Control method of multi-connected air conditioner, cloud server, multi-connected air conditioner and medium
By using a hybrid time-series optimization model and knowledge distillation technology combining multi-split air conditioning units with cloud servers, an adaptive optimal parameter table is generated, which solves the problems of insufficient operating condition coverage and equipment performance degradation in multi-split air conditioning systems, thereby improving energy efficiency and extending equipment life.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 青岛海尔暖通空调设备有限公司
- Filing Date
- 2026-01-26
- Publication Date
- 2026-05-29
AI Technical Summary
Existing multi-split air conditioning systems face problems such as insufficient coverage of operating conditions, equipment performance degradation, and lack of dynamic adaptability in practical applications, resulting in decreased energy efficiency, increased operating costs, and shortened equipment lifespan.
By communicating with a cloud server through a multi-split air conditioner, an optimal parameter table is generated based on historical operating data using a hybrid time-series optimization model, enabling adaptive parameter adjustment. Combined with knowledge distillation technology and equipment lifespan characteristics, performance degradation is dynamically compensated, and control strategies are optimized.
It enables automatic parameter adjustment of multi-split air conditioners under environmental changes, improving energy efficiency, extending equipment life, reducing operating costs, and overcoming the limitations of traditional control logic.
Smart Images
Figure CN122107526A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of air conditioning control technology, specifically providing a control method for a multi-split air conditioner, a cloud server, a multi-split air conditioner, and a medium. Background Technology
[0002] Existing multi-split central air conditioning systems typically rely on laboratory testing to obtain a fixed set of control parameters or empirical formulas (such as compressor start / stop criteria, throttle valve opening control coefficients, refrigerant flow coefficients, energy efficiency calculation correction coefficients, etc.) at the factory, and these parameters are then embedded in the control logic. However, this control strategy based on fixed parameters faces significant limitations in practical applications, mainly for the following three reasons:
[0003] Insufficient coverage of operating conditions: The laboratory testing environment is highly idealized, and the equipment configuration is simple, making it difficult to simulate complex field application scenarios.
[0004] Equipment performance degradation: With long-term operation, key system components (such as heat exchangers and compressors) will age, resulting in problems such as decreased heat exchange efficiency and changes in refrigerant charge, causing the actual performance of the equipment to deviate from the factory condition and the original control parameters to gradually become ineffective.
[0005] Lack of dynamic adaptability: Fixed parameters cannot be adjusted in real time according to external environment, load demand and seasonal changes, resulting in adverse effects such as decreased energy efficiency, increased operating costs and shortened equipment life.
[0006] Accordingly, there is a need in the field for a new control scheme for multi-split air conditioning systems to solve the above problems. Summary of the Invention
[0007] In order to overcome the above-mentioned defects, this application is made to provide a solution or at least a partial solution to the technical problem that existing multi-split air conditioning systems have parameters that are fixed for a long time and cannot be adaptively adjusted.
[0008] In a first aspect, this application provides a control method for a multi-split air conditioner, wherein the multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the cloud server. The method includes: for each operating cycle of the multi-split air conditioner, based on historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model to obtain an optimal parameter table for the current operating cycle; wherein the optimal parameter table includes multiple operating conditions and corresponding optimal parameter combinations; the hybrid time-series optimization model is obtained by training based on continuously uploaded operating data of the multi-split air conditioner with maximizing the energy efficiency ratio of the multi-split air conditioner as the optimization objective; and distributing the optimal parameter table to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to real-time operating conditions, obtain the optimal parameter combination adapted to the real-time operating conditions, and control operation based on the optimal parameter combination.
[0009] In one technical solution of the control method for the multi-split air conditioner described above, the step of obtaining the optimal parameter table for the current operating cycle based on the historical operating data uploaded by the multi-split air conditioner and using a preset hybrid time-series optimization model includes: obtaining the optimal parameter set for the current operating cycle based on the historical operating data and through the hybrid time-series optimization model; the historical operating data is the operating data for a preset historical operating cycle; and performing knowledge distillation processing on the optimal parameter set to generate the optimal parameter table.
[0010] In one technical solution of the control method for the multi-split air conditioner described above, the step of obtaining the optimal parameter set for the current operating cycle based on the historical operating data and through the hybrid time-series optimization model includes: inputting the historical operating data into the hybrid time-series optimization model, taking the maximization of the energy efficiency ratio of the multi-split air conditioner as the optimization objective, and generating multiple sets of Pareto optimal parameters based on regularization constraints; among the multiple sets of Pareto optimal parameters, selecting the Pareto optimal parameter with the highest energy efficiency ratio and meeting the preset stability requirements for each operating condition to obtain the optimal parameter set for the current operating cycle.
[0011] In one technical solution of the above-mentioned control method for multi-split air conditioners, the step of performing knowledge distillation on the optimal parameter set to generate an optimal parameter table includes: discretizing the optimal parameter set to obtain multiple sets of mapping relationships between operating conditions and parameters; supplementing missing data in the multiple sets of mapping relationships between operating conditions and parameters using an interpolation fitting algorithm, and merging similar operating conditions and parameters in the multiple sets of mapping relationships between operating conditions and parameters using a clustering optimization algorithm; performing device-side executability verification on the multiple sets of mapping relationships between operating conditions and parameters; and generating an optimal parameter table based on the executability-verified mapping relationships between the multiple sets of operating conditions and parameters.
[0012] In one technical solution of the above-mentioned control method for multi-split air conditioners, the method further includes: retraining the hybrid time-series optimization model based on the latest operating data uploaded by the multi-split air conditioner according to a preset training cycle; performing a step of obtaining an optimal parameter table based on the trained hybrid time-series optimization model to obtain a new optimal parameter table; predicting the energy efficiency ratio improvement value after the multi-split air conditioner implements the new optimal parameter table; if the energy efficiency ratio improvement value is greater than a preset improvement threshold, sending the new optimal parameter table to the multi-split air conditioner; obtaining a parameter table execution report fed back by the multi-split air conditioner after implementing the new optimal parameter table, the parameter table execution report including whether the new optimal parameter table is adopted and the actual energy efficiency ratio improvement value; and using the parameter table execution report as training data for the hybrid time-series optimization model in the next preset training cycle.
[0013] In one technical solution of the control method for the multi-split air conditioner described above, before obtaining the optimal parameter table for the current operating cycle using a preset hybrid time-series optimization model based on the historical operating data uploaded by the multi-split air conditioner, the method further includes: preprocessing the historical operating data, wherein the preprocessing includes at least one of the following: data cleaning and standardization processing, including noise reduction processing, data normalization processing, and abnormal data removal processing; constructing time-series features for time-series operating data in the operating data using the sliding window method; and / or obtaining equipment operating life features, and using the equipment operating life features and the preprocessed historical operating data as input data for the hybrid time-series optimization model; wherein the equipment operating life features include at least the cumulative operating hours of the multi-split air conditioner and the number of compressor start-stop cycles of the multi-split air conditioner.
[0014] In a second aspect, this application provides a control method for a multi-split air conditioner, wherein the multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the multi-split air conditioner. The method includes: obtaining an optimal parameter table issued by the cloud server; the optimal parameter table is obtained by the cloud server based on historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model, with the optimization objective of maximizing the energy efficiency ratio of the multi-split air conditioner; obtaining current operating condition information in real time; wherein the current operating condition information includes at least outdoor temperature, indoor unit load rate, and equipment operating time; obtaining an optimal parameter combination matching the current operating condition information from the optimal parameter table based on the current operating condition information; and controlling the operation of the multi-split air conditioner based on the optimal parameter combination.
[0015] In a third aspect, this application provides a cloud server, including a processor and a memory, the memory being adapted to store multiple program codes, the program codes being adapted to be loaded and run by the processor to execute the control method for multi-split air conditioners described in any of the above-described technical solutions.
[0016] In a fourth aspect, this application provides a multi-split air conditioner, including a processor and a memory, the memory being adapted to store a plurality of program codes, the program codes being adapted to be loaded and run by the processor to execute the multi-split air conditioner control method described in any of the above-described technical solutions.
[0017] In a fifth aspect, this application provides a computer-readable storage medium storing a plurality of program codes adapted to be loaded and run by a processor to perform the control method for a multi-split air conditioner described in any of the above-described technical solutions.
[0018] The above-described technical solutions of this application have at least one or more of the following beneficial effects:
[0019] This application discloses a control method for a multi-split air conditioner, wherein the multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the cloud server. The method includes: for each operating cycle of the multi-split air conditioner, based on historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model to obtain an optimal parameter table for the current operating cycle; wherein the optimal parameter table includes multiple operating conditions and corresponding optimal parameter combinations; the hybrid time-series optimization model is obtained by training based on continuously uploaded operating data of the multi-split air conditioner with maximizing the energy efficiency ratio of the multi-split air conditioner as the optimization objective; and the optimal parameter table is distributed to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to the real-time operating conditions to obtain the optimal parameter combination adapted to the real-time operating conditions, and operate based on the optimal parameter combination.
[0020] This application generates the optimal parameter table for the current operating cycle of the multi-split air conditioner based on the historical operating data uploaded by the multi-split air conditioner through a hybrid time-series optimization model preset on the cloud server. It can obtain the optimal parameter table that automatically evolves with environmental changes through the actual operating data of the multi-split air conditioner, and completely solves the limitation of traditional control logic parameters that only rely on short-cycle and ideal experimental environments.
[0021] Furthermore, this application uses knowledge distillation technology to transform the optimal parameter set output by the hybrid time series optimization model into an optimal parameter table that can be directly looked up and executed by the multi-split air conditioner. This eliminates the need for the multi-split air conditioner to undertake complex model calculation tasks, achieving adaptive parameter adjustment without additional hardware investment, thus balancing technical practicality and economy.
[0022] Furthermore, this application retrains the hybrid time-series optimization model according to a preset training cycle, and combines it with a closed-loop mechanism for energy efficiency verification feedback. This allows for continuous optimization of the model based on the actual operating effect feedback from the multi-split air conditioner, thereby continuously improving the accuracy of the generated optimal parameter table over time.
[0023] Furthermore, by introducing equipment operating life characteristics as model input data, this application enables the hybrid time-series optimization model to accurately identify the performance degradation trend of multi-split air conditioners and dynamically compensate for the impact of performance degradation, thereby avoiding equipment overload operation caused by parameter mismatch and significantly extending the service life of multi-split air conditioners. Attached Figure Description
[0024] The preferred embodiments of this application are described below with reference to the accompanying drawings, in which:
[0025] Figure 1This is a schematic flowchart of the main steps of a control method for a multi-split air conditioner according to an embodiment of this application;
[0026] Figure 2 This is a detailed flowchart illustrating the steps of a control method for a multi-split air conditioner according to an embodiment of this application;
[0027] Figure 3 This is a schematic flowchart of the main steps of a control method for a multi-split air conditioner according to another embodiment of this application;
[0028] Figure 4 This is a schematic diagram of the main structure of a cloud server according to an embodiment of this application;
[0029] Figure 5 This is a schematic diagram of the main structure of a multi-split air conditioner according to an embodiment of this application;
[0030] Figure 6 This is a schematic diagram illustrating the collaborative control process between a multi-split air conditioner and a cloud server according to an embodiment of this application.
[0031] List of reference numerals in the attached diagram:
[0032] 41, 51: Memory; 42, 52: Processor. Detailed Implementation
[0033] Some embodiments of this application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of this application and are not intended to limit the scope of protection of this application.
[0034] In the description of this application, "module" and "processor" can include hardware, software, or a combination of both. A module can include hardware circuitry, various suitable sensors, communication ports, memory, and can also include software components, such as program code, or a combination of software and hardware. A processor can be a central processing unit, microprocessor, image processor, digital signal processor, or any other suitable processor. The processor has data and / or signal processing capabilities. The processor can be implemented in software, in hardware, or a combination of both. Non-transitory computer-readable storage media includes any suitable medium capable of storing program code, such as magnetic disks, hard disks, optical disks, flash memory, read-only memory, random access memory, etc. The term "A and / or B" means all possible combinations of A and B, such as only A, only B, or A and B. The terms "at least one A or B" or "at least one of A and B" have a similar meaning to "A and / or B" and can include only A, only B, or A and B. The singular terms "a" or "this" can also include plural forms.
[0035] Existing central air conditioning multi-split systems rely on fixed control parameters or empirical formulas obtained from laboratory tests and are embedded in the control logic at the factory. However, this strategy has significant limitations. For example, it has insufficient coverage of operating conditions, and the ideal laboratory environment is difficult to simulate complex field scenarios; equipment performance deteriorates, and the aging of core components over long-term operation causes the original parameters to become invalid; it lacks dynamic adaptability, and the fixed parameters cannot match environmental, load, and seasonal changes, resulting in decreased energy efficiency, increased costs, and shortened equipment life.
[0036] Therefore, this application provides a control method for a multi-split air conditioner, wherein the multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the cloud server. The method includes: for each operating cycle of the multi-split air conditioner, based on historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model to obtain an optimal parameter table for the current operating cycle; wherein the optimal parameter table includes multiple operating conditions and corresponding optimal parameter combinations; the hybrid time-series optimization model is obtained by training based on the continuously uploaded operating data of the multi-split air conditioner with maximizing the energy efficiency ratio of the multi-split air conditioner as the optimization objective; and the optimal parameter table is sent to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to the real-time operating conditions to obtain the optimal parameter combination adapted to the real-time operating conditions, and operate based on the optimal parameter combination.
[0037] This application generates the optimal parameter table for the current operating cycle of the multi-split air conditioner based on the historical operating data uploaded by the multi-split air conditioner through a hybrid time-series optimization model preset on the cloud server. It can obtain the optimal parameter table that automatically evolves with environmental changes through the actual operating data of the multi-split air conditioner, and completely solves the limitation of traditional control logic parameters that only rely on short-cycle and ideal experimental environments.
[0038] See appendix Figure 1 , Figure 1 This is a schematic flowchart illustrating the main steps of a control method for a multi-split air conditioner according to an embodiment of this application. Figure 1 As shown in the embodiment of this application, the control method of a multi-split air conditioner is connected to a cloud server. The method is applied to the cloud server and mainly includes the following steps S101-S102.
[0039] Step S101: For each operating cycle of the multi-split air conditioner, based on the historical operating data uploaded by the multi-split air conditioner, a preset hybrid time-series optimization model is used to obtain the optimal parameter table for the current operating cycle; wherein, the optimal parameter table includes multiple operating conditions and corresponding optimal parameter combinations; the hybrid time-series optimization model is obtained by training based on the operating data continuously uploaded by the multi-split air conditioner with the optimization objective of maximizing the energy efficiency ratio of the multi-split air conditioner.
[0040] In this embodiment, the multi-split air conditioner continuously uploads historical operating data. Based on the historical operating data, the cloud server uses a preset hybrid time-series optimization model to obtain the optimal parameter table for the current operating cycle. The optimal parameter table includes multiple operating conditions and the optimal parameter combination corresponding to each operating condition. The current operating cycle includes the operating cycle for a future preset time period, which can be 30 days.
[0041] Step S102: Send the optimal parameter table to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to the real-time operating conditions, obtain the optimal parameter combination that adapts to the real-time operating conditions, and control the operation based on the optimal parameter combination.
[0042] In this embodiment, after the optimal parameter table for the current operating cycle is sent to the multi-split air conditioner, the multi-split air conditioner looks up the table based on the real-time operating conditions to obtain the optimal parameter combination that is suitable for the current real-time operating conditions, and controls the operation of the multi-split air conditioner based on the optimal parameter combination.
[0043] Based on the above steps S101-S102, this application generates the optimal parameter table for the current operating cycle of the multi-split air conditioner based on the historical operating data uploaded by the multi-split air conditioner through the hybrid time-series optimization model preset by the cloud server. It can obtain the optimal parameter table that automatically evolves with environmental changes through the actual operating data of the multi-split air conditioner, and completely solves the limitation of traditional control logic parameters that only rely on short-cycle and ideal experimental environments.
[0044] The following sections will further explain steps S101 and S102.
[0045] See appendix Figure 2 , Figure 2 This is a detailed flowchart illustrating the steps of a control method for a multi-split air conditioner according to an embodiment of this application.
[0046] In one embodiment, before obtaining the optimal parameter table for the current operating cycle using a preset hybrid time-series optimization model based on the historical operating data uploaded by the multi-split air conditioner, the method further includes: preprocessing the historical operating data, the preprocessing including at least one of the following: data cleaning and standardization processing, including noise reduction processing, data normalization processing, and outlier data removal processing; constructing time-series features for time-series operating data in the operating data using the sliding window method; and / or obtaining equipment operating life features, and using the equipment operating life features and the preprocessed historical operating data as input data for the hybrid time-series optimization model; wherein, the equipment operating life features include at least the cumulative operating hours of the multi-split air conditioner and the number of compressor start-stop cycles of the multi-split air conditioner.
[0047] Specifically, historical operating data refers to the operating data collected by edge devices and uploaded to the cloud server within a preset historical time period. The edge devices can be the main unit of the multi-split air conditioner or a sub-device built into the multi-split air conditioner. The preset historical time period can be the past 30 days. Operating data includes environmental conditions (such as outdoor temperature, humidity, seasonal information, and geographical location), system status (such as compressor frequency, current, exhaust / suction temperature, and refrigerant pressure), terminal loads (such as the load rate of each indoor unit and supply air temperature), control parameters (such as the current throttle valve opening, target superheat setting, and return gas pressure), and performance indicators (such as the Coefficient of Performance (COP)). In one embodiment, historical operating data is shown in Table 1.
[0048] Table 1
[0049] category Example data illustrate Environmental conditions Outdoor temperature, humidity, seasonal information, and geographical location Basic characteristics of environmental load System status Compressor frequency, current, discharge / suction temperature, refrigerant pressure Health status of reaction equipment End load Load rate of each indoor unit and supply air temperature Load dynamics Control parameters Current throttle valve opening, target superheat setting, return gas pressure Control layer input Performance indicators COP Optimization Objective
[0050] Before obtaining the optimal parameter table for the current running cycle, these historical running data are preprocessed, specifically including:
[0051] Denoising: Algorithms such as moving average filtering, median filtering, or wavelet denoising are used to denoise historical running data;
[0052] Data normalization processing: For data of different dimensions and magnitudes, a preset normalization algorithm is used to perform data normalization processing;
[0053] Outlier data removal: Data points that are outside the reasonable range, logically contradictory, or isolated are removed to prevent outlier data from misleading the model and causing parameter calculation errors.
[0054] Based on the temporal characteristics of multi-split air conditioning operation (such as the current operating condition being affected by load and temperature changes over a period of time), for all time-series operating data (such as environmental conditions, system status, terminal load, performance indicators, etc., which change over time), time-series features are extracted through a sliding window, transforming single-point data into interval trend features, forming a combined feature set of original features plus time-series features.
[0055] To address the issue of equipment performance degradation over time, this application acquires real-time statistics on the operational lifespan characteristics of multi-split air conditioners and periodically uploads them to a cloud server. These characteristics, along with pre-processed historical operational data, serve as input data for a hybrid time-series optimization model. This data supplements the model's reflection of equipment aging, allowing the model to learn the correlation between lifespan status and optimal parameters. Consequently, the model can adjust the parameters in the generated optimal parameter table based on the equipment's lifespan. By introducing equipment operational lifespan characteristics as model input data, this application enables the hybrid time-series optimization model to accurately identify the performance degradation trend of multi-split air conditioners and dynamically compensate for the impact of performance degradation. This avoids equipment overload operation due to parameter mismatch and significantly extends the service life of multi-split air conditioners.
[0056] Regarding step S101, in one embodiment, obtaining the optimal parameter table for the current operating cycle based on the historical operating data uploaded by the multi-split air conditioner using a preset hybrid time-series optimization model includes: obtaining the optimal parameter set for the current operating cycle based on the historical operating data and through the hybrid time-series optimization model; the historical operating data is the operating data of a preset historical operating cycle; and performing knowledge distillation processing on the optimal parameter set to generate the optimal parameter table.
[0057] Specifically, the characteristic data corresponding to the preprocessed historical operating data and the equipment operating life characteristics are input into the hybrid time series optimization model, and the hybrid time series optimization model outputs the optimal parameter set for the current operating cycle of the multi-split air conditioner.
[0058] The optimal parameter set is a high-dimensional result set, including various actual operating conditions, and each condition corresponds to a parameter combination containing 5 to 20 adjustable parameters. To adapt to the low computing power of multi-split air conditioners and meet the need for rapid table lookup and retrieval of optimal parameters on the device side, knowledge distillation technology is used to reduce the dimensionality of the high-dimensional optimal parameter set, transforming it into a two-dimensional or three-dimensional optimal parameter table that the multi-split air conditioner can directly execute.
[0059] For example, the two-dimensional optimal parameter table can be constructed in the form of "outdoor temperature × indoor unit load rate → optimal parameter combination". The intersection of any dimension interval corresponds to a unique optimal parameter combination. Multi-split air conditioners only need to match the real-time operating condition "outdoor temperature × indoor unit load rate" to complete the table lookup and parameter retrieval, without having to undertake the complex high-dimensional function calculation task.
[0060] In one implementation, obtaining the optimal parameter set for the current operating cycle based on the historical operating data and through the hybrid time-series optimization model includes: inputting the historical operating data into the hybrid time-series optimization model, taking the maximization of the energy efficiency ratio of the multi-split air conditioner as the optimization objective, generating multiple sets of Pareto optimal parameters based on regularization constraints; and selecting the Pareto optimal parameters with the highest energy efficiency ratio and meeting preset stability requirements for each operating condition from among the multiple sets of Pareto optimal parameters to obtain the optimal parameter set for the current operating cycle.
[0061] Specifically, the hybrid temporal optimization model is a hybrid model that combines the Transformer model with a Long Short-Term Memory (LSTM) network and an attention mechanism, and its output is a Pareto optimal solution set.
[0062] The preprocessed historical operating data and the equipment operating life characteristics reflecting the equipment performance degradation state are used as the input feature set of the hybrid time series optimization model and input into the hybrid time series optimization model.
[0063] The input to the model can be represented as:
[0064]
[0065] in, Outdoor temperature; Outdoor humidity; For compressor frequency; This is the condensation pressure; For evaporation pressure, For indoor unit load, For the current time (t), energy efficiency, Indicates other variables.
[0066] The hybrid time-series optimization model aims to maximize the coefficient of performance (COP) of multi-split air conditioners. The objective function expression is as follows:
[0067]
[0068] in, E[·] is the loss function of the model; E[·] is the expectation operator, and the negative sign is to transform the goal of maximizing energy efficiency ratio into the paradigm of minimizing loss in machine learning; The energy efficiency ratio at a future time (time t+1); These are the optimal parameters output by the model at time t+1; the model indirectly achieves the future energy efficiency ratio (COP) by minimizing the loss function L. t+1 Maximizing the expected value.
[0069] The hybrid time series optimization model introduces a regularization constraint term to ensure parameter smoothness. The mathematical formula for the regularization constraint term is as follows:
[0070]
[0071] in, E[·] is the total loss function of the model; E[·] is the expectation operator, and the negative sign is to transform the goal of maximizing the energy efficiency ratio into the paradigm of minimizing the loss in machine learning; Energy efficiency ratio; The regularization coefficient is used. These are the optimal parameters output by the model at time t+1; These are the optimal parameters output by the model at time t; This is a parameter smoothness constraint term.
[0072] Regular constraint terms By penalizing drastic changes in parameters between adjacent time steps, the smoothness of parameter updates is ensured. If the parameter differences between adjacent time steps are large, the value of the regularization constraint term increases significantly, leading to an increase in the total loss. To minimize the total loss, the model automatically allows... as close as possible This avoids sudden parameter changes; ultimately, it achieves smooth adjustment of hardware parameters in the output results, prevents system operation fluctuations, ensures stable operation of multi-split air conditioners, and extends equipment life.
[0073] Through model computation, multiple sets of Pareto optimal parameters are generated, and the model output can be expressed as follows:
[0074]
[0075] in, This is the correction factor for the throttle valve; This is the correction factor for the condensation side; This refers to the control coefficient of the electronic expansion valve. Indicates other parameters.
[0076] The operating conditions of the multi-split air conditioner are divided based on the model's input features (such as outdoor temperature and indoor unit load rate), resulting in multiple operating conditions. Based on the Pareto optimal parameter corresponding to each operating condition, the parameter with the highest COP value is selected as the candidate optimal parameter for that operating condition. A preset stability requirement check is performed on the candidate optimal parameter to ensure that it meets the preset stability requirements. Specific preset stability requirements may include:
[0077] Comfort requirements, such as setting a reasonable temperature range, should be considered to avoid a decrease in comfort due to excessive pursuit of energy efficiency.
[0078] Equipment operation stability requirements, such as setting the maximum value of compressor frequency fluctuation rate, the maximum value of throttle valve opening adjustment step size, and the maximum value of refrigerant pressure change rate;
[0079] Equipment operation safety requirements, such as setting the maximum and minimum values of compressor discharge pressure and suction pressure, and setting compressor discharge temperature, to prevent parameters from exceeding safe ranges, which could lead to equipment wear or failure.
[0080] Based on the candidate optimal parameters that meet the preset stability requirements under all operating conditions, the optimal parameter set for the current operating cycle is obtained.
[0081] In one embodiment, the step of performing knowledge distillation on the optimal parameter set to generate an optimal parameter table includes: discretizing the optimal parameter set to obtain multiple sets of mapping relationships between operating conditions and parameters; supplementing missing data in the multiple sets of mapping relationships between operating conditions and parameters using an interpolation fitting algorithm, and merging similar operating conditions and parameters in the multiple sets of mapping relationships between operating conditions and parameters using a clustering optimization algorithm; performing equipment-side executability verification on the multiple sets of mapping relationships between operating conditions and parameters; and generating an optimal parameter table based on the executability-verified mapping relationships between the multiple sets of mapping relationships between operating conditions and parameters.
[0082] Specifically, the optimal parameter set is discretized by segmenting and sampling the continuous operating condition range and the continuous parameter range within the optimal parameter set. This transforms the infinite combinations into a finite, enumerable set of operating conditions and parameters, obtaining multiple sets of mapping relationships between operating conditions and parameters. In one embodiment, the discretization process includes:
[0083] The operating conditions are divided into intervals and sampled. For example, the outdoor temperature is discretized into several fixed intervals, and the sampling points include: -10℃, -5℃, 0℃, 5℃, ..., 30℃, 35℃ (each interval is 5℃ apart, for a total of 10 discrete points); similarly, the indoor unit load is discretized into: 10%, 20%, ..., 100% (for a total of 10 discrete points); the original continuous combination (infinitely many) of outdoor temperature and indoor unit load becomes 10×10=100 discrete combinations of operating conditions.
[0084] The parameters are discretized. For each discrete combination of operating conditions, the most representative parameter value is selected from the optimal parameter set (for example, under a certain operating condition, the continuous optimal range of the throttle valve correction coefficient is 1.2~1.3, and after discretization, 1.25 is selected as the fixed parameter for that operating condition). After discretization, a correspondence between discrete operating conditions and discrete parameters is formed, laying the foundation for subsequent conversion into a two-dimensional or three-dimensional optimal parameter table.
[0085] By using interpolation fitting algorithms, missing data in multiple sets of working condition and parameter mapping relationships are supplemented to generate reasonable intermediate parameter values, ensuring the continuity of working condition coverage. At the same time, clustering algorithms (such as K-Means algorithm and DBSCAN algorithm) are combined to group and merge mapping relationships with similar working conditions and similar parameters, further compressing the data scale and reducing the storage and query pressure of multi-split air conditioners.
[0086] Then, the device-side executability verification is performed on the mapping relationship of multiple sets of operating conditions and parameters after interpolation fitting and cluster optimization. The verification content may include uniqueness verification of mapping relationship, validity verification of parameter value range, and smoothness verification of parameter changes in adjacent operating conditions, to ensure that the multi-split air conditioner is unambiguous when looking up the table, that parameter execution complies with the hardware limits of the equipment, and that parameter adjustment does not cause system operation fluctuations.
[0087] Finally, based on the multiple sets of operating conditions and parameter mapping relationships that have passed the executability verification, an optimal parameter table with a two-dimensional or three-dimensional structure is constructed. For example, a two-dimensional parameter table can construct a mapping structure of "outdoor temperature × indoor unit load → optimal parameter combination", and a three-dimensional parameter table can construct a mapping structure of "outdoor temperature × indoor unit load × ambient humidity → optimal parameter combination". This application uses knowledge distillation technology to transform the optimal parameter set output by the hybrid time-series optimization model into an optimal parameter table that the multi-split air conditioner can directly look up and execute. This eliminates the need for the multi-split air conditioner to undertake complex model calculation tasks, achieving adaptive parameter adjustment without additional hardware investment, thus balancing technical practicality and economy.
[0088] Regarding step S102, in one embodiment, the optimal parameter table is sent to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to the real-time operating conditions, obtain the optimal parameter combination that adapts to the real-time operating conditions, and control the operation based on the optimal parameter combination.
[0089] Specifically, the optimal parameter table is sent to the multi-split air conditioner in JSON format. When the multi-split air conditioner is running, it only needs to match the similar operating conditions in the optimal parameter table with the real-time operating conditions to quickly obtain the optimal parameter combination that is adapted to the real-time operating conditions.
[0090] In one implementation, the hybrid time-series optimization model can be retrained according to a preset training cycle based on the latest operating data uploaded by the multi-split air conditioner; based on the trained hybrid time-series optimization model, the step of obtaining the optimal parameter table is executed to obtain a new optimal parameter table; the energy efficiency ratio improvement value after the multi-split air conditioner implements the new optimal parameter table is predicted; if the energy efficiency ratio improvement value is greater than a preset improvement threshold, the new optimal parameter table is sent to the multi-split air conditioner; a parameter table execution report fed back by the multi-split air conditioner after implementing the new optimal parameter table is obtained, the parameter table execution report including whether the new optimal parameter table is adopted and the actual energy efficiency ratio improvement value; the parameter table execution report is used as the training data of the hybrid time-series optimization model in the next preset training cycle.
[0091] Specifically, the hybrid time-series optimization model is periodically trained based on the latest operating data continuously uploaded by the multi-split air conditioner, according to a preset training cycle. The preset training cycle is one quarter, and the hybrid time-series optimization model can be retrained once every quarter.
[0092] Based on the retrained hybrid time-series optimization model, the step of obtaining the optimal parameter table is performed according to the historical operating data at the current moment, thereby obtaining a new optimal parameter table. Subsequently, the energy efficiency ratio improvement value when the multi-split air conditioner implements the new optimal parameter table is predicted. If the energy efficiency ratio improvement value is greater than the preset improvement threshold, the new optimal parameter table is sent to the multi-split air conditioner.
[0093] In a specific example, the preset boost threshold can be set to 2%.
[0094] After receiving a new optimal parameter table, the multi-split air conditioner does not directly overwrite the old table with the new one. Instead, it enters a parallel verification phase. During this phase, both the "previous optimal parameter table" and the "new optimal parameter table" are stored simultaneously, and trial operation is conducted based on the new optimal parameter table. During the trial operation, operating data is monitored in real time, and the actual energy efficiency ratio (EER) improvement is calculated. If the actual EER improvement is less than or equal to a preset improvement threshold, the "new optimal parameter table" is discarded, and the "previous optimal parameter table" continues to be used. Simultaneously, a parameter table execution report is sent to the cloud server, which may include the actual EER improvement and the result of whether the new optimal parameter table was adopted.
[0095] When the actual energy efficiency ratio improvement of the "new optimal parameter table" is less than or equal to the preset improvement threshold, the new table cannot achieve significant energy efficiency optimization. In this case, the new table should be abandoned to avoid operational fluctuations or cost waste caused by ineffective adjustments.
[0096] After receiving the parameter table execution report uploaded by the multi-split air conditioner, the cloud server will incorporate the parameter table execution report into the training data for the next preset training cycle. This application, by retraining the hybrid time-series optimization model according to the preset training cycle and combining it with the closed-loop mechanism of energy efficiency verification feedback, can continuously optimize the model based on the actual operating effect feedback from the multi-split air conditioner, so that the accuracy of the generated optimal parameter table continuously improves with the running time.
[0097] See appendix Figure 3 , Figure 3 This is a schematic flowchart illustrating the main steps of a control method for a multi-split air conditioner according to an embodiment of this application. Figure 3 As shown in the embodiment of this application, the control method for a multi-split air conditioner is described. The multi-split air conditioner is communicatively connected to a cloud server. The method is applied to the multi-split air conditioner and includes steps S301 to S304:
[0098] Step S301: Obtain the optimal parameter table issued by the cloud server; the optimal parameter table is obtained by the cloud server based on the historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model, with the goal of maximizing the energy efficiency ratio of the multi-split air conditioner.
[0099] Step S302: Obtain current operating condition information in real time; wherein, the current operating condition information includes at least outdoor temperature, indoor unit load rate, and equipment running time.
[0100] Step S303: Based on the current operating condition information, obtain the optimal parameter combination that matches the current operating condition information from the optimal parameter table.
[0101] Step S304: Control the operation of the multi-split air conditioner based on the optimal parameter combination.
[0102] Specifically, the multi-split air conditioner establishes a communication connection with the cloud server through a preset communication protocol, receives the optimal parameter table sent by the cloud server, and stores it in the local memory.
[0103] Multi-split air conditioners acquire real-time operating condition information, which includes at least the outdoor temperature collected by the outdoor temperature sensor, the indoor unit load rate calculated based on the difference between the indoor unit's return air temperature and the set temperature and the fan speed, and the equipment running time counted by the built-in timer.
[0104] The collected current operating condition information is matched with the operating condition data in the optimal parameter table to obtain the corresponding optimal parameter combination. The optimal parameter combination includes at least the throttle valve correction coefficient, the condenser side correction coefficient, and the electronic expansion valve control coefficient. The optimal parameter combination obtained from the table is injected into the local control algorithm and converted into an executable hardware control signal to control the multi-split air conditioner. The specific control logic includes:
[0105] Adjust the opening of the outdoor unit's throttle valve according to the throttle valve correction coefficient to precisely control the refrigerant circulation flow rate;
[0106] Adjust the condenser fan speed according to the condenser side correction coefficient to optimize the refrigerant condensation effect;
[0107] The command is sent to the electronic expansion valve of the indoor unit according to the control coefficient of the electronic expansion valve to realize the rational distribution of refrigerant among different indoor units.
[0108] After controlling the multi-split air conditioner based on the optimal parameter combination, the system continuously monitors real-time operating data (such as compressor frequency, refrigerant pressure, and COP value) and uploads the real-time operating data to the cloud server for the next round of model training and optimal parameter table updates.
[0109] It should be noted that although the steps in the above embodiments are described in a specific order, those skilled in the art will understand that in order to achieve the effect of this application, different steps do not necessarily have to be executed in such an order. They can be executed simultaneously (in parallel) or in other orders, and these variations are all within the scope of protection of this application.
[0110] Those skilled in the art will understand that all or part of the processes in the method of the above-described embodiment can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium can include any entity or device capable of carrying the computer program code, a medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.
[0111] Furthermore, this application also provides a cloud server. In one embodiment of the cloud server according to this application, the cloud server includes a processor and a memory. The memory can be configured to store a program for executing the control method of a multi-split air conditioner according to the above-described method embodiments. The processor can be configured to execute the program in the memory, which includes, but is not limited to, a program for executing the control method of a multi-split air conditioner according to the above-described method embodiments. For ease of explanation, only the parts related to the embodiments of this application are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this application. The cloud server can be a cloud server device comprising various electronic devices. See Appendix Figure 4 , Figure 4The example shows a memory 41 and a processor 42 connected via a bus communication connection.
[0112] Furthermore, this application also provides a multi-split air conditioner. In one embodiment of the multi-split air conditioner according to this application, the multi-split air conditioner includes a processor and a memory. The memory can be configured to store a program for executing the control method of the multi-split air conditioner described in the above-described method embodiments. The processor can be configured to execute the program in the memory, which includes, but is not limited to, a program for executing the control method of the multi-split air conditioner described in the above-described method embodiments. For ease of explanation, only the parts related to the embodiments of this application are shown. For specific technical details not disclosed, please refer to the method section of the embodiments of this application. The multi-split air conditioner can be a multi-split air conditioning device comprising various electronic devices. See Appendix Figure 5 , Figure 5 The example shows a memory 51 and a processor 52 connected via a bus communication connection.
[0113] See appendix Figure 6 , Figure 6 This is a schematic diagram illustrating the collaborative control process between a multi-split air conditioner and a cloud server according to one embodiment of this application. Figure 6 As shown, the cloud server receives historical operating data uploaded by the multi-split air conditioner, performs calculations using a preset hybrid time-series optimization model, generates an optimal parameter set, and performs knowledge distillation on the output optimal parameter set to generate a two-dimensional / three-dimensional optimal parameter table that the multi-split air conditioner can directly look up and call, and then sends it to the multi-split air conditioner.
[0114] The multi-split air conditioner receives the optimal parameter table, obtains the optimal parameter combination matching the current operating conditions from the optimal parameter table, and controls the operation of the multi-split air conditioner based on the optimal parameter combination. Then, it continuously uploads the operating data to the cloud server for the next round of model training and optimal parameter table updates on the cloud server.
[0115] Furthermore, this application also provides a computer-readable storage medium. In one embodiment of the computer-readable storage medium according to this application, the computer-readable storage medium can be configured to store a program for executing the control method of a multi-split air conditioner in the above-described method embodiments. This program can be loaded and run by a processor to implement the control method of the multi-split air conditioner. For ease of explanation, only the parts related to the embodiments of this application are shown; for specific technical details not disclosed, please refer to the method section of the embodiments of this application. The computer-readable storage medium can be a memory device comprising various electronic devices. Optionally, in the embodiments of this application, the computer-readable storage medium is a non-transitory computer-readable storage medium.
[0116] Furthermore, it should be understood that since the various modules are only provided to illustrate the functional units of the device described in this application, the physical devices corresponding to these modules may be the processor itself, or a part of the processor's software, hardware, or a combination of both. Therefore, the number of modules shown in the figures is merely illustrative.
[0117] Those skilled in the art will understand that the various modules in the device can be adaptively split or combined. Such splitting or combining of specific modules will not cause the technical solution to deviate from the principles of this application; therefore, the technical solutions after splitting or combining will fall within the protection scope of this application.
[0118] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. Without departing from the principles of this application, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of this application.
Claims
1. A control method for a multi-split air conditioner, characterized in that, The multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the cloud server, the method comprising: For each of the multi-split air conditioners' operating cycles, based on the historical operating data uploaded by the multi-split air conditioner, a preset hybrid time-series optimization model is used to obtain the optimal parameter table for the current operating cycle; wherein, the optimal parameter table includes multiple operating conditions and corresponding optimal parameter combinations; the hybrid time-series optimization model is obtained by training based on the continuously uploaded operating data of the multi-split air conditioner with the optimization objective of maximizing the energy efficiency ratio of the multi-split air conditioner; The optimal parameter table is sent to the multi-split air conditioner so that the multi-split air conditioner can query the optimal parameter table according to the real-time operating conditions, obtain the optimal parameter combination that is suitable for the real-time operating conditions, and control the operation based on the optimal parameter combination.
2. The control method for a multi-split air conditioner according to claim 1, characterized in that, The step of obtaining the optimal parameter table for the current operating cycle based on the historical operating data uploaded by the multi-split air conditioner and using a preset hybrid time-series optimization model includes: Based on the historical operating data, the optimal parameter set for the current operating cycle is obtained through the hybrid time series optimization model; the historical operating data is the operating data for a preset historical operating cycle. The optimal parameter set is subjected to knowledge distillation to generate an optimal parameter table.
3. The control method for a multi-split air conditioner according to claim 2, characterized in that, The process of obtaining the optimal parameter set for the current operating cycle based on the historical operating data and through the hybrid time-series optimization model includes: The historical operating data is input into the hybrid time series optimization model, with the maximization of the energy efficiency ratio of the multi-split air conditioner as the optimization objective, and multiple sets of Pareto optimal parameters are generated based on regularization constraints. Among the multiple sets of Pareto optimal parameters, the Pareto optimal parameters with the highest energy efficiency ratio and meeting the preset stability requirements are selected for each operating condition to obtain the optimal parameter set for the current operating cycle.
4. The control method for a multi-split air conditioner according to claim 2, characterized in that, The step of performing knowledge distillation on the optimal parameter set to generate an optimal parameter table includes: The optimal parameter set is discretized to obtain multiple sets of mapping relationships between working conditions and parameters; The missing data in the mapping relationship between the multiple sets of working conditions and parameters is supplemented by interpolation fitting algorithm, and similar working conditions and parameters in the mapping relationship between the multiple sets of working conditions and parameters are merged by clustering optimization algorithm. The mapping relationship between the multiple sets of operating conditions and parameters is verified for device-side executability. Based on the mapping relationship between the multiple sets of working conditions and parameters after executability verification, an optimal parameter table is generated.
5. The control method for a multi-split air conditioner according to claim 1, characterized in that, The method further includes: According to the preset training cycle, the hybrid time-series optimization model is retrained based on the latest operating data uploaded by the multi-split air conditioner; Based on the trained hybrid time series optimization model, the step of obtaining the optimal parameter table is performed to obtain a new optimal parameter table; Predict the energy efficiency ratio improvement value of the multi-split air conditioner after implementing the new optimal parameter table; If the energy efficiency ratio improvement value is greater than the preset improvement threshold, the new optimal parameter table is sent to the multi-split air conditioner. Obtain the parameter table execution report fed back by the multi-split air conditioner after executing the new optimal parameter table. The parameter table execution report includes whether the new optimal parameter table was adopted and the actual energy efficiency ratio improvement value. The parameter table execution report will be used as the training data for the hybrid time-series optimization model in the next preset training cycle.
6. The control method for a multi-split air conditioner according to claim 1, characterized in that, Before obtaining the optimal parameter table for the current operating cycle based on the historical operating data uploaded by the multi-split air conditioner and using a preset hybrid time-series optimization model, the method further includes: The historical operational data is preprocessed, and the preprocessing includes at least one of the following: Data cleaning and standardization processes include noise reduction, data normalization, and outlier removal. For time-series runtime data within the runtime data, a sliding window method is used to construct time-series features; and / or, The equipment operating life characteristics are obtained, and the equipment operating life characteristics and preprocessed historical operating data are used as input data for the hybrid time series optimization model; wherein, the equipment operating life characteristics include at least the cumulative operating hours of the multi-split air conditioner and the number of compressor start-stop cycles of the multi-split air conditioner.
7. A control method for a multi-split air conditioner, characterized in that, The multi-split air conditioner is communicatively connected to a cloud server, and the method is applied to the multi-split air conditioner, the method comprising: Obtain the optimal parameter table issued by the cloud server; the optimal parameter table is obtained by the cloud server based on the historical operating data uploaded by the multi-split air conditioner, using a preset hybrid time-series optimization model, with the goal of maximizing the energy efficiency ratio of the multi-split air conditioner; Real-time acquisition of current operating condition information; wherein, the current operating condition information includes at least outdoor temperature, indoor unit load rate, and equipment operating time; Based on the current operating condition information, the optimal parameter combination that matches the current operating condition information is obtained from the optimal parameter table; The operation of the multi-split air conditioner is controlled based on the optimal parameter combination.
8. A cloud server, comprising a processor and a memory, said memory being adapted to store multiple lines of program code, characterized in that, The program code is adapted to be loaded and run by the processor to perform the control method for a multi-split air conditioner according to any one of claims 1 to 6.
9. A multi-split air conditioner, comprising a processor and a memory, the memory being adapted to store multiple lines of program code, characterized in that, The program code is adapted to be loaded and run by the processor to perform the control method for a multi-split air conditioner as described in claim 7.
10. A computer-readable storage medium storing a plurality of program codes, characterized in that, The program code is adapted to be loaded and run by a processor to perform the control method for a multi-split air conditioner according to any one of claims 1 to 6 or 7.