Control method of air conditioner and air conditioner

Abstract

The invention provides a control method of an air conditioner and the air conditioner. The method comprises the following steps: step S102, acquiring a current determination coefficient of a currently stored expansion valve opening prediction model; s104, when the current decision coefficient is smaller than a preset threshold value, if the compressor frequency changes, the opening degree of an expansion valve is controlled based on a predicted value of an expansion valve opening degree prediction model, actual measurement operation data before and after the compressor frequency changes are obtained, and the expansion valve opening degree prediction model is updated based on the actual measurement operation data to obtain a temporary prediction model; and S106, calculating a decision coefficient of the temporary prediction model, recording the decision coefficient as a temporary decision coefficient, when the temporary decision coefficient is greater than the current decision coefficient, deleting the expansion valve opening prediction model, taking the temporary prediction model as a new expansion valve opening prediction model, storing the new expansion valve opening prediction model, and returning to execute the step S102 until the current decision coefficient is greater than or equal to a preset threshold value. The method can prevent the accuracy of the prediction model from being affected by the abnormal measured data, and improves the prediction accuracy of the opening degree of the expansion valve.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, and in particular to a control method for an air conditioner and an air conditioner. Background Technology

[0002] Variable Refrigerant Flow (VRF) multi-split air conditioners typically consist of one outdoor unit connected to multiple indoor units. They automatically adjust their capacity according to the needs of each indoor unit, primarily by changing the compressor frequency. Under normal operating conditions, the expansion valve controls the suction superheat by adjusting its opening. Considering that changes in the expansion valve opening may cause a lag in refrigerant response, the expansion valve typically adjusts its opening every tens of seconds, with each adjustment being small, such as adjusting only one step at a time (for example, a valve that is fully open for 500 steps). However, when the load demand of the indoor units changes, the compressor frequency adjusts immediately, but the expansion valve opening adjustment cannot keep up in time. This causes the expansion valve opening to deviate from its optimal state, resulting in a deviation of the suction superheat from the target value, and consequently, a deviation of the actual capacity of the indoor units from their optimal value, affecting comfort. Therefore, when the compressor frequency changes, the expansion valve opening needs to be quickly adjusted to the ideal state.

[0003] Existing expansion valve control technologies fall into two categories: one uses a fixed formula to calculate the expansion valve opening based on compressor frequency changes, which is set in the factory-installed control software without adjustment based on the actual operating environment; the other involves collecting measured data from actual use and adjusting the prediction model for the expansion valve opening to establish a new prediction model. However, the collected measured data may contain outliers that deviate from the characteristics of the equipment. Since the prediction model is usually learned using the input measured data as the correct state data, the accuracy of the newly established prediction model is lower than that of the factory-installed prediction model, thus reducing the accuracy of the expansion valve opening prediction. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a control method and an air conditioner for an air conditioner, which enables the expansion valve opening prediction model to continuously maintain its learning effect during the long-term operation of the air conditioner, thereby avoiding the impact of abnormal measured data on the accuracy of the prediction model, improving the accuracy of expansion valve opening prediction, and improving the reliability of air conditioner control.

[0005] According to an embodiment of the present invention, a control method for an air conditioner is provided, comprising: Step S102: Obtain the determination coefficient of the currently stored expansion valve opening prediction model, denoted as the current determination coefficient, and determine whether the current determination coefficient is greater than or equal to a preset threshold; wherein, the expansion valve opening prediction model is used to predict the expansion valve opening change corresponding to the frequency change, and the determination coefficient is used to characterize the correlation between the predicted opening change and the measured opening change. Step S104: When the current determination coefficient is less than a preset threshold, the expansion valve opening is controlled based on the predicted value of the expansion valve opening prediction model when the compressor frequency changes. The measured operating data before and after the compressor frequency change is obtained, and the expansion valve opening prediction model is updated based on the measured operating data to obtain a temporary prediction model. Step S106: Calculate the determination coefficient of the temporary prediction model and record it as the temporary determination coefficient. When the temporary determination coefficient is greater than the current determination coefficient, delete the expansion valve opening prediction model, store the temporary prediction model as the new expansion valve opening prediction model, and return to execute step S102 until the current determination coefficient is greater than or equal to the preset threshold.

[0006] By adopting the above technical solution, the temporary prediction model is used as the new expansion valve opening prediction model and stored only when the determination coefficient of the temporary prediction model is better than that of the current expansion valve opening prediction model. This allows for continuous optimization of the expansion valve opening prediction model and avoids the impact of abnormal measured data on the accuracy of the prediction model. This improves the accuracy of expansion valve opening prediction and enhances the reliability of air conditioner control.

[0007] Preferably, the measured operating data includes the first and second measured opening degrees of the expansion valve before and after the compressor frequency change; the calculation steps for the determination coefficient include: The change in measured opening is calculated based on the first measured opening and the second measured opening. The compressor frequency change is input into the temporary prediction model to predict the opening change. The residual sum of squares and the total sum of squares are calculated based on the measured change in opening degree and the predicted change in opening degree. The coefficient of determination is calculated based on the residual sum of squares and the total sum of squares. The coefficient of determination is inversely correlated with the residual sum of squares.

[0008] Preferably, the formula for calculating the coefficient of determination is:

[0009] in, The determination coefficient is... The sum of squares of the residuals, The sum of squares is given.

[0010] Preferably, the expansion valve opening prediction model is a linear equation or a neural network model. The expansion valve opening prediction model is constructed or trained based on a stored learning dataset. The learning dataset includes multiple sets of learning data. Each set of learning data includes operating data in a stable operating state before the compressor frequency changes and operating data in a stable operating state after the compressor frequency changes. The operating data includes the compressor frequency and the expansion valve opening.

[0011] Preferably, the step of acquiring measured operating data before and after the compressor frequency change, and updating the expansion valve opening prediction model based on the measured operating data to obtain a temporary prediction model, includes: Acquire first measured operating data when the compressor is in a stable operating state before the frequency change and second measured operating data when the compressor is in a stable operating state after the frequency change; wherein, both the first measured operating data and the second measured operating data include the compressor frequency and the expansion valve opening; A temporary dataset is constructed based on the first measured operating data, the second measured operating data, and the learning dataset. The temporary prediction model is then constructed based on the temporary dataset, or the expansion valve opening prediction model is retrained based on the temporary dataset to obtain the temporary prediction model.

[0012] Preferably, the step of constructing a temporary dataset based on the first measured running data, the second measured running data, and the learning dataset includes: Calculate the squared residual for each group of learning data in the learning dataset, and take the learning data group with the largest squared residual in the learning dataset as the deviation data group; The temporary dataset is constructed based on the other data groups in the learning dataset excluding the deviation data group, the first measured running data, and the second measured running data.

[0013] Preferably, the control method for the air conditioner further includes: When the provisional determination coefficient is greater than the current determination coefficient, the deviation data group in the learning dataset is replaced by the measured data group consisting of the first measured running data and the second measured running data, forming a new learning dataset and storing it.

[0014] Preferably, the control method for the air conditioner further includes: When the provisional determination coefficient is less than or equal to the current determination coefficient, the expansion valve opening prediction model is retained, and the provisional prediction model and the measured data set are deleted.

[0015] Preferably, the stable operating state is that the compressor frequency and the expansion valve opening remain unchanged within a preset time period, and the fluctuation range of the exhaust temperature within the preset time period is less than a preset range.

[0016] According to an embodiment of the present invention, another aspect provides an air conditioner including a computer-readable storage medium storing a computer program and a processor, the computer program being read and executed by the processor to implement the method as described in any of the first aspects.

[0017] The present invention has the following beneficial effects: By updating the expansion valve opening prediction model based on measured operating data after each change in compressor frequency and control of the expansion valve opening based on the predicted value of the expansion valve opening prediction model during the actual use of the air conditioner, the expansion valve opening prediction model can continuously maintain its learning effect during the long-term operation of the air conditioner. Only when the determination coefficient of the temporary prediction model is better than that of the current expansion valve opening prediction model is the temporary prediction model used as the new expansion valve opening prediction model and stored, so as to continuously optimize the expansion valve opening prediction model. At the same time, it can avoid the accuracy of the prediction model being affected by abnormal measured data, thereby improving the accuracy of expansion valve opening prediction and improving the reliability of air conditioner control.

[0018] Other features and advantages of the embodiments of the present invention will be set forth in the following description, or some features and advantages may be inferred from the description or determined without doubt, or may be learned by practicing the techniques described above in the embodiments of the present invention.

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0020] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0021] The structures, proportions, sizes, etc. illustrated in this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0022] Figure 1 A flowchart of a control method for an air conditioner provided by the present invention; Figure 2 An example diagram for calculating the coefficient of determination provided by this invention; Figure 3 This invention provides a schematic diagram of the construction of an initial prediction function; Figure 4 This invention provides a schematic diagram of establishing a linear equation based on the least squares method; Figure 5 A schematic diagram of a neural network-based expansion valve opening prediction model provided by the present invention; Figure 6 A schematic diagram of a stable operating state provided by the present invention; Figure 7 This invention provides a control flowchart for an expansion valve. Detailed Implementation

[0023] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] This embodiment provides a control method for an air conditioner. This method can be applied to the controller of an air conditioner (such as a VRF air conditioner), see below. Figure 1 The flowchart of the control method for the air conditioner shown mainly includes the following steps S102 to S106: Step S102: Obtain the determination coefficient of the currently stored expansion valve opening prediction model, denoted as the current determination coefficient, and determine whether the current determination coefficient is greater than or equal to the preset threshold. Among them, the expansion valve opening prediction model is used to predict the expansion valve opening change corresponding to the frequency change, and the coefficient of determination is used to characterize the correlation between the predicted opening change and the measured opening change. The expansion valve opening prediction model and its corresponding determination coefficients are stored in the controller's memory (e.g., in EEPROM). When the air conditioner leaves the factory, a learning dataset consisting of n sets of learning data measured in the laboratory is stored in the EEPROM. An initial expansion valve opening prediction model (also called a prediction function) is created based on this learning dataset. In the controller's memory, using methods such as the least squares method or neural networks, the initial expansion valve opening prediction model is generated based on the learning dataset, and the determination coefficients of the initial expansion valve opening prediction model are calculated. This expansion valve opening prediction model can be a linear function mathematical model or a neural network model. When the compressor frequency changes, the frequency change is input into the expansion valve opening prediction model, which can predict the corresponding change in expansion valve opening. This change in expansion valve opening is the difference in opening of the expansion valve after the air conditioner returns to a stable operating state compared to before the frequency change.

[0026] The preset threshold value can be in the range of 0.9 to 0.99, preferably 0.95. By judging whether the current determination coefficient is greater than or equal to the preset threshold, the error of the current expansion valve opening prediction model is verified. If the current determination coefficient is greater than or equal to the preset threshold, it indicates that the determination coefficient has reached a sufficiently high accuracy, and the expansion valve opening prediction model does not need to be updated and learned in the future. The currently stored expansion valve opening prediction model is used as a fixed model.

[0027] Step S104: When the current determination coefficient is less than the preset threshold, if the compressor frequency changes, the expansion valve opening is controlled based on the predicted value of the expansion valve opening prediction model. The measured operating data before and after the compressor frequency change is obtained, and the expansion valve opening prediction model is updated based on the measured operating data to obtain a temporary prediction model. If the current determination coefficient is less than the preset threshold, it indicates that there is still room for improvement in the prediction accuracy of the expansion valve opening prediction model. When the compressor frequency changes, it is necessary to continue learning and updating based on the measured operating data.

[0028] The above measured operating data includes measured operating data of the compressor in a stable operating state before the compressor frequency change and measured operating data of the compressor in a stable operating state after the compressor frequency change. The operating data includes the compressor frequency and the opening degree of the expansion valve. Based on the above measured operating data, the change in frequency and the change in opening degree of the expansion valve before and after the compressor frequency change can be calculated.

[0029] When the above expansion valve opening prediction model is a linear function mathematical model, the slope and intercept parameters in the model are updated according to the frequency change and expansion valve opening change corresponding to the measured operating data. The updated model is recorded as the temporary prediction model.

[0030] When the above expansion valve opening prediction model is a neural network model, the above measured operating data may also include indoor ambient temperature, outdoor ambient temperature, exhaust pressure, intake pressure and intake temperature. The above temperature data, pressure data and the frequency change and expansion valve opening change corresponding to the measured operating data are input into the neural network model for training to obtain the updated model, which is denoted as the temporary prediction model.

[0031] Step S106: Calculate the determination coefficient of the temporary prediction model and record it as the temporary determination coefficient. When the temporary determination coefficient is greater than the current determination coefficient, delete the expansion valve opening prediction model, use the temporary prediction model as the new expansion valve opening prediction model and store it, and return to execute step S102 until the current determination coefficient is greater than or equal to the preset threshold.

[0032] The determination coefficient of the temporary prediction model is calculated based on the change in expansion valve opening predicted by the temporary prediction model and the actual change in expansion valve opening. This determination coefficient is denoted as the temporary determination coefficient.

[0033] Determine whether the temporary determination coefficient is greater than the determination coefficient of the currently stored expansion valve opening prediction model. If the temporary determination coefficient is greater than the determination coefficient of the currently stored expansion valve opening prediction model, it indicates that the prediction accuracy of the temporary prediction model is better than the prediction accuracy of the currently stored expansion valve opening prediction model. Delete the currently stored expansion valve opening prediction model and the measured operating data, and return to execute the above step S102 until the determination coefficient of the stored expansion valve opening prediction model is greater than or equal to the preset threshold.

[0034] When the determination coefficient of the expansion valve opening prediction model is greater than or equal to the preset threshold, the learning and updating of the expansion valve opening prediction model is terminated, and thereafter the fixed expansion valve opening prediction model is used for predictive control of the expansion valve opening.

[0035] The air conditioner control method provided in this embodiment updates the expansion valve opening prediction model based on measured operating data after each change in compressor frequency and control of the expansion valve opening based on the predicted value of the expansion valve opening prediction model during actual use of the air conditioner. This allows the expansion valve opening prediction model to continuously maintain its learning effect during long-term operation of the air conditioner. Only when the determination coefficient of the temporary prediction model is better than that of the current expansion valve opening prediction model is the temporary prediction model used as the new expansion valve opening prediction model and stored, so as to continuously optimize the expansion valve opening prediction model. At the same time, it can avoid the accuracy of the prediction model being affected by abnormal measured data, thereby improving the accuracy of expansion valve opening prediction and enhancing the reliability of air conditioner control.

[0036] In one embodiment, the aforementioned measured operating data includes the first and second measured opening degrees of the expansion valve before and after the compressor frequency change; this embodiment provides the calculation steps for the determination coefficient: The change in measured opening degree is calculated based on the first and second measured opening degrees. The compressor frequency change is input into the temporary prediction model to predict the opening change. The residual sum of squares and the total sum of squares are calculated based on the measured and predicted changes in opening degree. The coefficient of determination is then calculated based on the residual sum of squares and the total sum of squares. The coefficient of determination is inversely correlated with the residual sum of squares.

[0037] The difference between the first measured opening degree of the expansion valve before the compressor frequency change and the second measured opening degree of the expansion valve after the compressor frequency change is calculated to obtain the measured opening degree change. The compressor frequency change is input into the updated temporary prediction model to obtain the predicted opening degree change output by the temporary prediction model. The residual sum of squares and the total sum of squares are calculated based on the measured opening degree change and the predicted opening degree change.

[0038] The formula for calculating the sum of squared residuals is: SSR = Σ([measured ΔP] - [predicted ΔP]) 2 The measured ΔP is the measured change in opening, and the predicted ΔP is the predicted change in opening. The formula for calculating the total sum of squares (also known as the total variation) is: SST = Σ([measured ΔP] - [average measured ΔP]) 2 The measured average value of ΔP is the average value of the measured opening changes of multiple sets of data in the learning dataset.

[0039] The formula for calculating the coefficient of determination is as follows:

[0040] in, As the coefficient of determination, For the sum of squared residuals, This is the total sum of squares.

[0041] See also Figure 2 The following is an example diagram of the calculation of the coefficient of determination. Figure 2 The figure shows the sum of squared residuals (SSR), sum of total variation (SST), and coefficient of determination when the learning dataset includes 10 sets of laboratory data. The specific calculation method.

[0042] In one embodiment, the expansion valve opening prediction model provided in this embodiment is a linear equation or a neural network model. The expansion valve opening prediction model is constructed or trained based on a stored learning dataset. The learning dataset includes multiple sets of learning data. Each set of learning data includes operating data in a stable operating state before the compressor frequency changes and operating data in a stable operating state after the compressor frequency changes. All operating data include compressor frequency and expansion valve opening.

[0043] When the expansion valve opening prediction model is a linear equation, the aforementioned operating data includes compressor frequency and expansion valve opening. The linear equation of the expansion valve opening prediction model is obtained by linear fitting based on the frequency change and expansion valve opening change of multiple sets of data in the learning dataset, or by establishing a linear equation using the least squares method based on multiple sets of data in the learning dataset. When the expansion valve opening prediction model is a neural network model, the aforementioned operating data includes compressor frequency, expansion valve opening, and environmental parameters (outdoor ambient temperature, indoor ambient temperature, exhaust pressure, and intake pressure, etc.). Multiple sets of data in the learning dataset are input into the neural network model for model training to obtain the expansion valve opening prediction model.

[0044] See also Figure 3 The diagram illustrating the initial prediction function construction shows that, before the air conditioner leaves the factory, operating conditions are altered in a laboratory environment to acquire stable operating parameters before the frequency change (including compressor frequency, outdoor ambient temperature, indoor ambient temperature, exhaust pressure, suction pressure, and expansion valve opening, etc.) and stable operating frequency and expansion valve opening after the frequency change (e.g., 10 sets of operating status data can be acquired: data sets No. 1 to No. 10). This initial data is stored in an EEPROM so that it can be read from the controller's memory when the air conditioner is powered on. After reading the data, methods such as least squares or neural networks are used to establish the initial prediction function, which is used to calculate the change in expansion valve opening ΔP corresponding to the frequency change ΔF, thus completing the construction of the initial prediction function.

[0045] In one implementation, the expansion valve opening prediction model is a linear equation. The frequency change ΔF is used as the independent variable, and the expansion valve opening change ΔP is used as the computational variable to establish the linear equation. The slope of the linear equation is determined based on the experimental data of each group in the learning dataset. and intercept See also Figure 4 The diagram shown illustrates the establishment of a linear equation using the least squares method. Given the number of data sets, the linear equation slope and intercept The calculation formulas are as follows: = ; = ; In one embodiment, the above-mentioned expansion valve opening prediction model is a neural network model, and the functional expression of the neural network model can be:

[0046] See also Figure 5 The diagram shown illustrates the structure of a neural network-based expansion valve opening prediction model, with frequency variation. and environmental variables (Parameters including indoor ambient temperature, outdoor ambient temperature, exhaust pressure, intake pressure, and intake temperature) are used as input parameters for the expansion valve opening prediction model, and the change in expansion valve opening ΔP is used as the output parameter. For the successively updated learning parameters, the loss function L( It can be defined as:

[0047] By searching the loss function L( The minimum value obtained This allows us to construct a prediction model for the opening degree of the expansion valve.

[0048] In one embodiment, this embodiment provides a specific implementation method for obtaining measured operating data before and after compressor frequency changes, and updating the expansion valve opening prediction model based on the measured operating data to obtain a temporary prediction model: Acquire the first measured operating data when the compressor is in a stable operating state before the frequency change and the second measured operating data when the compressor is in a stable operating state after the frequency change; wherein, both the first and second measured operating data include the compressor frequency and the expansion valve opening; In one embodiment, the above-mentioned stable operating state is that the compressor frequency and the opening degree of the expansion valve remain unchanged within a preset time period, and the fluctuation range of the exhaust temperature within the preset time period is less than a preset range (such as -1℃ to +1℃).

[0049] Before and after the compressor frequency change, it is determined whether the air conditioner is in a stable operating state. When it is in a stable operating state, the first measured operating data before the compressor frequency change and the second measured operating data after the compressor frequency change are acquired. A stable operating state means that the compressor frequency remains constant (or the fluctuation amplitude is less than a preset threshold), while the expansion valve is precisely controlled according to the target superheat, and the expansion valve opening remains constant (or the fluctuation amplitude is less than a preset threshold). At this time, the temperature and pressure changes of various parts of the system also remain basically constant for a certain period of time. The preset duration can range from 5 minutes to 15 minutes, preferably 10 minutes. For example, see... Figure 6 The diagram showing the stable operating state indicates that if the compressor frequency and expansion valve opening remain stable for 10 minutes, and the exhaust temperature fluctuation does not exceed 1°C within 10 minutes, then the stable operating state has been achieved.

[0050] A temporary dataset is constructed based on the first measured operating data, the second measured operating data, and the learning dataset. A temporary prediction model is then constructed based on the temporary dataset, or the expansion valve opening prediction model is retrained based on the temporary dataset to obtain a temporary prediction model.

[0051] A temporary dataset is constructed by combining the first and second measured operating data with the training dataset. Alternatively, a temporary dataset can be built based on a subset of data groups from the training dataset (recently collected data or data groups with small deviations) along with the first and second measured operating data. A temporary prediction model is then constructed using linear fitting or the least squares method described above, or the temporary dataset can be input into the expansion valve opening prediction model for retraining. The aforementioned temporary prediction model is an assumed prediction function, and the first and second measured operating data are assumed sets of measured data. Further evaluation of the accuracy of this temporary prediction model is required.

[0052] In one specific implementation, the squared residuals corresponding to each group of learning data in the learning dataset are calculated, and the learning data group with the largest squared residuals in the learning dataset is taken as the deviation data group. A temporary dataset is constructed based on the other data groups in the learning dataset excluding the deviation data group, the first measured running data, and the second measured running data.

[0053] The squared residuals for each group of learning data in the learning dataset are calculated as follows: Squared Residual = ([Measured ΔP] - [Predicted ΔP])². The group with the largest squared residuals is designated as the deviation group, and this deviation group is used as the maximum deviation value of the prediction model, stored in the controller. This maximum deviation value will serve as a replacement candidate for subsequent data collected by actual equipment. A temporary dataset is then created by combining the learning dataset after removing the deviation groups with the measured array consisting of the first and second measured running data.

[0054] In one embodiment, the method provided in this embodiment further includes: when the temporary determination coefficient is greater than the current determination coefficient, replacing the deviation data group in the learning dataset with the measured data group composed of the first measured running data and the second measured running data to form a new learning dataset and store it.

[0055] When the temporary determination coefficient of the newly established temporary prediction model is greater than the current determination coefficient of the currently stored prediction model, the temporary prediction model is officially adopted for subsequent prediction of the expansion valve opening change. The currently stored expansion valve opening prediction model is deleted, and the temporary prediction model is stored as the new expansion valve opening prediction model. At the same time, the deviation data group in the learning dataset is replaced by the measured data group consisting of the first measured operating data and the second measured operating data to form a new learning dataset and store it in the controller, ensuring that the learning dataset and expansion valve opening prediction model data saved during power failure are not lost.

[0056] In one embodiment, the method provided in this embodiment further includes: when the temporary determination coefficient is less than or equal to the current determination coefficient, retaining the expansion valve opening prediction model and deleting the temporary prediction model and the measured data set.

[0057] When the temporary determination coefficient of the newly established temporary prediction model is less than the current determination coefficient of the currently stored prediction model, it indicates that the prediction accuracy of the currently stored prediction model is high. The temporary prediction model and the measured data are then discarded, and the currently stored prediction model is used to control the expansion valve. At the same time, the expansion valve opening prediction model is updated and learned after the compressor frequency changes, so as to gradually improve the prediction accuracy of the expansion valve opening prediction model in actual operation.

[0058] The control method for the air conditioner provided in this embodiment can predict the opening degree of the expansion valve in the subsequent stable operating state when the compressor frequency changes, and control the expansion valve according to the predicted value. By learning and optimizing the expansion valve control of the VRF air conditioner during operation, the transient instability phenomenon when the compressor frequency changes can be effectively suppressed, the system's start-up performance, comfort and energy saving can be greatly improved, and the system can also automatically adapt to the field environment, with extremely high control reliability.

[0059] Based on the foregoing embodiments, this embodiment provides an example of applying the aforementioned control method for an air conditioner, see, for example... Figure 7 The expansion valve control flowchart shown below can be followed as follows: Step 1: Read 10 sets of data before and after the frequency change from the EEPROM to form a learning dataset and store it in memory.

[0060] A regression equation or neural network model is constructed based on a memory-based learning dataset to obtain a prediction function, which is used to predict ΔF corresponding to ΔP.

[0061] Step 2, calculate the coefficient of determination R² of the prediction function. When R² 2 When the value is less than 0.95, the data with the largest squared residual in the learning dataset is used as the deviation data.

[0062] Step 3: When the air conditioner is in a stable operating state, temporarily store the measured operating data before the stable frequency change in memory.

[0063] Step 4: When the compressor frequency changes, the expansion valve opening is controlled according to the predicted frequency change amount by the prediction function. When the air conditioner is in a stable operating state, the measured operating data after the frequency change is temporarily stored in memory.

[0064] Step 5: Swap the data with the largest deviation in the learning dataset with the actual running data, and calculate the temporary prediction function and the temporary coefficient of determination R².

[0065] Step 6: Determine whether the temporary R² of the temporary prediction function is greater than the R² of the current prediction function. If so, use the temporary prediction function as the subsequent prediction function and apply it to the expansion valve control. The measured running data will be stored in the EEPROM to replace the data group with the largest deviation in the learning dataset.

[0066] Step 7: If not, abandon the use of measured operating data and the temporary prediction function generated based on the measured operating data, and continue to use the original prediction function for expansion valve control.

[0067] Step 8: When R² is greater than or equal to 0.95, the prediction function is considered to have been learned sufficiently, and the prediction function is no longer updated.

[0068] Corresponding to the air conditioner control method provided in the above embodiments, this embodiment provides an air conditioner that includes a computer-readable storage medium storing a computer program and a processor. The computer program is read and executed by the processor to implement the air conditioner control method provided in the above embodiments.

[0069] This embodiment also provides a computer-readable storage medium storing a computer program. When executed by a processor, this computer program implements the various processes of the control method embodiment for the air conditioner described above, and achieves the same technical effect. To avoid repetition, it will not be described again here. The computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0070] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by computer-controlled devices. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the above method embodiments. The storage medium can be a memory, a disk, an optical disk, etc.

[0071] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

[0072] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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 a process, method, article, or apparatus. Without further limitations, 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 said element. The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the air conditioner disclosed in the embodiments, since it corresponds to the control method of the air conditioner disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section. The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0073] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A control method for an air conditioner, characterized in that, include: Step S102: Obtain the determination coefficient of the currently stored expansion valve opening prediction model, denoted as the current determination coefficient, and determine whether the current determination coefficient is greater than or equal to a preset threshold; wherein, the expansion valve opening prediction model is used to predict the expansion valve opening change corresponding to the frequency change, and the determination coefficient is used to characterize the correlation between the predicted opening change and the measured opening change. Step S104: When the current determination coefficient is less than a preset threshold, the expansion valve opening is controlled based on the predicted value of the expansion valve opening prediction model when the compressor frequency changes. The measured operating data before and after the compressor frequency change is obtained, and the expansion valve opening prediction model is updated based on the measured operating data to obtain a temporary prediction model. Step S106: Calculate the determination coefficient of the temporary prediction model and record it as the temporary determination coefficient. When the temporary determination coefficient is greater than the current determination coefficient, delete the expansion valve opening prediction model, store the temporary prediction model as the new expansion valve opening prediction model, and return to execute step S102 until the current determination coefficient is greater than or equal to the preset threshold.

2. The method according to claim 1, characterized in that, The measured operating data includes the first and second measured opening degrees of the expansion valve before and after the compressor frequency change; The steps for calculating the coefficient of determination include: The change in measured opening is calculated based on the first measured opening and the second measured opening. The compressor frequency change is input into the temporary prediction model to predict the opening change. The residual sum of squares and the total sum of squares are calculated based on the measured change in opening degree and the predicted change in opening degree. The coefficient of determination is calculated based on the residual sum of squares and the total sum of squares. The coefficient of determination is inversely correlated with the residual sum of squares.

3. The method according to claim 2, characterized in that, The formula for calculating the coefficient of determination is: in, The determination coefficient is... The sum of squares of the residuals, The sum of squares is given.

4. The method according to claim 1, characterized in that, The expansion valve opening prediction model is a linear equation or a neural network model. The expansion valve opening prediction model is constructed or trained based on a stored learning dataset. The learning dataset includes multiple sets of learning data. Each set of learning data includes operating data in a stable operating state before the compressor frequency changes and operating data in a stable operating state after the compressor frequency changes. The operating data includes the compressor frequency and the expansion valve opening.

5. The method according to claim 4, characterized in that, The step of acquiring measured operating data before and after compressor frequency change, and updating the expansion valve opening prediction model based on the measured operating data to obtain a temporary prediction model, includes: Acquire first measured operating data when the compressor is in a stable operating state before the frequency change and second measured operating data when the compressor is in a stable operating state after the frequency change; wherein, both the first measured operating data and the second measured operating data include the compressor frequency and the expansion valve opening; A temporary dataset is constructed based on the first measured operating data, the second measured operating data, and the learning dataset. The temporary prediction model is then constructed based on the temporary dataset, or the expansion valve opening prediction model is retrained based on the temporary dataset to obtain the temporary prediction model.

6. The method according to claim 5, characterized in that, The step of constructing a temporary dataset based on the first measured running data, the second measured running data, and the learning dataset includes: Calculate the squared residual for each group of learning data in the learning dataset, and take the learning data group with the largest squared residual in the learning dataset as the deviation data group; The temporary dataset is constructed based on the other data groups in the learning dataset excluding the deviation data group, the first measured running data, and the second measured running data.

7. The method according to claim 6, characterized in that, Also includes: When the provisional determination coefficient is greater than the current determination coefficient, the deviation data group in the learning dataset is replaced by the measured data group consisting of the first measured running data and the second measured running data, forming a new learning dataset and storing it.

8. The method according to claim 7, characterized in that, Also includes: When the provisional determination coefficient is less than or equal to the current determination coefficient, the expansion valve opening prediction model is retained, and the provisional prediction model and the measured data set are deleted.

9. The method according to claim 5, characterized in that, The stable operating state is characterized by the compressor frequency and expansion valve opening remaining constant within a preset time period, and the fluctuation range of the exhaust temperature within the preset time period being less than a preset range.

10. An air conditioner, characterized in that, The method includes a computer-readable storage medium storing a computer program, which is read and executed by the processor to implement the method as described in any one of claims 1-9.

Images