Air conditioner compressor frequency control method based on gradient descent

Abstract

The invention discloses an air conditioner compressor frequency control method based on gradient descent, and the method comprises the steps: carrying out the data preprocessing of the operation IOT data of an air conditioner, so as to guarantee the quality and availability of the data; the change of the indoor temperature is predicted through cooling curve fitting; and determining the compressor frequency according to curve gradient calculation. The operation frequency of the air conditioner compressor can be adjusted in real time, and it is ensured that the system can keep excellent robustness and stability under various environment changes. The start-up frequency and the operation frequency are intelligently adjusted through a self-adaptive mechanism, it is ensured that the air conditioner system always operates in the most energy-saving mode, and the energy consumption performance is optimized. The running frequency of the compressor is dynamically controlled by combining real-time prediction of the indoor temperature and gradient analysis of the cooling curve. According to the method, the energy efficiency is improved, the indoor temperature can reach the set value more smoothly by continuously optimizing the operation curve, and the comfortable experience of a user is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of household appliances, and in particular to a frequency control method for an air-conditioning compressor based on gradient descent. Background Art

[0002] Existing air conditioner compressor frequency control usually adopts PID controller. The difficulty of PID control lies in the adjustment of its parameters, including proportion, differential and integral. The setting of these parameters is directly related to the accuracy of temperature control. In addition, the traditional PID controller has weak ability to handle the control of nonlinear systems, slow response speed, and difficulty in quickly adapting to temperature changes.

[0003] At present, there are existing energy-saving control methods and systems, such as patent CN111520883, entitled "An air conditioning control method and system based on target room load prediction", which proposes a method for predicting the load of an air-conditioned room. The invention includes a device-side system and a cloud system: the device-side system includes an air conditioner and a communication module and a data acquisition module installed in the air conditioning control panel; the cloud system includes load model establishment and heating and cooling capacity prediction. The communication module uploads the various data collected by the data acquisition module to the cloud server. The cloud server establishes a target room load prediction model based on the uploaded data, calculates the model weight, and predicts the first instantaneous heating and cooling capacity based on the real-time collected data through the target room load prediction model. The communication module sends the heating and cooling capacity sent from the cloud to the control module, and adjusts the total cooling and heating capacity of the air conditioner through the control module, thereby reducing energy consumption.

[0004] However, the system does not directly control the compression frequency by only changing the set temperature. Its limitations are as follows: first, it is only applicable to multi-split air-conditioning systems, not general household air-conditioning, which limits its usage scenarios; second, the system uses neural networks for load forecasting, which increases computing and storage costs, and has poor adaptability and versatility to different scenarios. Summary of the invention

[0005] The purpose of the present invention is to provide an air-conditioning compressor frequency control method based on gradient descent in order to solve the above problems. The present invention is based on the principle of gradient descent and aims to achieve efficient, accurate and real-time calculation of the operating frequency of the air-conditioning compressor. The method preprocesses the IOT data of the air-conditioning operation to ensure the quality and availability of the data; predicts the change of indoor temperature by fitting the cooling curve; and determines the compressor frequency according to the curve gradient calculation. Compared with the traditional PID control method, the present method can eliminate the hysteresis when adjusting the frequency, reduce the energy waste caused by excessive dehumidification and temperature over-adjustment, and thus achieve the purpose of energy saving.

[0006] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0007] A method for controlling the frequency of an air conditioner compressor based on gradient descent comprises the following steps:

[0008] Step 1, data collection: obtain the temperature and humidity of the room where the air conditioner is located and the operating status data of the air conditioner itself through sensors, and upload them to the cloud for processing in real time through the network;

[0009] Step 2: Data preprocessing: Process the data collected in step 1 to determine whether there are missing values ​​or abnormal values ​​in the data;

[0010] Step 3, performing temperature curve fitting on the data preprocessed in step 2;

[0011] Step 4, for the calculated gradient after the temperature curve fitting in step 3, the frequency is increased or decreased within the time window according to the gradient;

[0012] Step 5, compare the gradient calculated in step 4 with the previous window to calculate the compressor frequency value.

[0013] A further solution is that in step 2, the specific determination rules are as follows:

[0014] Step 2.1 Calculate the number of missing values. If the number of missing values ​​is greater than 80% of the total, delete the data.

[0015] Step 2.2 determines the type of missing values ​​of the feature in step 2.1. If the missing value is a categorical feature, use the mode to fill the missing value; otherwise, use the mean to fill the missing value;

[0016] Step 2.3 uses the quartile method to remove outliers. The daily sampling data is processed separately according to the feature dimensions, sorted by size, and divided into 4 equal parts. The IQR (the difference between the upper quartile Q3 and the lower quartile Q1) is calculated:

[0017] IQR=Q3-Q1

[0018] Remove the data distributed in the interval [-∞, Q1-3*IQR] and [Q3+3*IQR, +∞].

[0019] A further solution is that in step 3, curve fitting is performed on the indoor temperature after data preprocessing, including exponential model, nonlinear transformation, and time window calculation.

[0020] A further solution is an exponential model: an exponential model is used to fit the curve during the cooling process. The model expression is:

[0021] ΔT=T0·e -t / a

[0022] Where ΔT is the initial temperature difference, T0 is the initial temperature value, e -t / a is the calculation coefficient.

[0023] A further solution is that the nonlinear transformation first converts the nonlinear function into a linear function for regression analysis;

[0024]

[0025] The goal is to minimize the sum of squared errors as shown below:

[0026]

[0027] The gradient descent method is used to obtain a. The specific method is as follows:

[0028]

[0029] A further solution is that the time window is calculated as the time window length of each fitting calculation.

[0030] A further solution is that in step 4, the exponential model ΔT = T0·e -t / a Fit the cooling curve and calculate the temperature difference ΔT(k-1) and ΔT(k) before and after the two fittings;

[0031] The temperature drop effect after frequency modulation is determined by the ratio of ΔT(k-1) and ΔT(k), and the change in the required operating frequency Δf is adjusted according to the temperature drop effect, as shown below:

[0032] Where f0 is a fixed coefficient of 12;

[0033] The calculation method for the compressor frequency f(k) sent this time is determined by the previous compression rate f(k-1) and the frequency increment Δf, specifically f(k)=f(k-1)±Δf, where the positive or negative value of Δf is determined by the sign of the temperature change.

[0034] The beneficial effects of the present invention are:

[0035] Real-time control and stability: The operating frequency of the air-conditioning compressor can be adjusted in real time to ensure that the system can maintain excellent robustness and stability under various environmental changes.

[0036] Adaptive frequency adjustment: The startup frequency and operating frequency are intelligently adjusted through the adaptive mechanism to ensure that the air-conditioning system always runs in the most energy-saving mode and optimize energy consumption performance.

[0037] Temperature prediction and curve optimization: Combined with real-time prediction of indoor temperature and gradient analysis of cooling curve, the operating frequency of the compressor is dynamically controlled. This method not only improves energy efficiency, but also continuously optimizes the operating curve to achieve a smoother indoor temperature reaching the set value, enhancing the user's comfort experience.

[0038] Improved comfort: Compared with traditional control methods, the present invention can adjust the indoor temperature more smoothly, reducing the discomfort caused by temperature fluctuations, thereby significantly improving the user's physical comfort. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 It is a flow chart of the air-conditioning compressor frequency control method based on gradient descent of the present invention;

[0041] Figure 2 It is a schematic diagram of the data sliding window processing method of the present invention;

[0042] Figure 3 It is a schematic diagram of the cooling curve fitting method of the present invention;

[0043] Figure 4 It is a schematic diagram of the heating curve fitting method of the present invention. DETAILED DESCRIPTION

[0044] To make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described in detail below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0045] 1. Data collection: The temperature and humidity of the room where the air conditioner is located and the operating status data of the air conditioner itself are obtained through sensors, and uploaded to the cloud for processing in real time through the network;

[0046] 2. Data preprocessing: Process the collected data to determine whether there are missing values ​​or abnormal values ​​in the data. The specific judgment rules are as follows:

[0047] 2.1 Calculate the number of missing values. If the number of missing values ​​is greater than 80% of the total, delete the data;

[0048] 2.2 Determine the type of missing values ​​of the feature in 2.1. If the missing value is a categorical feature, use the mode to fill the missing value; otherwise, use the mean to fill the missing value;

[0049] 2.3 The outliers described in 2 above are removed by the quartile method in this invention. The sampled data of each day is processed separately according to the feature dimension, sorted by size, and divided into 4 equal parts, and the IQR (the difference between the upper quartile Q3 and the lower quartile Q1) is calculated:

[0050] IQR=Q3-Q1

[0051] Remove the data distributed in the interval [-∞, Q1-3*IQR] and [Q3+3*IQR, +∞].

[0052] 3. Cooling curve fitting: Perform curve fitting on the indoor temperature after data preprocessing, including exponential model, nonlinear transformation, and time window calculation. The specific methods are as follows:

[0053] 3.1 Exponential model: The exponential model is used to fit the curve during the cooling process. The model expression is:

[0054] ΔT=T0·e -t / a

[0055] Where ΔT is the initial temperature difference, T0 is the initial temperature value, e -t / a is the calculation coefficient. Figure 3 As shown, the fitting curve can coincide with the actual cooling curve with a small error. This model can be applied to scenarios where the air conditioning temperature difference changes over time.

[0056] 3.2 The nonlinear transformation described in 3 above, according to the regression calculation principle, first converts the nonlinear function into a linear function for regression analysis;

[0057]

[0058] The goal is to minimize the sum of squared errors as shown below:

[0059]

[0060] The gradient descent method is used to obtain a. The specific method is as follows:

[0061]

[0062] 3.3 The time window calculation described in 3 above is as follows: Figure 2 , 3 Shown is the time window length for each fitting calculation, and the curve is fitted every 2 minutes;

[0063] 4. Frequency calculation: After fitting the room temperature curve, the operating frequency of the compressor is calculated by comparing the temperature difference between the two fitting curves. The specific method is as follows:

[0064] 4.1 Compare the temperature difference values ​​ΔT(k-1) and ΔT(k) obtained by exponential fitting, and adjust the increment of the operating frequency according to their ratio, as shown below:

[0065] Where f0 is a fixed coefficient of 12;

[0066] 4.2 Calculate the frequency adjustment value based on the temperature change and the fixed coefficient to ensure that the air conditioner operates in the most energy-saving state. The specific calculation method of the compressor frequency f(k) is:

[0067] f(k)=f(k-1)±Δf, where the sign of Δf is determined by the sign of the temperature change.

[0068] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not further explain various possible combinations. In addition, the various different embodiments of the present invention can also be combined arbitrarily, as long as they do not violate the concept of the present invention, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A method for controlling the frequency of an air-conditioning compressor based on gradient descent, characterized in that: The following steps are involved: Step 1, data collection: obtain the temperature and humidity of the room where the air conditioner is located and the operating status data of the air conditioner itself through sensors, and upload them to the cloud for processing in real time through the network; Step 2: Data preprocessing: Process the data collected in step 1 to determine whether there are missing values ​​or abnormal values ​​in the data; Step 3, performing temperature curve fitting on the data preprocessed in step 2; Step 4, for the calculated gradient after the temperature curve fitting in step 3, the frequency is increased or decreased within the time window according to the gradient; Step 5, compare the gradient calculated in step 4 with the previous window to calculate the compressor frequency value.

2. The air conditioning compressor frequency control method based on gradient descent according to claim 1, characterized in that: In step 2, the specific determination rules are as follows: Step 2.1 Calculate the number of missing values. If the number of missing values ​​is greater than 80% of the total, delete the data. Step 2.2 determines the type of missing values ​​of the feature in step 2.

1. If the missing value is a categorical feature, use the mode to fill the missing value; otherwise, use the mean to fill the missing value; Step 2.3 uses the quartile method to remove outliers. The daily sampling data is processed separately according to the feature dimensions, sorted by size, and divided into 4 equal parts. The IQR (the difference between the upper quartile Q3 and the lower quartile Q1) is calculated: IQR=Q3-Q1 Remove the data distributed in the interval [-∞, Q1-3*IQR] and [Q3+3*IQR, +∞].

3. The air conditioning compressor frequency control method based on gradient descent according to claim 1, characterized in that: In the step 3, curve fitting is performed on the indoor temperature after data preprocessing, including exponential model, nonlinear transformation, and time window calculation.

4. The air conditioning compressor frequency control method based on gradient descent according to claim 3, characterized in that: The exponential model is used to fit the curve during the cooling process, and the model expression is: ΔT=T0·e -t / a Where ΔT is the initial temperature difference, T0 is the initial temperature value, e -t / a is the calculation coefficient.

5. The air-conditioning compressor frequency control method based on gradient descent according to claim 3, characterized in that: The nonlinear transformation first converts the nonlinear function into a linear function for regression analysis; The goal is to minimize the sum of squared errors as shown below: The gradient descent method is used to obtain a. The specific method is as follows:

6. The air conditioning compressor frequency control method based on gradient descent according to claim 3, characterized in that: The time window calculation is the time window length of each fitting calculation.

7. The air conditioning compressor frequency control method based on gradient descent according to claim 1, characterized in that: In step 4, the exponential model ΔT = T0·e -t / a Fit the cooling curve and calculate the temperature difference ΔT(k-1) and ΔT(k) before and after the two fittings; The temperature drop effect after frequency modulation is determined by the ratio of ΔT(k-1) and ΔT(k), and the change in the required operating frequency Δf is adjusted according to the temperature drop effect, as shown below: Where f0 is a fixed coefficient of 12; The calculation method for the compressor frequency f(k) sent this time is determined by the previous compression rate f(k-1) and the frequency increment Δf, specifically f(k)=f(k-1)±Δf, where the positive or negative value of Δf is determined by the sign of the temperature change.

Images