Air conditioner, control method and device thereof, storage medium and computer program product

Abstract

This invention discloses a control method, device, air conditioner, storage medium, and computer program product for an air conditioner. The method includes: acquiring the temperature at at least two locations within a room; determining the average room temperature based on the temperatures at the at least two locations; predicting the actual cooling capacity of the air conditioner using a preset artificial intelligence model; calculating the room heat load based on the average room temperature, the actual cooling capacity, and a preset room heat capacity; determining the target cooling capacity of the air conditioner based on the room heat load and the average room temperature; and outputting control commands to the air conditioner based on the target cooling capacity to adjust the air conditioner's operating state. This solution achieves precise matching between cooling capacity and room load, effectively improving temperature control accuracy, reducing temperature fluctuations, and lowering equipment energy consumption.

Description

Technical Field

[0001] This invention belongs to the field of air conditioning technology, and specifically relates to an air conditioning control method, device, air conditioner, storage medium, and computer program product. Background Technology

[0002] Air conditioners typically employ a single-point temperature acquisition method at the return air vent combined with PID closed-loop control. This method adjusts the compressor and fan operation based solely on single-point temperature feedback, lacking effective perception of the overall room temperature distribution. While some air conditioners are equipped with infrared sensors, these are only used for human presence detection; they do not analyze the room's temperature field distribution, calculate the room's heat load in real time, or predict the air conditioner's cooling capacity. This results in an inaccurate match between the air conditioner's cooling output and the actual room heat load. Consequently, air conditioners exhibit low temperature control accuracy, large temperature fluctuations, and a tendency to overshoot. They also generate significant amounts of ineffective cooling output, leading to high energy consumption and poor energy efficiency. This has become a core technological bottleneck restricting the improvement of high-precision temperature control and high-efficiency energy-saving performance in air conditioners.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The purpose of this invention is to provide an air conditioning control method, device, air conditioner, storage medium, and computer program product to solve the problems in related solutions where the air conditioner lacks global temperature field sensing, real-time heat load calculation, and cooling capacity prediction, resulting in a mismatch between cooling capacity and room load, low temperature control accuracy, large fluctuations, and high energy consumption. This invention achieves precise matching between cooling capacity and room load, effectively improves temperature control accuracy, reduces temperature fluctuations, and lowers equipment energy consumption.

[0005] This invention provides a method for controlling an air conditioner, comprising: acquiring the temperature at at least two locations within a room; determining the average room temperature based on the temperatures at the at least two locations; predicting the actual cooling capacity of the air conditioner using a preset artificial intelligence model; calculating the room heat load based on the average room temperature, the actual cooling capacity, and a preset room heat capacity; determining the target cooling capacity of the air conditioner based on the room heat load and the average room temperature; and outputting control commands for the air conditioner based on the target cooling capacity to adjust the operating state of the air conditioner.

[0006] In some implementations, the temperatures at the at least two locations are collected by an infrared array sensor; determining the average room temperature based on the temperatures at the at least two locations includes: detecting the location of a person in the room and determining the temperature at that location from the temperatures at the at least two locations; and calculating the average room temperature by weighting the temperatures at the at least two locations; wherein the temperature at the location of the person has a greater weight than the temperatures at other locations.

[0007] In some implementations, the artificial intelligence model is a hybrid network model that combines convolutional neural networks and long short-term memory networks.

[0008] In some embodiments, the method further includes: acquiring the room heat load at multiple consecutive times before determining the target cooling capacity of the air conditioner based on the room heat load; and averaging the room heat load at the multiple consecutive times to obtain a smoothed room heat load.

[0009] In some implementations, determining the target cooling capacity of the air conditioner based on the room heat load and the room average temperature includes: determining the magnitude of the absolute value of the difference between the average temperature and the set temperature; if the absolute value of the difference is less than or equal to a preset difference, then the room heat load is used as the target cooling capacity of the air conditioner; if the absolute value of the difference is greater than the preset difference, then the target cooling capacity of the air conditioner is calculated based on the room heat load using a preset PID control method.

[0010] In some implementations, outputting control commands for the air conditioner based on the target cooling capacity includes: limiting the adjustment range of the target cooling capacity according to a preset rate of change; and outputting corresponding control commands based on the limited target cooling capacity.

[0011] In conjunction with the above method, another aspect of the present invention provides an air conditioner control device, comprising: an acquisition unit configured to acquire temperatures at at least two locations within a room; a processing unit configured to determine an average room temperature based on the temperatures at the at least two locations; the processing unit further configured to predict the actual cooling capacity of the air conditioner using a preset artificial intelligence model; the processing unit further configured to calculate a room heat load based on the average room temperature, the actual cooling capacity, and a preset room heat capacity; the processing unit further configured to determine a target cooling capacity of the air conditioner based on the room heat load and the average room temperature; and a control unit configured to output control commands for the air conditioner based on the target cooling capacity to adjust the operating state of the air conditioner.

[0012] In some implementations, the temperatures at the at least two locations are collected by an infrared array sensor; the processing unit determines the average room temperature based on the temperatures at the at least two locations, including: detecting the location of a person in the room, determining the temperature of the person's location from the temperatures at the at least two locations; and calculating the average room temperature based on the temperatures at the at least two locations by weighting; wherein the weight of the temperature of the person's location is greater than the weight of the temperatures at other locations.

[0013] In some implementations, the artificial intelligence model is a hybrid network model that combines convolutional neural networks and long short-term memory networks.

[0014] In some embodiments, the processing unit is further configured to: acquire the room heat load at multiple consecutive times before determining the target cooling capacity of the air conditioner based on the room heat load; and average the room heat load at the multiple consecutive times to obtain a smoothed room heat load.

[0015] In some implementations, the processing unit determines the target cooling capacity of the air conditioner based on the room heat load and the room average temperature, including: determining the magnitude of the absolute value of the difference between the average temperature and the set temperature; if the absolute value of the difference is less than or equal to a preset difference, then the room heat load is used as the target cooling capacity of the air conditioner; if the absolute value of the difference is greater than the preset difference, then the target cooling capacity of the air conditioner is calculated based on the room heat load using a preset PID control method.

[0016] In some implementations, the control unit outputs control commands for the air conditioner based on the target cooling capacity, including: limiting the adjustment range of the target cooling capacity according to a preset rate of change; and outputting corresponding control commands based on the limited target cooling capacity.

[0017] In conjunction with the above-described device, the present invention further provides an air conditioner, comprising: the control device for the air conditioner described above.

[0018] In conjunction with the above method, the present invention further provides a storage medium comprising a stored program, wherein, when the program is executed, the device on which the storage medium is located controls the air conditioner control method described above to be performed.

[0019] In conjunction with the above method, the present invention further provides a computer program product comprising a computer program that, when processed and executed, implements the steps of the above-described air conditioner control method.

[0020] The solution of this invention collects temperature data from multiple indoor points and calculates the average room temperature. It then uses a model to predict the actual cooling capacity of the air conditioner, calculates the real-time room heat load based on the principles of room heat capacity and heat balance, determines the target cooling capacity based on the heat load and room temperature, and finally adjusts the air conditioner's operating status through control commands. This achieves precise matching between cooling capacity and room load, significantly improves temperature control accuracy, reduces temperature fluctuations, and lowers equipment energy consumption.

[0021] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention.

[0022] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating an embodiment of the air conditioner control method of the present invention;

[0024] Figure 2 This is a schematic diagram of the structure of an embodiment of the air conditioner control device of the present invention;

[0025] Figure 3 This is a schematic diagram of a neural network model for artificial intelligence.

[0026] Figure 4 This is a flowchart illustrating an air conditioning energy-saving control method.

[0027] Referring to the accompanying drawings, the reference numerals in the embodiments of the present invention are as follows:

[0028] 102-Acquisition unit; 102-Processing unit; 103-Control unit. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0030] According to embodiments of the present invention, an air conditioning control method is provided, such as... Figure 1 The flowchart of an embodiment of the method of the present invention is shown. The air conditioner control method may include steps S110 to S160.

[0031] In step S110, the temperature at at least two locations in the room is obtained.

[0032] The temperatures at at least two locations are collected using an infrared array sensor. The infrared array sensor consists of multiple temperature measuring units arranged in an array, which can simultaneously collect temperature values ​​at multiple locations within the space, enabling large-scale, multi-point temperature data collection indoors.

[0033] Traditional air conditioners only collect temperature data from a single location, failing to comprehensively reflect the indoor temperature distribution and prone to measurement errors. In contrast, infrared array sensors collect temperature data from multiple points indoors, comprehensively and accurately capturing temperature information from different areas, thus improving control accuracy.

[0034] The infrared array sensor is specifically a 32×32 infrared array with a sampling frequency of 1Hz (Δt=1s); each frame outputs 32×32=1024 temperature points T_ir (i,j,t). The sensors are distributed and installed in the center of the room ceiling to ensure that the detection range covers the entire area.

[0035] In step S120, the average room temperature is determined based on the temperatures at the at least two locations.

[0036] In some implementations, determining the average room temperature based on the temperatures at the at least two locations includes: detecting the location of a person in the room and determining the temperature at that location from the temperatures at the at least two locations; and calculating the average room temperature by weighting the temperatures at the at least two locations, wherein the temperature at the location of the person has a greater weight than the temperatures at other locations.

[0037] Specifically, firstly, the system identifies and locates the specific position and corresponding area of ​​the person inside the room. Secondly, the system initiates a weighted calculation process to average all the collected temperature data. During the calculation, the weight assigned to the temperature of the person's location is greater than the weight assigned to the temperature of other locations in the room, so that the final average room temperature is closer to the human body's comfort needs, effectively improving the comfort of the people inside the room.

[0038] Optionally, before determining the average room temperature, a preprocessing procedure is included for the collected temperatures, which includes bad pixel removal and time low-pass smoothing.

[0039] Defect removal (3×3 median filtering): T_med(i,j,t) = median temperature of 8 neighboring points; if |T_ir(i,j,t)-T_med(i,j,t)|>2℃, then replace the original value with T_med. Setting a 3×3 median filter can effectively remove isolated defective pixels without disrupting the overall temperature field distribution; the temperature difference between adjacent pixels in a normal room will not exceed 2℃, otherwise it will be judged as a defective pixel.

[0040] Time-based low-pass smoothing (coefficient α=0.3): T_filt(i,j,t)=0.3×T_ir(i,j,t)+0.7×T_filt(i,j,t-1). Too small an α (e.g., <0.2) will result in a delayed temperature response, failing to capture temperature changes in a timely manner; too large an α (e.g., >0.4) will lead to insufficient smoothing effect, failing to suppress instantaneous fluctuations; α=0.3 balances real-time performance and smoothness.

[0041] Preprocessing before calculating the average room temperature ensures the reliability of the temperature data. For example, for the human body area: 30℃≤T_filt(i,j,t)≤36℃, weight w=1.2; for other areas, weight w=1.0; the average room temperature T_room(t)=[the sum of all points w(i,j,t)×T_filt(i,j,t)]÷[the sum of all points w(i,j,t)].

[0042] In step S130, the actual cooling capacity of the air conditioner is predicted using a preset artificial intelligence model.

[0043] In some implementations, the artificial intelligence model is a hybrid network model that combines convolutional neural networks and long short-term memory networks.

[0044] Convolutional neural networks excel at processing multi-dimensional structured data, efficiently extracting key features that affect cooling capacity; long short-term memory networks excel at processing time-series data, learning the dynamic patterns of air conditioner operating status changes over time. When these two are combined, the model can capture both the spatial correlation between multiple parameters and cooling capacity, and the temporal correlation between changes in operating conditions and cooling capacity, thus more accurately predicting the actual cooling capacity of the air conditioner under different operating conditions. Figure 3 The diagram illustrates the hybrid network model. First, the input layer receives multi-dimensional parameters of the air conditioner's operation and the environment. The two-layer convolutional neural network (CNN) of the feature extraction layer extracts the coupled features of the environment and load parameters. Then, the processed temporal features are fed into the long short-term memory network (LSTM) of the temporal processing layer to capture the slow changing trends of parameters such as temperature and humidity. Finally, the output layer outputs the prediction results.

[0045] The model's inputs include environmental and operational parameters, and its outputs are cooling / heating capacity (Q_cool / Q_heat). Environmental parameters include: indoor equipment area temperature (T_eq), with a sampling accuracy of ±0.1℃ and a physical range of 0~45℃, collected by distributed temperature probes avoiding equipment heat dissipation vents; air conditioner indoor air inlet temperature (T_in), with a sampling accuracy of ±0.1℃ and a physical range of 0~45℃, collected by the air conditioner's built-in temperature sensor; outdoor ambient temperature (T_out), with a sampling accuracy of ±0.2℃ and a physical range of -10~48℃, collected by an outdoor temperature sensor; and indoor relative humidity (RH_in), with a sampling accuracy of ±3%RH and a physical range of 0%~100%RH, collected by a humidity sensor.

[0046] Operating parameters include the current compressor frequency (f_comp), with a sampling accuracy of ±1Hz and a physical range of 30~120Hz for inverter air conditioners, read through communication with the air conditioner controller; the current indoor fan speed (v_fan_in), with a sampling accuracy of ±10rpm and a physical range of 0~2000rpm, collected by the fan speed sensor; the current outdoor fan speed (v_fan_out), with a sampling accuracy of ±10rpm and a physical range of 0~2000rpm, collected by the fan speed sensor; and the current expansion valve opening (K_valve), with a sampling accuracy of ±1% and a physical range of 20%~80% of the mechanical limit, read through communication with the expansion valve controller.

[0047] The model training process involves collecting at least 10,000 sets of historical steady-state data covering the entire operating range, and removing invalid data such as start-stop transition states and sensor malfunctions to ensure data representativeness and purity. In the data preprocessing stage, Z-score normalization is used to convert each feature into a standard distribution with a mean of 0 and a variance of 1 to eliminate dimensional differences between different parameters. Then, Pearson correlation analysis is used to remove redundant features; for example, if the correlation between the indoor air inlet temperature and the indoor equipment area temperature is greater than 0.9, one of them can be retained to reduce the model's computational cost.

[0048] The model training uses the mean squared error (MSE), which performs well in regression prediction, as the loss function, and selects the Adam optimizer. The initial learning rate is set to 0.001, and the decay rate is set to 0.0001 to achieve fast and stable convergence and avoid training oscillations and overfitting risks. The maximum number of training epochs is set to 200. Training is stopped when the validation set error does not exceed 3% to ensure that the model accuracy meets engineering requirements. To further improve the model's generalization ability, a Dropout layer is added to the network with a dropout rate of 0.2. During training, some neurons are randomly deactivated to prevent the model from over-relying on training data, thereby solving the problems of overfitting and poor generalization ability, and ensuring that the model can accurately predict cooling capacity under different operating conditions.

[0049] The collection of 10,000 sets of data aims to cover all operating conditions with different ambient temperatures, humidity levels, and air conditioning operating parameters, laying the foundation for the model's generalization ability. Eliminating transitional and fault data prevents invalid data from interfering with model accuracy. Z-score normalization and correlation analysis optimize model input from two dimensions: data standardization and feature simplification, improving training efficiency. A reasonable loss function, optimizer configuration, and Dropout layer design collectively ensure the model's convergence speed, prediction accuracy, and long-term stability.

[0050] In step S140, the room heat load is calculated based on the average room temperature, the actual cooling capacity, and the preset room heat capacity.

[0051] The room's heat capacity determines how quickly the indoor temperature changes with heat, the room's average temperature reflects the current overall temperature level, and the air conditioner's actual cooling capacity represents its current cooling output capability. By combining these three parameters, the total heat generated in real time indoors can be accurately calculated, i.e., the room's heat load.

[0052] The formula for estimating room heat capacity is: C_room = 1.2 × air density × room volume × air specific heat × 1.2. Where, air density is taken as a fixed value of 1.2 kg / m³ under standard operating conditions. 3 The specific heat of air is taken as a fixed value of °C under standard operating conditions. Using these fixed values ​​simplifies the calculation process while meeting engineering accuracy requirements. The two coefficients 1.2 in the formula serve different purposes: the first 1.2 is a wall heat capacity correction coefficient, considering that wall heat storage accounts for approximately 20% of the total room heat storage; this coefficient corrects for wall heat storage and incorporates it into the overall estimate. The second 1.2 is a safety factor, used to avoid estimation deviations caused by differences in room structure. Taking a room with dimensions of 5m × 4m × 3m as an example, substituting into the formula yields a room heat capacity of approximately 86.83 kJ / °C. In practical applications, the main control module can automatically complete the calculation by inputting the room's length, width, and height parameters, thus solving the problem of inaccurate room heat capacity acquisition and providing reliable basic parameters for subsequent heat load calculations.

[0053] The calculation of room heat load is based on the law of conservation of energy, that is, the heat corresponding to the change in room temperature is equal to the difference between the room heat load and the air conditioning cooling capacity. The energy balance equation is: C_room×(T_room(t)-T_room(t-1)) / Δt=Q_load(t)-Q_cool(t). Considering the time interval Δt=1s in the control sequence, the formula can be simplified to a direct calculation of the real-time load: Q_load(t)=Q_cool(t)+C_room×(T_room(t)-T_room(t-1)). In this formula, the unit of Q_load(t) is kW. This simplified form avoids complex differential calculations, effectively reducing the system's computing power requirements. In actual operation, the main control module can read the current average room temperature T_room and the actual air conditioning cooling capacity Q_cool data in real time, substitute them into the formula to complete the calculation, and solve the problem of inaccurate real-time heat load detection in traditional control.

[0054] In some embodiments, the method further includes: acquiring the room heat load at multiple consecutive times before determining the target cooling capacity of the air conditioner based on the room heat load; and averaging the room heat load at the multiple consecutive times to obtain a smoothed room heat load.

[0055] In actual operation, factors such as indoor temperature and equipment operating status can fluctuate instantaneously, leading to significant random fluctuations in the real-time calculated room heat load. By averaging the room heat load over multiple consecutive moments before determining the target cooling capacity, random noise and instantaneous fluctuations in the data can be effectively filtered out, resulting in a more stable and representative heat load value. This reduces the need for frequent adjustments to the air conditioning system's operating status, resulting in more stable cooling output. This improves the accuracy and comfort of indoor temperature control and avoids additional wear and tear on air conditioning components caused by frequent adjustments.

[0056] A 5-point moving average algorithm can be used to smooth the heat load data. The calculation formula is: Q_load_smooth(t) = [Q_load(t) + Q_load(t-1) + Q_load(t-2) + Q_load(t-3) + Q_load(t-4)] / 5. By caching the continuous heat load data of the current moment and the previous four moments (a total of 5 seconds), the arithmetic mean is calculated to obtain the smoothed heat load value Q_load_smooth. The time window corresponding to the 5 sampling points can effectively filter out heat load abrupt changes and instantaneous noise caused by factors such as human movement and equipment start-up and shutdown. At the same time, the window length is short, which will not cause obvious response lag, thus balancing the data smoothing effect and the dynamic response speed of the system.

[0057] In step S150, the target cooling capacity of the air conditioner is determined based on the room heat load and the room average temperature.

[0058] The system combines the total indoor heat production with the current overall temperature to determine the optimal cooling output target for the current operating conditions. When the indoor temperature is close to the ideal state, the target cooling capacity is mainly set to match the room's heat load to ensure a balance between heat supply and demand. When the indoor temperature deviates from the ideal state, the target cooling capacity will be adjusted synchronously according to the temperature deviation, taking into account both temperature control and heat balance requirements, so that the air conditioner's cooling output matches the actual indoor demand.

[0059] In some implementations, step S150, determining the target cooling capacity of the air conditioner based on the room heat load and the room average temperature, includes: determining the magnitude of the absolute value of the difference between the average temperature and the set temperature; if the absolute value of the difference is less than or equal to a preset difference, then the room heat load is used as the target cooling capacity of the air conditioner; if the absolute value of the difference is greater than the preset difference, then the target cooling capacity of the air conditioner is calculated based on the room heat load using a preset PID control method.

[0060] Specifically, the system acquires the current average room temperature in real time and compares it with the user-set temperature. It calculates the absolute value of the difference and then compares this absolute value with a preset difference to determine the control status of the indoor temperature. When the absolute value of the difference between the average room temperature and the set temperature is less than or equal to the preset difference, it indicates that the indoor temperature is close to the target state. In this case, the air conditioner does not need to significantly adjust its cooling output; the system directly uses the room heat load as the target cooling capacity of the air conditioner, ensuring that the cooling output of the air conditioner is roughly equal to the amount of heat generated indoors, thus maintaining a stable indoor temperature. When the absolute value of the difference between the average room temperature and the set temperature is greater than the preset difference, it indicates that the indoor temperature deviates significantly from the target state. In this case, it is necessary to dynamically adjust the cooling output to quickly bring the temperature closer to the target value. The system calculates the target cooling capacity of the air conditioner based on the room heat load using a preset PID control method, and dynamically adjusts the cooling output in conjunction with the magnitude and trend of the temperature deviation, balancing temperature adjustment speed and control stability.

[0061] For example, the difference e(t) between the average room temperature T_room and the user-set temperature T_set can be calculated using the formula e(t) = T_room(t) - T_set.

[0062] A steady state is defined as when |e(t)| ≤ 0.1℃ and this state lasts for 30 seconds. Under steady-state conditions, the target cooling capacity equals the smoothed room heat load, i.e., Q_target(t) = Q_load_smooth(t). Setting a deviation threshold of 0.1℃ meets the requirements for high-precision temperature control; the 30-second duration requirement avoids misjudging the steady state due to instantaneous temperature fluctuations. Using a direct balance between cooling capacity and heat load under steady-state conditions maintains stable room temperature while avoiding redundant or insufficient cooling capacity, thus achieving energy savings.

[0063] Under unsteady-state conditions, the system uses a PID control algorithm to dynamically calculate the target cooling capacity. The formula is: Q_target(t)=Q_load_smooth(t)+Kp×e(t)+Ki×(e(0)+e(1)+…+e(t))×Δt+Kd×(e(t)-e(t-1)) / Δt. This is suitable for a 1.5 HP air conditioner, compatible with 50~70m². 2For a room as an example, the recommended PID parameters are: Kp = 0.8 kW / ℃, Ki = 0.05 kW / (℃·s), Kd = 0.1 kW·s / ℃. To protect the equipment, the target cooling capacity must be limited to the air conditioner's operating range, i.e., Q_min ≤ Q_target(t) ≤ Q_max, such as the common range of 0.3~3.5 kW. The proportional coefficient Kp is used for rapid response to temperature deviations to avoid cumulative deviations; the integral coefficient Ki is used to eliminate long-term temperature control offsets, and a smaller value can prevent integral saturation; the derivative coefficient Kd is used to predict the trend of deviation changes, suppress temperature overshoot, and avoid excessive cooling capacity adjustment. By programming the PID algorithm and judging the cooling capacity limit, the problems of low temperature control accuracy and easy overshoot under dynamic operating conditions can be effectively solved, while protecting the air conditioner compressor from damage caused by operating outside its range.

[0064] In step S160, a control command for the air conditioner is output based on the target cooling capacity to adjust the operating status of the air conditioner.

[0065] The system converts the calculated target cooling capacity into operating commands that can be recognized by each component of the air conditioner. These commands adjust operating parameters such as compressor operating frequency, fan speed, and throttling device opening, thereby changing the actual cooling output of the air conditioner. After completing the state adjustment, the air conditioner continues to feed back operating data to the front-end process. The entire control process is executed cyclically to achieve continuous regulation of the indoor temperature.

[0066] In some implementations, outputting control commands for the air conditioner based on the target cooling capacity includes: limiting the adjustment range of the target cooling capacity according to a preset rate of change; and outputting corresponding control commands based on the limited target cooling capacity.

[0067] Specifically, firstly, after determining the target cooling capacity of the air conditioner, the system does not directly convert this value into a control command. Instead, it limits the adjustment range of the target cooling capacity according to a preset rate of change. For example, the system compares the current target cooling capacity with the actual cooling output value of the air conditioner at the previous moment, calculates the difference between the two, and ensures that the change range of cooling capacity per unit time does not exceed the preset maximum value. If the change range of the target cooling capacity exceeds the preset rate of change, the system will adjust it to the maximum allowable change range to obtain the limited target cooling capacity value; if the change range is within the preset range, the target cooling capacity value is directly retained.

[0068] Secondly, the system converts the limited target cooling capacity into operating commands that can be recognized by each component of the air conditioner, thereby adjusting operating parameters such as compressor operating frequency, fan speed, and throttling device opening, and changing the actual cooling output of the air conditioner.

[0069] For example, setting the maximum rate of change of cooling capacity to ΔQ_rate = 0.3kW / s is suitable for the adjustment capability of a 1.5HP inverter air conditioner compressor, preventing equipment damage and temperature overshoot caused by excessively rapid frequency adjustments. The formula for calculating the actual cooling capacity is: Q_apply(t) = Q_apply(t-1) + limit(Q_target(t) - Q_apply(t-1), -0.3, +0.3). In this formula, the limit operation restricts the adjustment range of cooling capacity per unit time to within ±0.3kW / s. Setting the initial cooling capacity of the system to the minimum cooling capacity Q_min of the air conditioner avoids a sudden drop in room temperature due to excessive cooling capacity at startup.

[0070] In actual operation, the core components of an air conditioner, such as the compressor and fan, cannot withstand drastic, instantaneous changes in cooling capacity. If the target cooling capacity is directly converted into a control command, significant fluctuations in the target cooling capacity will cause drastic changes in compressor frequency and fan speed within a short period. This not only easily leads to mechanical wear on equipment components but also causes drastic fluctuations in air conditioning system pressure and temperature, affecting the stability and lifespan of the equipment. By limiting the adjustment range of the target cooling capacity through preset change rates, the cooling output of the air conditioner can be made more stable, avoiding additional wear on components due to frequent and drastic adjustments, extending the equipment's lifespan, and ensuring gradual changes in indoor temperature, thus improving user comfort.

[0071] In some embodiments, the process also includes temperature control accuracy assessment and online correction of room heat capacity.

[0072] Temperature control accuracy assessment: The system assesses temperature control accuracy every 10 minutes (600 seconds), using the root mean square error (RMSE) as the evaluation metric. The calculation formula is: RMSE = square root [average of (T_room - T_set)² over the most recent 600 seconds]. The assessment period is set to 10 minutes to avoid the problems of too short a period being susceptible to instantaneous fluctuations and too long a period failing to optimize parameters in a timely manner. Based on the assessment results, the system adaptively adjusts the PID parameters. When RMSE > 0.3℃, it indicates insufficient temperature control accuracy, and the system increases the proportional coefficient Kp to improve the response speed to temperature deviations. When RMSE < 0.1℃ but the temperature overshoot is greater than 0.5℃, it indicates excessive derivative adjustment, and the system decreases the derivative coefficient Kd to suppress temperature overshoot.

[0073] Online room heat capacity correction: Under steady-state conditions, the rate of change of room temperature dT_room / dt≈0, at which point the air conditioning cooling capacity and room heat load are approximately equal. The system can use historical data from steady-state operation to infer the room heat capacity, achieving online parameter correction. The correction adopts a weighted update method: C_room_new = room heat capacity calculated from steady-state data, C_room = 0.9 × old C_room + 0.1 × C_room_new. In the formula, the old value is weighted at 0.9 to ensure the stability of the room heat capacity parameter; the new value is weighted at 0.1 to achieve slow adaptive correction and avoid parameter abrupt changes caused by single calculation deviations. This mechanism solves the problem of load calculation deviation caused by changes in room heat capacity due to environmental factors such as humidity and furniture placement. Through steady-state data caching and weighted calculation, it effectively improves the long-term temperature control accuracy of the system.

[0074] This solution relies on artificial intelligence technology and the principle of thermal balance to accurately calculate cooling capacity and room heat load. It can dynamically match the air conditioner's cooling output according to the indoor heat and temperature status, effectively avoiding data deviation caused by single temperature measurement. This allows the air conditioner's cooling output to be accurately matched with the actual indoor heat demand, which not only improves the control accuracy of indoor temperature and reduces the temperature fluctuation range, but also reduces the ineffective cooling work of the air conditioner and reduces the energy consumption of the equipment.

[0075] For example, temperature data is continuously collected at three locations in the living room: a wall, the sofa area, and near the window. The system integrates the temperature data from these three locations to calculate the average temperature of the living room. Then, information such as the outdoor temperature and the air conditioner's operating level is input into an artificial intelligence model to determine the actual cooling capacity that the air conditioner can currently output. Combining the living room's inherent heat capacity, the current average temperature, and the air conditioner's actual cooling capacity, the system calculates the room's heat load generated by people, electrical appliances, and sunlight from outside. Based on this heat load and the average temperature of the living room, the system determines the target cooling capacity that the air conditioner needs to achieve, and then issues corresponding operating instructions to adjust the air conditioner's compressor speed so that the cooling output of the air conditioner perfectly matches the heat generation of the living room, maintaining a stable living room temperature.

[0076] Figure 4 The flowchart of the air conditioning energy-saving control method is shown, specifically including steps 1 to 9.

[0077] Step 1: After the system starts up, the initial configuration is completed first, including calibrating the infrared sensor and temperature probe through a standard temperature source to eliminate equipment system errors; estimating the room heat capacity C_room; and setting the initial cooling capacity of the air conditioner Q_apply to the minimum value Q_min.

[0078] Step 2: The infrared array sensor continuously collects temperature data from multiple points indoors, performs noise reduction and smoothing on the data, calculates the average room temperature T_room based on the human body area weighting algorithm, and outputs indoor temperature distribution information.

[0079] Step 3: The system inputs the air conditioner operating parameters and environmental data into the pre-trained AI model, and the model outputs the real-time cooling capacity Q_cool of the air conditioner under the current operating conditions.

[0080] Step 4: Based on the law of conservation of energy, using the room heat capacity C_room, the average room temperature change, and the real-time cooling capacity of the air conditioner Q_cool, calculate the real-time room heat load Q_load through the heat balance equation.

[0081] Step 5: Perform a 5-point moving average on the calculated real-time heat load Q_load to obtain a smoothed heat load Q_load_smooth, filtering out instantaneous fluctuations and noise.

[0082] Step 6: Combine the smoothed heat load Q_load_smooth with the temperature deviation, and calculate the target cooling capacity Q_target by combining feedforward control and PID control: In steady state, the heat load value is directly used, and in dynamic conditions, the temperature deviation is quickly reduced by PID adjustment, taking into account both temperature control stability and response speed.

[0083] Step 7: Perform rate-of-change limiting processing on the target cooling capacity Q_target to limit the adjustment range of cooling capacity per unit time within a preset range, so as to avoid compressor wear and temperature overshoot caused by sudden changes in cooling capacity, and ensure a smooth transition of cooling output.

[0084] Step 8: Convert the limited cooling capacity Q_apply into control commands, adjust the air conditioner compressor frequency, fan speed and other components, output the actual cooling capacity, and achieve precise adjustment of indoor temperature.

[0085] Step 9: The system re-collects indoor temperature data, evaluates the temperature control accuracy in 10-minute cycles, and adaptively adjusts the PID parameters. At the same time, under steady-state conditions, the room heat capacity C_room is corrected online using historical data to adapt to changes in the room environment and equipment status, forming a long-term stable closed-loop control.

[0086] Through the above control, the steady-state temperature control accuracy is achieved as follows: |T_room-T_set|≤0.2℃; temperature overshoot: <0.5℃; energy saving rate: 10%~15% energy saving compared to ordinary inverter air conditioners, and 20%~35% energy saving compared to fixed frequency air conditioners.

[0087] The technical solution adopted in this embodiment collects indoor temperatures at multiple points, calculates the average temperature, predicts the air conditioning cooling capacity using a model, and calculates the real-time heat load based on the principles of room heat capacity and heat balance. Based on this, the target cooling capacity is dynamically determined and the air conditioning operation is adjusted, forming a closed-loop control. This effectively solves the problems of traditional air conditioners, which lack comprehensive temperature sensing, real-time heat load calculation, and cooling capacity prediction, leading to mismatches between cooling capacity and room load, low temperature control accuracy, large fluctuations, and high energy consumption. It achieves precise matching between cooling capacity and room load, significantly improves temperature control accuracy, reduces temperature fluctuations, and simultaneously lowers equipment energy consumption.

[0088] According to an embodiment of the present invention, an air conditioner control device corresponding to the air conditioner control method is also provided. See also Figure 2 The schematic diagram shown is a structural diagram of an embodiment of the device of the present invention. The control device of the air conditioner may include: an acquisition unit 101, a processing unit 102, and a control unit 103.

[0089] Acquisition unit 101 is configured to acquire the temperature at at least two locations within the room.

[0090] The temperatures at at least two locations are collected using an infrared array sensor. The infrared array sensor consists of multiple temperature measuring units arranged in an array, which can simultaneously collect temperature values ​​at multiple locations within the space, enabling large-scale, multi-point temperature data collection indoors.

[0091] Traditional air conditioners only collect temperature data from a single location, failing to comprehensively reflect the indoor temperature distribution and prone to measurement errors. In contrast, infrared array sensors collect temperature data from multiple points indoors, comprehensively and accurately capturing temperature information from different areas, thus improving control accuracy.

[0092] The infrared array sensor is specifically a 32×32 infrared array with a sampling frequency of 1Hz (Δt=1s); each frame outputs 32×32=1024 temperature points T_ir (i,j,t). The sensors are distributed and installed in the center of the room ceiling to ensure that the detection range covers the entire area.

[0093] The processing unit 102 is configured to determine the average room temperature based on the temperatures at the at least two locations.

[0094] In some embodiments, the processing unit 102 determines the average room temperature based on the temperatures at the at least two locations, including: detecting the location of a person in the room and determining the temperature of the person's location from the temperatures at the at least two locations; and calculating the average room temperature by weighting the temperatures at the at least two locations; wherein the weight of the temperature of the person's location is greater than the weight of the temperatures at other locations.

[0095] Specifically, firstly, the system identifies and locates the specific position and corresponding area of ​​the person inside the room. Secondly, the system initiates a weighted calculation process to average all the collected temperature data. During the calculation, the weight assigned to the temperature of the person's location is greater than the weight assigned to the temperature of other locations in the room, so that the final average room temperature is closer to the human body's comfort needs, effectively improving the comfort of the people inside the room.

[0096] Optionally, before determining the average room temperature, a preprocessing procedure is included for the collected temperatures, which includes bad pixel removal and time low-pass smoothing.

[0097] Defect removal (3×3 median filtering): T_med(i,j,t) = median temperature of 8 neighboring points; if |T_ir(i,j,t)-T_med(i,j,t)|>2℃, then replace the original value with T_med. Setting a 3×3 median filter can effectively remove isolated defective pixels without disrupting the overall temperature field distribution; the temperature difference between adjacent pixels in a normal room will not exceed 2℃, otherwise it will be judged as a defective pixel.

[0098] Time-based low-pass smoothing (coefficient α=0.3): T_filt(i,j,t)=0.3×T_ir(i,j,t)+0.7×T_filt(i,j,t-1). Too small an α (e.g., <0.2) will result in a delayed temperature response, failing to capture temperature changes in a timely manner; too large an α (e.g., >0.4) will lead to insufficient smoothing effect, failing to suppress instantaneous fluctuations; α=0.3 balances real-time performance and smoothness.

[0099] Preprocessing before calculating the average room temperature ensures the reliability of the temperature data. For example, for the human body area: 30℃≤T_filt(i,j,t)≤36℃, weight w=1.2; for other areas, weight w=1.0; the average room temperature T_room(t)=[the sum of all points w(i,j,t)×T_filt(i,j,t)]÷[the sum of all points w(i,j,t)].

[0100] The processing unit 102 is also configured to predict the actual cooling capacity of the air conditioner using a preset artificial intelligence model.

[0101] In some implementations, the artificial intelligence model is a hybrid network model that combines convolutional neural networks and long short-term memory networks.

[0102] Convolutional neural networks excel at processing multi-dimensional structured data, efficiently extracting key features that affect cooling capacity; long short-term memory networks excel at processing time-series data, learning the dynamic patterns of air conditioner operating status changes over time. When these two are combined, the model can capture both the spatial correlation between multiple parameters and cooling capacity, and the temporal correlation between changes in operating conditions and cooling capacity, thus more accurately predicting the actual cooling capacity of the air conditioner under different operating conditions. Figure 3 The diagram illustrates the hybrid network model. First, the input layer receives multi-dimensional parameters of the air conditioner's operation and the environment. The two-layer convolutional neural network (CNN) of the feature extraction layer extracts the coupled features of the environment and load parameters. Then, the processed temporal features are fed into the long short-term memory network (LSTM) of the temporal processing layer to capture the slow changing trends of parameters such as temperature and humidity. Finally, the output layer outputs the prediction results.

[0103] The model's inputs include environmental and operational parameters, and its outputs are cooling / heating capacity (Q_cool / Q_heat). Environmental parameters include: indoor equipment area temperature (T_eq), with a sampling accuracy of ±0.1℃ and a physical range of 0~45℃, collected by distributed temperature probes avoiding equipment heat dissipation vents; air conditioner indoor air inlet temperature (T_in), with a sampling accuracy of ±0.1℃ and a physical range of 0~45℃, collected by the air conditioner's built-in temperature sensor; outdoor ambient temperature (T_out), with a sampling accuracy of ±0.2℃ and a physical range of -10~48℃, collected by an outdoor temperature sensor; and indoor relative humidity (RH_in), with a sampling accuracy of ±3%RH and a physical range of 0%~100%RH, collected by a humidity sensor.

[0104] Operating parameters include the current compressor frequency (f_comp), with a sampling accuracy of ±1Hz and a physical range of 30~120Hz for inverter air conditioners, read through communication with the air conditioner controller; the current indoor fan speed (v_fan_in), with a sampling accuracy of ±10rpm and a physical range of 0~2000rpm, collected by the fan speed sensor; the current outdoor fan speed (v_fan_out), with a sampling accuracy of ±10rpm and a physical range of 0~2000rpm, collected by the fan speed sensor; and the current expansion valve opening (K_valve), with a sampling accuracy of ±1% and a physical range of 20%~80% of the mechanical limit, read through communication with the expansion valve controller.

[0105] The model training process involves collecting at least 10,000 sets of historical steady-state data covering the entire operating range, and removing invalid data such as start-stop transition states and sensor malfunctions to ensure data representativeness and purity. In the data preprocessing stage, Z-score normalization is used to convert each feature into a standard distribution with a mean of 0 and a variance of 1 to eliminate dimensional differences between different parameters. Then, Pearson correlation analysis is used to remove redundant features; for example, if the correlation between the indoor air inlet temperature and the indoor equipment area temperature is greater than 0.9, one of them can be retained to reduce the model's computational cost.

[0106] The model training uses the mean squared error (MSE), which performs well in regression prediction, as the loss function, and selects the Adam optimizer. The initial learning rate is set to 0.001, and the decay rate is set to 0.0001 to achieve fast and stable convergence and avoid training oscillations and overfitting risks. The maximum number of training epochs is set to 200. Training is stopped when the validation set error does not exceed 3% to ensure that the model accuracy meets engineering requirements. To further improve the model's generalization ability, a Dropout layer is added to the network with a dropout rate of 0.2. During training, some neurons are randomly deactivated to prevent the model from over-relying on training data, thereby solving the problems of overfitting and poor generalization ability, and ensuring that the model can accurately predict cooling capacity under different operating conditions.

[0107] The collection of 10,000 sets of data aims to cover all operating conditions with different ambient temperatures, humidity levels, and air conditioning operating parameters, laying the foundation for the model's generalization ability. Eliminating transitional and fault data prevents invalid data from interfering with model accuracy. Z-score normalization and correlation analysis optimize model input from two dimensions: data standardization and feature simplification, improving training efficiency. A reasonable loss function, optimizer configuration, and Dropout layer design collectively ensure the model's convergence speed, prediction accuracy, and long-term stability.

[0108] The processing unit 102 is also configured to calculate the room heat load based on the average room temperature, the actual cooling capacity, and the preset room heat capacity.

[0109] The room's heat capacity determines how quickly the indoor temperature changes with heat, the room's average temperature reflects the current overall temperature level, and the air conditioner's actual cooling capacity represents its current cooling output capability. By combining these three parameters, the total heat generated in real time indoors can be accurately calculated, i.e., the room's heat load.

[0110] The formula for estimating room heat capacity is: C_room = 1.2 × air density × room volume × air specific heat × 1.2. Where, air density is taken as a fixed value of 1.2 kg / m³ under standard operating conditions. 3The specific heat of air is taken as a fixed value of °C under standard operating conditions. Using these fixed values ​​simplifies the calculation process while meeting engineering accuracy requirements. The two coefficients 1.2 in the formula serve different purposes: the first 1.2 is a wall heat capacity correction coefficient, considering that wall heat storage accounts for approximately 20% of the total room heat storage; this coefficient corrects for wall heat storage and incorporates it into the overall estimate. The second 1.2 is a safety factor, used to avoid estimation deviations caused by differences in room structure. Taking a room with dimensions of 5m × 4m × 3m as an example, substituting into the formula yields a room heat capacity of approximately 86.83 kJ / °C. In practical applications, the main control module can automatically complete the calculation by inputting the room's length, width, and height parameters, thus solving the problem of inaccurate room heat capacity acquisition and providing reliable basic parameters for subsequent heat load calculations.

[0111] The calculation of room heat load is based on the law of conservation of energy, that is, the heat corresponding to the change in room temperature is equal to the difference between the room heat load and the air conditioning cooling capacity. The energy balance equation is: C_room×(T_room(t)-T_room(t-1)) / Δt=Q_load(t)-Q_cool(t). Considering the time interval Δt=1s in the control sequence, the formula can be simplified to a direct calculation of the real-time load: Q_load(t)=Q_cool(t)+C_room×(T_room(t)-T_room(t-1)). In this formula, the unit of Q_load(t) is kW. This simplified form avoids complex differential calculations, effectively reducing the system's computing power requirements. In actual operation, the main control module can read the current average room temperature T_room and the actual air conditioning cooling capacity Q_cool data in real time, substitute them into the formula to complete the calculation, and solve the problem of inaccurate real-time heat load detection in traditional control.

[0112] In some embodiments, the processing unit 102 is further configured to: acquire the room heat load at multiple consecutive times before determining the target cooling capacity of the air conditioner based on the room heat load; and average the room heat load at the multiple consecutive times to obtain a smoothed room heat load.

[0113] In actual operation, factors such as indoor temperature and equipment operating status can fluctuate instantaneously, leading to significant random fluctuations in the real-time calculated room heat load. By averaging the room heat load over multiple consecutive moments before determining the target cooling capacity, random noise and instantaneous fluctuations in the data can be effectively filtered out, resulting in a more stable and representative heat load value. This reduces the need for frequent adjustments to the air conditioning system's operating status, resulting in more stable cooling output. This improves the accuracy and comfort of indoor temperature control and avoids additional wear and tear on air conditioning components caused by frequent adjustments.

[0114] A 5-point moving average algorithm can be used to smooth the heat load data. The calculation formula is: Q_load_smooth(t) = [Q_load(t) + Q_load(t-1) + Q_load(t-2) + Q_load(t-3) + Q_load(t-4)] / 5. By caching the continuous heat load data of the current moment and the previous four moments (a total of 5 seconds), the arithmetic mean is calculated to obtain the smoothed heat load value Q_load_smooth. The time window corresponding to the 5 sampling points can effectively filter out heat load abrupt changes and instantaneous noise caused by factors such as human movement and equipment start-up and shutdown. At the same time, the window length is short, which will not cause obvious response lag, thus balancing the data smoothing effect and the dynamic response speed of the system.

[0115] The processing unit 102 is also configured to determine the target cooling capacity of the air conditioner based on the room heat load and the room average temperature.

[0116] The system combines the total indoor heat production with the current overall temperature to determine the optimal cooling output target for the current operating conditions. When the indoor temperature is close to the ideal state, the target cooling capacity is mainly set to match the room's heat load to ensure a balance between heat supply and demand. When the indoor temperature deviates from the ideal state, the target cooling capacity will be adjusted synchronously according to the temperature deviation, taking into account both temperature control and heat balance requirements, so that the air conditioner's cooling output matches the actual indoor demand.

[0117] In some embodiments, the processing unit 102 determines the target cooling capacity of the air conditioner based on the room heat load and the room average temperature, including: determining the magnitude of the absolute value of the difference between the average temperature and the set temperature; if the absolute value of the difference is less than or equal to a preset difference, then the room heat load is used as the target cooling capacity of the air conditioner; if the absolute value of the difference is greater than the preset difference, then the target cooling capacity of the air conditioner is calculated based on the room heat load using a preset PID control method.

[0118] Specifically, the system acquires the current average room temperature in real time and compares it with the user-set temperature. It calculates the absolute value of the difference and then compares this absolute value with a preset difference to determine the control status of the indoor temperature. When the absolute value of the difference between the average room temperature and the set temperature is less than or equal to the preset difference, it indicates that the indoor temperature is close to the target state. In this case, the air conditioner does not need to significantly adjust its cooling output; the system directly uses the room heat load as the target cooling capacity of the air conditioner, ensuring that the cooling output of the air conditioner is roughly equal to the amount of heat generated indoors, thus maintaining a stable indoor temperature. When the absolute value of the difference between the average room temperature and the set temperature is greater than the preset difference, it indicates that the indoor temperature deviates significantly from the target state. In this case, it is necessary to dynamically adjust the cooling output to quickly bring the temperature closer to the target value. The system calculates the target cooling capacity of the air conditioner based on the room heat load using a preset PID control method, and dynamically adjusts the cooling output in conjunction with the magnitude and trend of the temperature deviation, balancing temperature adjustment speed and control stability.

[0119] For example, the difference e(t) between the average room temperature T_room and the user-set temperature T_set can be calculated using the formula e(t) = T_room(t) - T_set.

[0120] A steady state is defined as when |e(t)| ≤ 0.1℃ and this state lasts for 30 seconds. Under steady-state conditions, the target cooling capacity equals the smoothed room heat load, i.e., Q_target(t) = Q_load_smooth(t). Setting a deviation threshold of 0.1℃ meets the requirements for high-precision temperature control; the 30-second duration requirement avoids misjudging the steady state due to instantaneous temperature fluctuations. Using a direct balance between cooling capacity and heat load under steady-state conditions maintains stable room temperature while avoiding redundant or insufficient cooling capacity, thus achieving energy savings.

[0121] Under unsteady-state conditions, the system uses a PID control algorithm to dynamically calculate the target cooling capacity. The formula is: Q_target(t)=Q_load_smooth(t)+Kp×e(t)+Ki×(e(0)+e(1)+…+e(t))×Δt+Kd×(e(t)-e(t-1)) / Δt. This is suitable for a 1.5 HP air conditioner, compatible with 50~70m². 2For a room as an example, the recommended PID parameters are: Kp = 0.8 kW / ℃, Ki = 0.05 kW / (℃·s), Kd = 0.1 kW·s / ℃. To protect the equipment, the target cooling capacity must be limited to the air conditioner's operating range, i.e., Q_min ≤ Q_target(t) ≤ Q_max, such as the common range of 0.3~3.5 kW. The proportional coefficient Kp is used for rapid response to temperature deviations to avoid cumulative deviations; the integral coefficient Ki is used to eliminate long-term temperature control offsets, and a smaller value can prevent integral saturation; the derivative coefficient Kd is used to predict the trend of deviation changes, suppress temperature overshoot, and avoid excessive cooling capacity adjustment. By programming the PID algorithm and judging the cooling capacity limit, the problems of low temperature control accuracy and easy overshoot under dynamic operating conditions can be effectively solved, while protecting the air conditioner compressor from damage caused by operating outside its range.

[0122] The control unit 103 is configured to output control commands for the air conditioner based on the target cooling capacity, so as to adjust the operating state of the air conditioner.

[0123] The system converts the calculated target cooling capacity into operating commands that can be recognized by each component of the air conditioner. These commands adjust operating parameters such as compressor operating frequency, fan speed, and throttling device opening, thereby changing the actual cooling output of the air conditioner. After completing the state adjustment, the air conditioner continues to feed back operating data to the front-end process. The entire control process is executed cyclically to achieve continuous regulation of the indoor temperature.

[0124] In some implementations, the control unit 103 outputs control commands for the air conditioner based on the target cooling capacity, including: limiting the adjustment range of the target cooling capacity according to a preset rate of change; and outputting corresponding control commands based on the limited target cooling capacity.

[0125] Specifically, firstly, after determining the target cooling capacity of the air conditioner, the system does not directly convert this value into a control command. Instead, it limits the adjustment range of the target cooling capacity according to a preset rate of change. For example, the system compares the current target cooling capacity with the actual cooling output value of the air conditioner at the previous moment, calculates the difference between the two, and ensures that the change range of cooling capacity per unit time does not exceed the preset maximum value. If the change range of the target cooling capacity exceeds the preset rate of change, the system will adjust it to the maximum allowable change range to obtain the limited target cooling capacity value; if the change range is within the preset range, the target cooling capacity value is directly retained.

[0126] Secondly, the system converts the limited target cooling capacity into operating commands that can be recognized by each component of the air conditioner, thereby adjusting operating parameters such as compressor operating frequency, fan speed, and throttling device opening, and changing the actual cooling output of the air conditioner.

[0127] For example, setting the maximum rate of change of cooling capacity to ΔQ_rate = 0.3kW / s is suitable for the adjustment capability of a 1.5HP inverter air conditioner compressor, preventing equipment damage and temperature overshoot caused by excessively rapid frequency adjustments. The formula for calculating the actual cooling capacity is: Q_apply(t) = Q_apply(t-1) + limit(Q_target(t) - Q_apply(t-1), -0.3, +0.3). In this formula, the limit operation restricts the adjustment range of cooling capacity per unit time to within ±0.3kW / s. Setting the initial cooling capacity of the system to the minimum cooling capacity Q_min of the air conditioner avoids a sudden drop in room temperature due to excessive cooling capacity at startup.

[0128] In actual operation, the core components of an air conditioner, such as the compressor and fan, cannot withstand drastic, instantaneous changes in cooling capacity. If the target cooling capacity is directly converted into a control command, significant fluctuations in the target cooling capacity will cause drastic changes in compressor frequency and fan speed within a short period. This not only easily leads to mechanical wear on equipment components but also causes drastic fluctuations in air conditioning system pressure and temperature, affecting the stability and lifespan of the equipment. By limiting the adjustment range of the target cooling capacity through preset change rates, the cooling output of the air conditioner can be made more stable, avoiding additional wear on components due to frequent and drastic adjustments, extending the equipment's lifespan, and ensuring gradual changes in indoor temperature, thus improving user comfort.

[0129] This solution relies on artificial intelligence technology and the principle of thermal balance to accurately calculate cooling capacity and room heat load. It can dynamically match the air conditioner's cooling output according to the indoor heat and temperature status, effectively avoiding data deviation caused by single temperature measurement. This allows the air conditioner's cooling output to be accurately matched with the actual indoor heat demand, which not only improves the control accuracy of indoor temperature and reduces the temperature fluctuation range, but also reduces the ineffective cooling work of the air conditioner and reduces the energy consumption of the equipment.

[0130] Since the processing and functions implemented by the device in this embodiment are basically the same as the embodiments, principles and examples of the aforementioned methods, any details not covered in the description of this embodiment can be found in the relevant descriptions in the aforementioned embodiments, and will not be repeated here.

[0131] By adopting the technical solution of this invention, closed-loop temperature control is achieved through multi-point temperature measurement, cooling capacity prediction, heat load calculation and dynamic cooling capacity adjustment. This enables precise matching of load and cooling capacity, improves temperature control accuracy, reduces temperature fluctuations, and lowers operating energy consumption.

[0132] According to an embodiment of the present invention, an air conditioner corresponding to an air conditioner control device is also provided. This air conditioner may include the air conditioner control device described above.

[0133] Since the processing and functions implemented by the air conditioner in this embodiment are basically the same as the embodiments, principles and examples of the aforementioned device, any details not covered in the description of this embodiment can be found in the relevant descriptions in the aforementioned embodiments, and will not be repeated here.

[0134] By employing the technical solution of this invention, a closed-loop control logic of multi-point temperature acquisition, cooling capacity prediction, heat load calculation and dynamic adjustment is used to achieve precise matching between cooling capacity and room load, effectively improving temperature control accuracy, reducing temperature fluctuations, and lowering equipment energy consumption.

[0135] According to an embodiment of the present invention, a storage medium corresponding to an air conditioner control method is also provided, the storage medium including a stored program, wherein the program controls the device where the storage medium is located to execute the air conditioner control method described above when it is executed.

[0136] Since the processing and functions implemented by the storage medium in this embodiment are basically the same as the embodiments, principles and examples of the aforementioned methods, any details not covered in this embodiment can be found in the relevant descriptions in the aforementioned embodiments, and will not be repeated here.

[0137] By employing the technical solution of this invention, a closed-loop control logic of multi-point temperature acquisition, cooling capacity prediction, heat load calculation and dynamic adjustment is used to achieve precise matching between cooling capacity and room load, effectively improving temperature control accuracy, reducing temperature fluctuations, and lowering equipment energy consumption.

[0138] According to an embodiment of the present invention, a computer program product corresponding to the control method for an air conditioner is also provided. The computer program product includes a computer program that, when processed and executed, implements the steps of the control method for the air conditioner described above.

[0139] Since the processing and functions implemented by the computer program product in this embodiment are basically corresponding to the embodiments, principles and examples of the aforementioned methods, any details not covered in the description of this embodiment can be found in the relevant descriptions in the aforementioned embodiments, and will not be repeated here.

[0140] By employing the technical solution of this invention, a closed-loop control logic of multi-point temperature acquisition, cooling capacity prediction, heat load calculation and dynamic adjustment is used to achieve precise matching between cooling capacity and room load, effectively improving temperature control accuracy, reducing temperature fluctuations, and lowering equipment energy consumption.

[0141] In summary, it is readily understood by those skilled in the art that, without conflict, the aforementioned advantageous methods can be freely combined and superimposed.

[0142] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for controlling an air conditioner, characterized in that, The method includes: Obtain the temperature at at least two locations within the room; The average room temperature is determined based on the temperatures at at least two locations. Predict the actual cooling capacity of the air conditioner using a pre-set artificial intelligence model; The room heat load is calculated based on the average room temperature, the actual cooling capacity, and the preset room heat capacity. The target cooling capacity of the air conditioner is determined based on the room heat load and the room average temperature. Based on the target cooling capacity, control commands are output to the air conditioner to adjust its operating status.

2. The air conditioning control method according to claim 1, characterized in that, The temperatures at at least two locations were collected using an infrared array sensor. Determining the average room temperature based on the temperatures at the at least two locations includes: Detect the location of people in the room, and determine the temperature at the location of the people from the temperatures at at least two locations; The average room temperature is calculated by weighting the temperatures at at least two locations; wherein the temperature at the location where a person is located has a greater weight than the temperature at other locations.

3. The air conditioning control method according to claim 1, characterized in that, The artificial intelligence model is a hybrid network model that combines convolutional neural networks and long short-term memory networks.

4. The air conditioning control method according to claim 1, characterized in that, The method further includes: Before determining the target cooling capacity of the air conditioner based on the room heat load, the room heat load is obtained at multiple consecutive time points; The room heat load at multiple consecutive time points is averaged to obtain a smoothed room heat load.

5. The air conditioning control method according to claim 1 or 4, characterized in that, Determining the target cooling capacity of the air conditioner based on the room heat load and the room average temperature includes: Determine the absolute value of the difference between the average temperature and the set temperature; If the absolute value of the difference is less than or equal to the preset difference, then the room heat load is taken as the target cooling capacity of the air conditioner. If the absolute value of the difference is greater than the preset difference, the target cooling capacity of the air conditioner is calculated based on the room heat load using a preset PID control method.

6. The air conditioning control method according to claim 1, characterized in that, Based on the target cooling capacity, control commands for the air conditioner are output, including: The adjustment range of the target cooling capacity is limited according to a preset rate of change; The corresponding control command is output based on the target cooling capacity after the limit is set.

7. A control device for an air conditioner, characterized in that, include: The acquisition unit is configured to acquire the temperature at at least two locations within the room; The processing unit is configured to determine the average room temperature based on the temperatures at the at least two locations; The processing unit is also configured to predict the actual cooling capacity of the air conditioner using a preset artificial intelligence model. The processing unit is also configured to calculate the room heat load based on the average room temperature, the actual cooling capacity, and the preset room heat capacity. The processing unit is also configured to determine the target cooling capacity of the air conditioner based on the room heat load and the room average temperature. The control unit is configured to output control commands for the air conditioner based on the target cooling capacity, so as to adjust the operating status of the air conditioner.

8. An air conditioner, characterized in that, include: The air conditioning control device as described in claim 7.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the storage medium to perform the air conditioning control method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the air conditioning control method according to any one of claims 1 to 6.

Images