical solution into natural language suggestions and negotiate with the user for implementation. For example, if the filter lifespan is exhausted (usage time > 500 hours), the system will automatically reduce the fresh air ratio from 30% to 15% (to reduce dust entry), and then push a message to the user: "The filter in the East Zone has expired and is recommended to be replaced immediately. Temporary solution: reduce the fresh air volume, the indoor CO2 concentration will rise to 800ppm (still in compliance with national standards). Do you agree? Let the user choose

Replace immediately

Delay to 18:00

[0092] In some alternative implementations, the second operating unit includes an output module and a generation module. The output module outputs multiple suggested options, allowing the user to select one of them. Each suggested option includes at least a replacement option and a retention option. The replacement option indicates replacing the air conditioner component whose reliability parameter is less than the preset threshold, and the retention option indicates retaining the air conditioner component whose reliability parameter is less than the preset threshold. The generation module generates the alternative control program if the user selects the retention option. This device generates alternative control programs through user-participatory decision-making via the above steps, balancing equipment maintenance and user experience.

[0093] Specifically, the system outputs multiple suggested options for the user to choose from: When the first neural network model analyzes that the reliability parameters of one or more components are below a preset threshold, the system generates and displays multiple suggested options to the user. These options typically include replacement and retention options, allowing the user to make a decision based on the current situation and preferences. The system suggests replacing the unreliable component to ensure system stability and extend equipment lifespan. Alternative control strategies are provided for situations where immediate component replacement is not necessary, such as reducing the component's load, adjusting the operating mode to reduce wear, or activating a backup component. When the user selects the retention option, an alternative control program is generated: If the user selects the retention option, an alternative control program is generated based on this decision, aiming to reduce the risk of component failure by changing the control strategy, while minimizing the impact on user comfort and system performance.

[0094] In some optional embodiments, the device further includes a fourth analysis unit and a generation unit. The fourth analysis unit is used to analyze the operating data and the equipment status data using a second neural network model to obtain a predicted energy consumption value. The second neural network model is trained using multiple sets of fourth data, each set including historical operating data, historical equipment status data, and a corresponding historical predicted energy consumption value. The generation unit is used to generate an energy consumption warning message when the predicted energy consumption value exceeds a preset energy consumption value, thus alerting the air conditioner to high energy consumption. This device, by predicting energy consumption values, can balance user comfort with energy efficiency.

[0095] Specifically, using the aforementioned historical operating data and historical equipment status data, energy consumption can be predicted through a second neural network model. For example, it can output the predicted energy consumption for the next hour, recommend optimal energy efficiency control parameters (such as chilled water temperature setpoint), and generate energy consumption warning information to alert the air conditioner to high energy consumption. Furthermore, a causal model can be constructed to analyze factors that significantly impact energy consumption, such as the influence intensity of "filter cleaning" on energy consumption (Δ = 0.35).

[0096] The control device of the air conditioner includes a processor and a memory. The first acquisition unit, the first analysis unit, and the execution unit are all stored as program units in the memory. The processor executes the program units stored in the memory to achieve the corresponding functions. All of the above modules are located in the same processor; or, the above modules are located in different processors in any combination.

[0097] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and user experience can be improved by adjusting kernel parameters.

[0098] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0099] This invention provides a computer-readable storage medium including a stored program, wherein the program, when executed, controls the device containing the computer-readable storage medium to perform a control method for an air conditioner.

[0100] Specifically, the control methods for air conditioners include:

[0101] Step S201: Obtain user interaction data, wherein the user interaction data includes at least voice data;

[0102] Step S202: Analyze the user interaction data using a first language model to obtain expected data, wherein the expected data is data representing the user's adjustment expectations and includes at least the target temperature. The first language model is trained using multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data.

[0103] Step S203: Generate an initial control program based on the expected data, simulate the operation of the initial control program, obtain the simulation operation results, and adjust the initial control program until the simulation operation results indicate that the air conditioner is abnormal, so as to obtain a target control program and execute the target control program to control the operation of the air conditioner.

[0104] This invention provides an electronic device, including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs at least the following steps:

[0105] Step S201: Obtain user interaction data, wherein the user interaction data includes at least voice data;

[0106] Step S202: Analyze the user interaction data using a first language model to obtain expected data, wherein the expected data is data representing the user's adjustment expectations and includes at least the target temperature. The first language model is trained using multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data.

[0107] Step S203: Generate an initial control program based on the expected data, simulate the operation of the initial control program, obtain the simulation operation results, and adjust the initial control program until the simulation operation results indicate that the air conditioner is abnormal, so as to obtain a target control program and execute the target control program to control the operation of the air conditioner.

[0108] The devices mentioned in this article can be servers, PCs, tablets, mobile phones, etc.

[0109] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the methods described in various embodiments of this application:

[0110] Step S201: Obtain user interaction data, wherein the user interaction data includes at least voice data;

[0111] Step S202: Analyze the user interaction data using a first language model to obtain expected data, wherein the expected data is data representing the user's adjustment expectations and includes at least the target temperature. The first language model is trained using multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data.

[0112] Step S203: Generate an initial control program based on the expected data, simulate the operation of the initial control program, obtain the simulation operation results, and adjust the initial control program until the simulation operation results indicate that the air conditioner is abnormal, so as to obtain a target control program and execute the target control program to control the operation of the air conditioner.

[0113] It is obvious to those skilled in the art that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. They can be implemented using computer-executable program code, and thus can be stored in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those described herein, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0114] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0115] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0116] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0117] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0118] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0119] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0120] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0121] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0122] As can be seen from the above description, the embodiments of this application achieve the following technical effects:

[0123] 1) In the air conditioner control method of this application, user interaction data is acquired, analyzed using a first language model to obtain expected data, and an initial control program is generated based on the expected data. The initial control program is simulated and run to obtain simulation results. If the simulation results indicate an air conditioner malfunction, the initial control program is adjusted until the simulation results indicate the air conditioner is functioning normally, resulting in a target control program. The target control program is then executed to control the air conditioner's operation. Compared to existing technologies where air conditioners are adjusted directly based on manually set parameters, leading to a poor user experience, the air conditioner control method of this application can control based on user interaction data and simulate the running results to verify whether the initial control program generated from the expected data will cause air conditioner malfunctions. This avoids causing air conditioner malfunctions and a poor user experience during adjustments based on user interaction data, and eliminates the need for manual adjustments by the user, thus improving the user experience. Therefore, it can solve the problem of poor user experience in existing technologies and achieve the effect of improving user experience.

[0124] 2) In the air conditioner control device of this application, user interaction data is acquired, analyzed using a first language model to obtain expected data, and an initial control program is generated based on the expected data. The initial control program is simulated and run to obtain simulation results. If the simulation results indicate an air conditioner malfunction, the initial control program is adjusted until the simulation results indicate the air conditioner is functioning normally, resulting in a target control program. The target control program is then executed to control the air conditioner's operation. Compared to existing technologies where air conditioners are adjusted directly based on manually set parameters, leading to a poor user experience, the air conditioner control device of this application can control based on user interaction data and simultaneously simulate the running results to verify whether the initial control program generated from the expected data will cause air conditioner malfunctions. This avoids causing air conditioner malfunctions and a negative user experience during adjustments based on user interaction data, and eliminates the need for manual user adjustments, thus improving the user experience. Therefore, it solves the problem of poor user experience in existing technologies and achieves the effect of improving user experience.

[0125] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A control method for an air conditioner, characterized in that, include: Acquire user interaction data, wherein the user interaction data includes at least voice data; The user interaction data is analyzed by a first language model to obtain expected data, wherein the expected data is data representing the user's adjustment expectations and includes at least the target temperature. The first language model is trained by multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data. An initial control program is generated based on the expected data. The initial control program is simulated and run to obtain simulation results. If the simulation results indicate that the air conditioner is abnormal, the initial control program is adjusted until the simulation results indicate that the air conditioner is normal, thus obtaining a target control program. The target control program is then executed to control the operation of the air conditioner. The method further includes: acquiring operating data and equipment status data of the air conditioner, wherein the operating data includes at least the historical operating frequency of the air conditioner, and the equipment status data represents the operating status of the components in the air conditioner; analyzing the operating data and the equipment status data through a first neural network model to obtain a reliability parameter, wherein the reliability parameter represents the reliability of the components in the air conditioner, and the first neural network model is trained using multiple sets of third data, each set of third data including: historical operating data, historical equipment status data, and corresponding historical reliability parameters; and generating an alternative control program and executing the alternative control program to improve the reliability of the air conditioner when the reliability parameter is less than a preset threshold.

2. The control method for an air conditioner according to claim 1, characterized in that, Simulate the execution of the initial control program, including: Obtain the structural parameters of the predetermined space and the location parameters of the air conditioner, wherein the predetermined space represents the space where the air conditioner is located, and the structural parameters include at least the area; A virtual operating model is constructed based at least on the structural parameters and the position parameters, and the initial control program is simulated and run on the virtual operating model, wherein the virtual operating model is used to simulate the operation of the air conditioner in the predetermined space.

3. The control method for an air conditioner according to claim 2, characterized in that, At least based on the structural parameters and the positional parameters, a virtual operating model is constructed, including: A physical model of the predetermined space is constructed based on the structural parameters and the positional parameters of the predetermined space; A spatial transformation matrix is ​​determined, and the physical model is converted into a virtual platform using the spatial transformation matrix. The spatial transformation matrix represents the positional transformation relationship between the physical model and the virtual platform, and the virtual platform includes a data transmission interface. The desired data is input to the virtual platform through the data transmission interface to obtain the virtual operating model.

4. The control method for an air conditioner according to claim 1, characterized in that, Generate an initial control program based on the desired data, including: Obtain the current temperature, and if the current temperature is different from the target temperature, calculate the absolute value of the difference between the current temperature and the target temperature; An initial control strategy is generated based at least on the absolute value of the difference, and an initial control program is generated based on the initial control strategy, wherein the initial control strategy includes at least a strategy for adjusting from the current temperature to the target temperature.

5. The control method for an air conditioner according to claim 1, characterized in that, After executing the target control program to control the operation of the air conditioner, the method further includes: Obtain user feedback data, wherein the user feedback data represents the user's feedback on the adjustment results of the air conditioner; The user feedback data is analyzed using a second language model to obtain fine-tuning parameters, which represent parameters for adjusting the operation of the air conditioner. The second language model is trained using multiple sets of second data, each set of which includes historical user feedback data and historical fine-tuning parameters. The operation of the air conditioner is controlled according to the fine-tuning parameters to meet the user's needs.

6. The control method for an air conditioner according to claim 1, characterized in that, Generate alternative control procedures, including: Multiple suggested options are output, allowing the user to select one of them. The suggested options include at least a replacement option and a retention option. The replacement option indicates that the air conditioner component with the reliability parameter less than the preset threshold should be replaced, and the retention option indicates that the air conditioner component with the reliability parameter less than the preset threshold should be retained. If the user selects the reserved option, the alternative control procedure is generated.

7. The control method for an air conditioner according to claim 1, characterized in that, The method further includes: The second neural network model is used to analyze the operating data and the equipment status data to obtain the predicted energy consumption value. The second neural network model is trained by multiple sets of fourth data, and each set of fourth data includes: historical operating data, historical equipment status data and corresponding historical predicted energy consumption value. If the predicted energy consumption value is greater than the preset energy consumption value, an energy consumption warning message is generated to alert the air conditioner to its high energy consumption.

8. A control device for an air conditioner, characterized in that, include: The first acquisition unit is used to acquire user interaction data, wherein the user interaction data includes at least voice data; The first analysis unit is used to analyze the user interaction data through a first language model to obtain expected data, wherein the expected data is data that characterizes the user's adjustment expectations and includes at least the target temperature. The first language model is trained by multiple sets of first data, and each set of first data includes: historical user interaction data and historical expected data. An execution unit is configured to generate an initial control program based on the desired data, simulate the operation of the initial control program, obtain simulation results, and adjust the initial control program if the simulation results indicate that the air conditioner is abnormal, until the simulation results indicate that the air conditioner is normal, thereby obtaining a target control program, and executing the target control program to control the operation of the air conditioner. The device further includes a second acquisition unit, a third analysis unit, and a second operation unit. The second acquisition unit acquires the air conditioner's operating data and equipment status data. The operating data includes at least the air conditioner's historical operating frequency, and the equipment status data represents the operating status of the components in the air conditioner. The third analysis unit analyzes the operating data and equipment status data using a first neural network model to obtain reliability parameters. The reliability parameters represent the reliability of the components in the air conditioner. The first neural network model is trained using multiple sets of third data, each set including historical operating data, historical equipment status data, and corresponding historical reliability parameters. The second operation unit generates an alternative control program and executes it to improve the air conditioner's reliability when the reliability parameters are less than a preset threshold.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the control method of the air conditioner according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including a control method for performing an air conditioner according to any one of claims 1 to 7.

Images