CN-121993887-A - Control method of air conditioner and air conditioner

CN121993887ACN 121993887 ACN121993887 ACN 121993887ACN-121993887-A

Abstract

The invention provides a control method of an air conditioner and the air conditioner, the method comprises the steps of S102, obtaining a current decision coefficient of a currently stored expansion valve opening prediction model, S104, controlling the expansion valve opening based on a predicted value of the expansion valve opening prediction model if the change of compressor frequency is smaller than a preset threshold value when the current decision coefficient is smaller than the preset threshold value, obtaining actual measurement operation data before and after the change of the compressor frequency, updating the expansion valve opening prediction model based on the actual measurement operation data to obtain a temporary prediction model, S106, calculating the decision coefficient of the temporary prediction model, recording the decision coefficient as the temporary decision coefficient, deleting the expansion valve opening prediction model when the temporary decision coefficient is larger than the current decision coefficient, storing the temporary prediction model as a new expansion valve opening prediction model, and returning to the step S102 until the current decision coefficient is larger than or equal to the preset threshold value. The method and the device can avoid influencing the accuracy of the prediction model due to the abnormality of the measured data, and improve the accuracy of the prediction of the opening of the expansion valve.

Inventors

Xue Xiufu

Assignees

宁波奥克斯电气有限公司
奥克斯空调股份有限公司

Dates

Publication Date: 20260508
Application Date: 20251224

Claims (10)

1. A control method of an air conditioner, comprising: Step S102, obtaining a currently stored decision coefficient of an expansion valve opening prediction model, and recording the decision coefficient as a current decision coefficient, and judging whether the current decision coefficient is larger than or equal to a preset threshold value, wherein the expansion valve opening prediction model is used for predicting an expansion valve opening variation corresponding to a frequency variation, and the decision coefficient is used for representing the correlation degree of the predicted opening variation and an actually measured opening variation; Step S104, when the current decision coefficient is smaller than a preset threshold, controlling the opening of the expansion valve based on the predicted value of the expansion valve opening prediction model when the frequency of the compressor changes, obtaining actual measurement operation data before and after the frequency of the compressor changes, and updating the expansion valve opening prediction model based on the actual measurement operation data to obtain a temporary prediction model; And step S106, calculating a decision coefficient of the temporary prediction model, recording the decision coefficient as a temporary decision coefficient, deleting the expansion valve opening prediction model when the temporary decision coefficient is larger than the current decision coefficient, taking the temporary prediction model as a new expansion valve opening prediction model, storing the new expansion valve opening prediction model, and returning to the step S102 until the current decision coefficient is larger than or equal to the preset threshold value.
2. The method of claim 1, wherein the measured operation data includes a first measured opening and a second measured opening of the expansion valve before and after the compressor frequency is changed, and wherein the calculating of the determination coefficient includes: calculating a measured opening variation based on the first measured opening and the second measured opening; inputting the compressor frequency variation into the temporary prediction model to predict to obtain a predicted opening variation; calculating a residual square sum and a total square sum based on the actually measured opening change amount and the predicted opening change amount, and calculating the decision coefficient based on the residual square sum and the total square sum, wherein the decision coefficient is inversely related to the residual square sum.
3. The method of claim 2, wherein the decision coefficient is calculated by: Wherein, the For the said decision coefficient(s), For the sum of squares of the residuals, Is the sum of the total squares.
4. The method of claim 1, wherein the expansion valve opening degree prediction model is a linear equation or a neural network model, the expansion valve opening degree prediction model is constructed or trained based on a stored learning data set, the learning data set includes a plurality of sets of learning data, each set of learning data includes operation data in a steady operation state before a change in a compressor frequency and operation data in a steady operation state after the change in the compressor frequency, and the operation data includes both a compressor frequency and an expansion valve opening degree.
5. The method of claim 4, wherein the step of obtaining actual measurement operation data before and after the compressor frequency change, and updating the expansion valve opening degree prediction model based on the actual measurement operation data to obtain a temporary prediction model, comprises: acquiring first actual measurement operation data when the compressor is in a stable operation state before the frequency change and second actual measurement operation data when the compressor is in the stable operation state after the frequency change, wherein the first actual measurement operation data and the second actual measurement operation data comprise the compressor frequency and the opening of an expansion valve; and constructing a temporary data set based on the first measured operation data, the second measured operation data and the learning data set, constructing the temporary prediction model based on the temporary data set, or retraining the expansion valve opening degree prediction model based on the temporary data set to obtain the temporary prediction model.
6. The method of claim 5, wherein the step of constructing a temporary dataset based on the first measured operational data, the second measured operational data, and the learning dataset comprises: calculating residual error squares corresponding to each group of learning data in the learning data set, and taking the learning data set with the largest data residual error square in the learning data set as a deviation data set; The temporary data set is constructed based on the other data set in the learning data set than the offset data set, the first measured operation data, and the second measured operation data.
7. The method as recited in claim 6, further comprising: And when the temporary decision coefficient is larger than the current decision coefficient, replacing the deviation data set in the learning data set with a measured data set formed by the first measured operation data and the second measured operation data to form a new learning data set and storing the new learning data set.
8. The method as recited in claim 7, further comprising: And when the temporary decision coefficient is smaller than or equal to the current decision coefficient, reserving the expansion valve opening prediction model, and deleting the temporary prediction model and the actual measurement data set.
9. The method of claim 5, wherein the steady state is where the compressor frequency and expansion valve opening remain unchanged for a preset period of time and an amplitude of fluctuation in the discharge temperature for the preset period of time is less than a preset amplitude.
10. An air conditioner comprising a computer readable storage medium storing a computer program and a processor, the computer program implementing the method of any one of claims 1-9 when read and run by the processor.

Description

Control method of air conditioner and air conditioner Technical Field The invention relates to the technical field of air conditioners, in particular to a control method of an air conditioner and the air conditioner. Background Refrigerant variable flow (Variable Refrigerant Flow, VRF) multi-connected air conditioner is usually connected with a plurality of indoor units by an outdoor unit, and the capacity is adjusted by changing the frequency of a compressor according to the requirement of each indoor unit. In a normal operation state, the expansion valve controls the suction superheat degree by adjusting the opening degree, and the expansion valve is usually adjusted in a period of tens of seconds, for example, only 1 step at a time (taking a fully-opened 500-step valve as an example) in consideration of the fact that the refrigerant reaction is delayed due to the change of the opening degree of the expansion valve. However, when the load demand of the indoor unit changes, the compressor frequency is immediately adjusted, and the opening adjustment of the expansion valve is difficult to keep up in time, so that the opening deviates from the optimal state, the suction superheat deviates from the target value, and the actual capacity of the indoor unit deviates from the optimal value, thereby affecting the comfort. Therefore, when the compressor frequency is changed, it is necessary to quickly adjust the opening degree of the expansion valve to an ideal state. The existing expansion valve control technology is that a fixed calculation formula for calculating the opening of the expansion valve according to the frequency variation of the compressor is arranged in control software when leaving a factory, the calculation formula is not adjusted according to the actual use environment, the prediction model of the opening of the expansion valve is adjusted by collecting actual measurement data acquired in the actual use process so as to establish a new prediction model, the collected actual measurement data may be mixed with abnormal values deviating from the characteristics of equipment, but the prediction model is usually learned by taking the input actual measurement data as correct state data, so that the prediction accuracy of the newly established prediction model is lower than that of the prediction model when leaving the factory, and the accuracy of the expansion valve opening prediction is reduced. Disclosure of Invention Accordingly, the present invention is directed to a control method of an air conditioner and an air conditioner, which can keep a learning effect of an expansion valve opening prediction model continuously during a long-term operation of the air conditioner, thereby avoiding influence on accuracy of the prediction model due to abnormality of measured data, improving accuracy of expansion valve opening prediction, and improving reliability of air conditioner control. According to an embodiment of the present invention, there is provided a control method of an air conditioner, including: Step S102, obtaining a currently stored decision coefficient of an expansion valve opening prediction model, and recording the decision coefficient as a current decision coefficient, and judging whether the current decision coefficient is larger than or equal to a preset threshold value, wherein the expansion valve opening prediction model is used for predicting an expansion valve opening variation corresponding to a frequency variation, and the decision coefficient is used for representing the correlation degree of the predicted opening variation and an actually measured opening variation; Step S104, when the current decision coefficient is smaller than a preset threshold, controlling the opening of the expansion valve based on the predicted value of the expansion valve opening prediction model when the frequency of the compressor changes, obtaining actual measurement operation data before and after the frequency of the compressor changes, and updating the expansion valve opening prediction model based on the actual measurement operation data to obtain a temporary prediction model; And step S106, calculating a decision coefficient of the temporary prediction model, recording the decision coefficient as a temporary decision coefficient, deleting the expansion valve opening prediction model when the temporary decision coefficient is larger than the current decision coefficient, taking the temporary prediction model as a new expansion valve opening prediction model, storing the new expansion valve opening prediction model, and returning to the step S102 until the current decision coefficient is larger than or equal to the preset threshold value. By adopting the technical scheme, when the decision coefficient of the temporary prediction model is better than that of the current expansion valve opening prediction model, the temporary prediction model is used as a new expansion valve opening prediction model an