Intelligent air conditioner cooling method and system

Abstract

The application discloses an intelligent air conditioner cooling method and system, and belongs to the technical field of air conditioning. In view of the problems of the existing whole-house centralized air conditioner temperature control without partition adaptation capability and the disorderly diffusion of cold quantity, the application provides an air conditioner cooling control method, which comprises the following steps: collecting the air outlet wind speed of a cabinet, the space natural wind speed and the required temperature value of the heat exchange temperature difference, transmitting the collected data to a working condition recognition unit, so as to determine the working condition mode of the current cabinet according to the wind speed difference and the heat exchange temperature difference; obtaining the real-time active power and temperature data of the cabinet according to the working condition mode, calculating the real-time heat load of a single cabinet, dividing different heat emission partitions according to the heat load and outputting the cold quantity distribution result of each partition; and controlling the temperature control execution unit and the double-cavity ceiling cold-heat isolation unit to supply cold to each partition according to the cold quantity distribution result. Through working condition recognition, dynamic partition cooling and double-cavity ceiling cold-heat isolation, the application realizes the on-demand directional delivery of cold quantity, and improves the cold quantity utilization rate.

Description

Technical Field

[0001] This application belongs to the field of air conditioning technology, specifically relating to an intelligent air conditioning cooling method and system. Background Technology

[0002] Currently, two types of core server racks are commonly deployed in equipment rooms, data centers, and server rooms: actively exhaust racks and airtight racks. Actively exhaust racks have built-in cooling fans that actively exhaust air from the top or rear; airtight racks use a fully sealed structure with no active exhaust mechanism, relying solely on natural convection for heat dissipation. For temperature control in these scenarios, the industry mainstream adopts centralized whole-room air conditioning, which uses cabinet or ceiling-mounted central air conditioning units to uniformly control the temperature of the entire equipment space. This method is simple and direct, eliminating the need for separate control of different racks, and can maintain a relatively stable overall temperature in the server room to a certain extent.

[0003] However, centralized whole-house air conditioning temperature control has significant technical flaws. Lacking any zone adaptation capability, it cannot precisely adjust cooling based on the real-time heat load differences of different areas and server racks. In actual operation, each server rack operates at different states; some racks are under high load and generate a large amount of heat, while others are under low load or in standby mode and generate less heat. Centralized air conditioning can only output a uniform amount of cooling, leading to disordered diffusion of cooling capacity. A large amount of cooling capacity is lost due to mixing with hot airflow, resulting in significant waste of cooling capacity, high energy consumption, and an imbalance problem of "localized rack overheating and localized overcooling." Overheated areas may cause heat buildup in equipment, increasing the risk of downtime; overcooled areas waste energy. Furthermore, the airflow method of centralized air conditioning cannot achieve directional cooling, causing cooling capacity to diffuse to non-heat-generating areas, further reducing cooling capacity utilization and failing to meet the needs of precise temperature control for server racks. Therefore, there is an urgent need for an air conditioning cooling system that can precisely provide cooling to different zones based on the actual heat dissipation needs of the server racks. Summary of the Invention

[0004] The purpose of this application is to provide an intelligent air conditioning cooling method and system that can achieve precise zoned cooling and improve the utilization rate of cooling capacity.

[0005] This application provides an intelligent air conditioning cooling method applicable to air conditioning cooling control systems. The air conditioning cooling control system includes a refrigeration unit, a dual-cavity ceiling-mounted cold and heat isolation unit, a main control unit, a temperature control execution unit, and a working condition identification unit. The refrigeration unit serves as the cold source, and the dual-cavity ceiling-mounted cold and heat isolation unit serves as the cold and hot airflow isolation carrier. The dual-cavity ceiling-mounted cold and heat isolation unit includes a supply air positive pressure static pressure chamber, an exhaust air negative pressure heat collection chamber, a supply air pressure regulating fan, and an exhaust air pressure regulating fan. The main control unit serves as the decision-making and control subject. The temperature control execution unit provides cooling, and the temperature control execution unit includes a directional cooling... Air nozzles and Coanda attached nozzles; operating condition determination is performed using an operating condition identification unit; the supply air pressure regulating fan is connected to the supply air positive pressure static pressure chamber, the exhaust air pressure regulating fan is connected to the exhaust air negative pressure heat collection chamber, the supply air pressure regulating fan is connected to the outdoor refrigerant pipeline, and the exhaust air pressure regulating fan is connected to the outdoor exhaust pipeline; both the directional air supply nozzles and the Coanda attached nozzles are located on the lower panel of the supply air positive pressure static pressure chamber; the refrigeration unit, the dual-chamber ceiling heat and cold insulation unit, the temperature control execution unit, and the operating condition identification unit are respectively connected to the main control unit; the air conditioning cooling control method includes: Collect the operating condition data of the cabinet; transmit the operating condition data to the operating condition identification unit; the operating condition data includes the air outlet wind speed, the natural wind speed in the space, and the temperature value required for heat exchange temperature difference; The first operating mode of the current cabinet is determined based on the wind speed difference and the heat exchange temperature difference, so that the first operating mode is transmitted to the main control unit; the first operating mode includes an active air outlet mode and a no air outlet mode; the wind speed difference is the difference between the air outlet wind speed of the cabinet and the natural wind speed in the space. The real-time active power and temperature data of the cabinet are obtained according to the first working mode, the real-time heat load of a single cabinet is calculated, and different heat-generating zones are dynamically divided according to the real-time heat load of the single cabinet and the cooling capacity allocation results of each zone are output. The temperature control unit and the dual-cavity ceiling heat and cold isolation unit provide cooling to each zone according to the cooling capacity allocation results.

[0006] In this embodiment, a dual-cavity ceiling-mounted thermal isolation unit is used to achieve physical isolation of hot and cold airflows. A condition identification unit distinguishes between active air-discharge cabinets and closed cabinets without airflow. A main control unit dynamically divides heat-generating zones and allocates cooling capacity based on the real-time heat load of each cabinet. A temperature control execution unit performs condition-specific air delivery. Compared to existing whole-house centralized air conditioning systems with unified air delivery and disordered cold air diffusion, this application, through the synergistic effect of thermal isolation, condition identification, and dynamic zoned cooling, enables the cooling capacity to be directionally delivered to different heat-generating zones as needed. This fundamentally avoids the mixing and loss of cold and hot airflows, thus improving cooling capacity utilization.

[0007] Further, based on the first operating mode, the real-time active power and temperature data of the rack are obtained, the real-time heat load of a single rack is calculated, and different heat-generating zones are dynamically divided according to the real-time heat load of the single rack, and the cooling capacity allocation results of each zone are output, including: When the first operating mode is active air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the heat dissipation efficiency coefficient by the real-time active power of the cabinet; when the first operating mode is no air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the convective heat transfer coefficient by the effective heat dissipation area of ​​the cabinet and then by the difference between the center temperature of the main heat dissipation surface of the cabinet and the ambient reference temperature. Based on the ratio of the real-time heat load of a single cabinet to the rated maximum heat load of the cabinet, the cabinet is dynamically divided into high-heat zone, medium-heat zone and low-heat zone. The cooling capacity allocation results for each zone are output according to the heat load ratio of each zone.

[0008] In this embodiment, different heat load calculation formulas are used for active air supply mode and no air supply mode: in active air supply mode, the calculation is based on active power and heat dissipation efficiency coefficient; in no air supply mode, the calculation is based on convection heat transfer, effective heat dissipation area, and temperature difference. The cabinet is dynamically divided into high, medium, and low heat generation zones according to the ratio of the heat load to the rated maximum heat load, and then the cooling capacity is allocated proportionally. Compared with existing technologies that use fixed power thresholds or uniform cooling, this application uses heat load calculation driven by the physical mechanism of different modes and proportional zoned cooling to accurately match the cooling capacity with the actual heat generation of the cabinet, avoiding over-cooling or under-cooling.

[0009] Furthermore, the operating condition identification unit determines the first operating condition mode of the current cabinet based on the wind speed difference and heat exchange temperature difference, including: Obtain the second operating mode; the second operating mode is the current first operating mode; When the second operating mode is the no-airflow mode, the difference between the rack outlet air temperature and the rack inlet air temperature is used as the heat exchange temperature difference; if the wind speed difference is greater than or equal to the first wind speed threshold, the heat exchange temperature difference is greater than or equal to the first temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the first preset number of cycles, then the second operating mode is determined to be the active airflow mode; otherwise, the first operating mode remains unchanged. When the second operating mode is the active air outlet mode, the difference between the cabinet surface temperature and the ambient reference temperature is used as the heat exchange temperature difference; if the wind speed difference is less than or equal to the second wind speed threshold, the heat exchange temperature difference is less than or equal to the second temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the second preset number of cycles, then the second operating mode is determined to be the no air outlet mode; otherwise, the first operating mode remains unchanged. Wherein, the first wind speed threshold is greater than the second wind speed threshold, and the first temperature difference threshold is greater than the second temperature difference threshold.

[0010] In this embodiment, after obtaining the second operating mode, the wind speed difference is calculated, and the corresponding temperature difference calculation method is selected according to the current mode—in the no-airflow mode, the difference between the rack outlet temperature and the inlet temperature is used as the heat exchange temperature difference; in the active airflow mode, the difference between the rack surface temperature and the ambient reference temperature is used as the heat exchange temperature difference. Furthermore, a hysteresis comparison logic is adopted, where the first wind speed threshold is greater than the second wind speed threshold and the first temperature difference threshold is greater than the second temperature difference threshold. The mode is switched only after the conditions are maintained continuously for a preset number of cycles. Compared with the existing technology that uses a single threshold for instantaneous determination and is prone to frequent mode switching due to short-term fluctuations in wind speed or temperature, this application significantly improves the stability and anti-interference capability of operating mode identification through a composite logic of mode anchoring, asymmetric threshold hysteresis, and delayed confirmation.

[0011] Furthermore, the temperature control unit and the dual-cavity ceiling thermal insulation unit provide cooling to each zone according to the cooling capacity allocation results, including: The dual-cavity ceiling heat and cold isolation unit controls the distribution of cold air supplied by the cold source to the top of each zone cabinet through the positive pressure static pressure chamber. When in active air supply mode, the directional air supply nozzle of the temperature control unit is opened to deliver cold air to the air inlet of the active air supply cabinet. When in no-airflow mode, the Coanda attachment nozzle of the temperature control execution unit is opened to blow cold air onto the surface of the airless sealed cabinet; the inner side of the Coanda attachment nozzle is provided with an arc-shaped wall-mounting guide plate to allow the cold air to adhere to the surface of the cabinet and form an attachment laminar flow cold film. The dual-cavity ceiling heat and cold isolation unit actively captures and discharges the hot airflow generated by each cabinet through the exhaust negative pressure heat collection cavity.

[0012] In this embodiment, cold air is evenly distributed above each zone of the server racks via a positive pressure static pressure chamber. The corresponding nozzle is selected according to the first operating mode: in active airflow mode, the directional airflow nozzle delivers cold air to the rack's air inlet; in no-airflow mode, the Coanda attached nozzle blows cold air onto the rack surface. The exhaust negative pressure heat capture chamber actively captures the exhausted hot airflow. Compared to existing technologies with single-outlet uniform airflow and disordered mixing of hot and cold air, this application, through a physical isolation structure of lower-level airflow and upper-level heat exhaust, combined with operating mode-specific nozzle airflow, forms a complete directional airflow path from the cold source to the rack and then to the heat exhaust, eliminating the mixing of hot and cold air.

[0013] Furthermore, after cooling each zone, the process also includes: Collect real-time temperature inside the cabinet; The real-time temperature is compared with the target temperature to obtain the temperature deviation; When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments according to the correction signal.

[0014] In this embodiment, after cooling each zone, the real-time temperature inside the cabinet is collected and compared with the target temperature. When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit, enabling the main control unit to make corresponding adjustments based on the correction signal. Compared with existing technologies that rely solely on ambient temperature or outlet air temperature for open-loop or semi-open-loop control, this application directly uses the cabinet's internal temperature as feedback, forming a closed-loop control chain of "cooling—collection—comparison—correction." This allows the main control unit to perceive changes in the thermal state inside the cabinet in real time and proactively intervene to adjust when the deviation exceeds the allowable range. This prevents the cabinet temperature from deviating from the target value due to sudden load changes, environmental disturbances, or other factors, thereby improving the accuracy of temperature control and the system's responsiveness, and providing a triggering prerequisite for subsequent graded corrections.

[0015] Furthermore, when the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments according to the correction signal, including: When the absolute value of the temperature deviation is greater than the first temperature threshold and not greater than the second temperature threshold, the nozzle air volume corresponding to the first working mode is adjusted by the first adjustment step; wherein, when the first working mode is the active air outlet mode, the directional air outlet is adjusted, and when the first working mode is the no air outlet mode, the Coanda attached nozzle is adjusted. When the absolute value of the temperature deviation is greater than the second temperature threshold, the PID control parameters, the output power of the refrigeration unit, and the wind speed of the Coanda attached nozzle are adjusted according to the second adjustment step size. Wherein, the first adjustment step size is smaller than the second adjustment step size.

[0016] In this embodiment, when the absolute value of the temperature deviation is greater than a first temperature threshold but not greater than a second temperature threshold, only the nozzle airflow is adjusted. When the absolute value of the deviation is greater than the second temperature threshold, the PID parameters, the power of the cooling unit, and the Coanda nozzle airflow are adjusted simultaneously, with the first adjustment step size being smaller than the second adjustment step size. Compared with the existing single fixed gain adjustment scheme, this application performs correction actions of different magnitudes according to the degree of deviation, with fine-tuning for small deviations and coarse-tuning for large deviations, avoiding temperature control oscillations and improving the temperature control stability and response rationality of the system.

[0017] Furthermore, after outputting the cooling capacity allocation results for each zone, it also includes: Historical zone heat load data from multiple past control cycles is obtained, and the historical zone heat load data is input into a time-series prediction model to predict the zone heat load at future times. When the predicted value is higher than the preset ratio of the current zone heat load, the upper limit of the reference output of the refrigeration unit is pre-adjusted to be higher. When the predicted value is lower than the preset ratio of the current zone heat load, the baseline output lower limit of the refrigeration unit will be lowered.

[0018] In this embodiment, after outputting the cooling capacity allocation result, historical zone heat load data from multiple control cycles is acquired. A time-series prediction model is used to predict the zone heat load at future times. Based on the predicted value relative to the current load, the upper or lower limit of the refrigeration unit's baseline output is adjusted in advance. Compared to existing technologies that only respond retrospectively based on the current load, this application implements a pre-cooling strategy through heat load prediction, adjusting the refrigeration unit's output in advance and avoiding temperature control delays caused by lag in the refrigeration system's response.

[0019] Furthermore, the time-series prediction model is an LSTM model; before acquiring historical zonal heat load data from multiple past control cycles and predicting future zonal heat loads using the time-series prediction model, the following steps are also included: Collect historical heat load data for multiple consecutive days in different zones and divide it into training and validation sets according to a preset ratio; The historical zone heat load data in the training set are input into the LSTM model in time sequence to obtain the heat load prediction value; The LSTM model is trained using the mean square error between the predicted heat load and the actual collected zone heat load data as the loss function, and the Adam optimization algorithm is employed.

[0020] In this embodiment, the time-series prediction model employs an LSTM model. Historical heat load data from multiple consecutive days is collected and divided into training and validation sets according to a preset ratio. The mean square error between the predicted and actual heat load values ​​is used as the loss function, and the Adam optimization algorithm is employed to train the model. During the operational phase, the latest data is periodically collected for fine-tuning. Compared to existing technologies that use fixed-parameter models or manually set thresholds, this application, through data-driven LSTM model training and periodic fine-tuning updates, enables the prediction model to adapt to the long-term changing characteristics of the data center load, maintaining prediction accuracy.

[0021] Furthermore, the aforementioned air conditioning cooling control method further includes: Obtain the actual static pressure of the positive pressure static pressure chamber for air supply, and calculate the deviation between the actual static pressure and the target static pressure for air supply; When the absolute value of the deviation is greater than the first preset threshold, the speed of the air supply pressure regulating fan is adjusted to reduce the deviation; Obtain the actual negative pressure of the exhaust negative pressure heat collection chamber, and calculate the deviation between the actual negative pressure and the target negative pressure of the exhaust. When the absolute value of the deviation is greater than the second preset threshold, the speed of the exhaust pressure regulating fan is adjusted to reduce the deviation.

[0022] In this embodiment, the actual static pressure of the positive pressure static pressure chamber is acquired and its deviation from the target static pressure is calculated. When the deviation exceeds a threshold, the speed of the supply air pressure regulating fan is adjusted to reduce the deviation. Simultaneously, the actual negative pressure of the negative pressure heat capture chamber is acquired and its deviation from the target negative pressure is calculated. When the deviation exceeds a threshold, the speed of the exhaust air pressure regulating fan is adjusted to reduce the deviation. Compared with existing static pressure chambers or negative pressure chambers without pressure closed-loop control, this application uses independent pressure closed-loop control to compensate for pressure fluctuations caused by changes in air resistance, filter blockage, etc., in real time, ensuring the uniformity of cooling in the positive pressure chamber and the heat plume capture effect in the negative pressure chamber.

[0023] Another embodiment of this application provides a cooling control system, including: Refrigeration unit, dual-cavity ceiling heat and cold insulation unit, main control unit, temperature control execution unit and operating condition identification unit; The dual-cavity ceiling thermal insulation unit includes a positive pressure static pressure chamber for air supply, a negative pressure heat collection chamber for exhaust, an air supply pressure regulating fan, and an exhaust pressure regulating fan. The temperature control unit includes a directional air supply nozzle and a Coanda attached nozzle. The air supply pressure regulating fan is connected to the positive pressure static pressure chamber for air supply, and the exhaust pressure regulating fan is connected to the negative pressure heat collection chamber for exhaust. The air supply pressure regulating fan is connected to an outdoor refrigerant pipeline, and the exhaust pressure regulating fan is connected to an outdoor exhaust pipeline. The directional air supply nozzle and the Coanda attached nozzle are both located on the lower panel of the positive pressure static pressure chamber for air supply. The refrigeration unit, the dual-cavity ceiling thermal insulation unit, the temperature control unit, and the operating condition identification unit are all connected to the main control unit. The main control unit is used to execute the air conditioning cooling control method described in this application.

[0024] In this embodiment, the system includes a refrigeration unit, a dual-cavity ceiling-mounted thermal isolation unit, an operating condition identification unit, a main control unit, and a temperature control execution unit, with each unit connected to the main control unit. This application integrates physical isolation of hot and cold airflow, automatic identification of cabinet operating conditions, dynamic zoned cooling based on heat load, precise air delivery based on operating conditions, and adaptive closed-loop feedback into a complete hardware system architecture through multi-unit collaboration and modular connection. This provides clear structural support and connection relationships for the aforementioned methods, making the system as a whole feasible and deployable. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a schematic diagram illustrating the steps of an air conditioning cooling control method according to an embodiment of this application; Figure 2 This is a schematic diagram of an air conditioning cooling control system provided in one embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application. It should be understood that the step numbers used herein are only for convenience of description and are not intended to limit the order of execution of the steps.

[0028] To make the above-mentioned objectives, features, and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. In an exemplary embodiment, such as Figure 1 As shown, this application provides an air conditioning cooling control method, applicable to an air conditioning cooling control system. The air conditioning cooling control system includes a refrigeration unit, a dual-cavity ceiling cold and heat isolation unit, a main control unit, a temperature control execution unit, and an operating condition identification unit. The dual-cavity ceiling thermal insulation unit includes a positive pressure static pressure chamber for air supply, a negative pressure heat collection chamber for exhaust, an air supply pressure regulating fan, and an exhaust pressure regulating fan. The temperature control unit includes a directional air supply nozzle and a Coanda attached nozzle. The air supply pressure regulating fan is connected to the positive pressure static pressure chamber for air supply, and the exhaust pressure regulating fan is connected to the negative pressure heat collection chamber for exhaust. The air supply pressure regulating fan is connected to an outdoor refrigerant pipeline, and the exhaust pressure regulating fan is connected to an outdoor exhaust pipeline. The directional air supply nozzle and the Coanda attached nozzle are both located on the lower panel of the positive pressure static pressure chamber for air supply. The refrigeration unit, the dual-cavity ceiling thermal insulation unit, the temperature control unit, and the operating condition identification unit are all connected to the main control unit. The air conditioning cooling control method includes steps S101 to S104.

[0029] Step S101: Collect the operating condition data of the cabinet; the operating condition data includes the air outlet wind speed, the natural wind speed in the space, and the temperature value required for heat exchange temperature difference.

[0030] The air velocity at the rack outlet is used to determine the active exhaust intensity of the rack, and is collected by a wireless infrared wind speed sensor installed at the rack outlet. The natural wind speed in the space is used to eliminate interference from ambient airflow, and is collected by a wireless infrared wind speed sensor of the same specification installed in an open and unobstructed area of ​​the equipment room (away from the rack outlet or cold source).

[0031] The required temperature values ​​for the heat exchange temperature difference include: the rack inlet air temperature, which serves as the benchmark for calculating the heat exchange temperature difference in active air outlet mode, and is collected by a wireless temperature sensor installed inside the inlet grille of the active air outlet rack; the rack outlet air temperature, i.e., the surface temperature, which serves as the benchmark for calculating the temperature difference to determine operating conditions, and is collected by a wireless temperature sensor installed inside the air outlet of the active air outlet rack, and at the center of the main heat dissipation surface of the rack in the non-air outlet rack; and the equipment room ambient reference temperature, which serves as the benchmark for calculating the heat exchange temperature difference in the non-air outlet mode, and is collected by a wireless temperature sensor installed 1m below the ceiling, away from the cold source and the air outlet. Step S102: The operating condition data is transmitted to the operating condition identification unit so that the operating condition identification unit determines the first operating condition mode of the current cabinet based on the wind speed difference and heat exchange temperature difference, and transmits the first operating condition mode to the main control unit; the first operating condition mode includes active air outlet mode or no air outlet mode; the wind speed difference is the difference between the air outlet wind speed of the cabinet and the natural wind speed in the space.

[0032] Specifically, the heat exchange temperature difference is calculated using the corresponding formula based on the operating mode, matching the heat dissipation characteristics of different cabinets. The calculation formula is as follows: in, This refers to the rack intake air temperature. For rack exhaust / surface temperature, This is the reference temperature for the equipment room environment.

[0033] Furthermore, the operating condition identification unit determines the first operating condition mode of the current cabinet based on the wind speed difference and heat exchange temperature difference, including: Obtain the second operating mode; the second operating mode is the current first operating mode; When the second operating mode is the no-airflow mode, the difference between the rack outlet air temperature and the rack inlet air temperature is used as the heat exchange temperature difference; if the wind speed difference is greater than or equal to the first wind speed threshold, the heat exchange temperature difference is greater than or equal to the first temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the first preset number of cycles, then the second operating mode is determined to be the active airflow mode; otherwise, the first operating mode remains unchanged. When the second operating mode is the active air outlet mode, the difference between the cabinet surface temperature and the ambient reference temperature is used as the heat exchange temperature difference; if the wind speed difference is less than or equal to the second wind speed threshold, the heat exchange temperature difference is less than or equal to the second temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the second preset number of cycles, then the second operating mode is determined to be the no air outlet mode; otherwise, the first operating mode remains unchanged. Wherein, the first wind speed threshold is greater than the second wind speed threshold, and the first temperature difference threshold is greater than the second temperature difference threshold.

[0034] Specifically, it solves the problems of frequent boundary value switching and mode dead zone by using a dual judgment of "wind speed difference + temperature difference between modes" and built-in hysteresis anti-shake logic with current mode anchoring. The specific judgment rules are as follows: The system default initial mode is no airflow mode. ; If the current mode is no air outlet Switch to active air supply Conditions: must be met simultaneously ( And the condition is met for three consecutive control cycles (each cycle is 0.5s, for a total of 1.5s); If the current mode is active air supply Switch to no air outlet Conditions: must be met simultaneously ( And it is satisfied for 3 consecutive control cycles; Under non-switching conditions, the current operating mode remains unchanged, and the hysteresis width of 0.2m / s and 1℃ ensures stable operating condition judgment without frequent switching.

[0035] In this embodiment, after obtaining the second operating mode, the wind speed difference is calculated, and the corresponding temperature difference calculation method is selected according to the current mode—in the no-airflow mode, the difference between the rack outlet temperature and the inlet temperature is used as the heat exchange temperature difference; in the active airflow mode, the difference between the rack surface temperature and the ambient reference temperature is used as the heat exchange temperature difference. Furthermore, a hysteresis comparison logic is adopted, where the first wind speed threshold is greater than the second wind speed threshold and the first temperature difference threshold is greater than the second temperature difference threshold. The mode is switched only after the conditions are maintained continuously for a preset number of cycles. Compared with the existing technology that uses a single threshold for instantaneous determination and is prone to frequent mode switching due to short-term fluctuations in wind speed or temperature, this application significantly improves the stability and anti-interference capability of operating mode identification through a composite logic of mode anchoring, asymmetric threshold hysteresis, and delayed confirmation.

[0036] Step S103: Based on the first operating mode, obtain the real-time active power and temperature data of the cabinet, calculate the real-time heat load of a single cabinet, dynamically divide different heat generation zones according to the real-time heat load of the single cabinet, and output the cooling capacity allocation results of each zone.

[0037] Further, based on the first operating mode, the real-time active power and temperature data of the rack are obtained, the real-time heat load of a single rack is calculated, and different heat-generating zones are dynamically divided according to the real-time heat load of the single rack, and the cooling capacity allocation results of each zone are output, including: When the first operating mode is active air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the heat dissipation efficiency coefficient by the real-time active power of the cabinet; when the first operating mode is no air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the convective heat transfer coefficient by the effective heat dissipation area of ​​the cabinet and then by the difference between the center temperature of the main heat dissipation surface of the cabinet and the ambient reference temperature. Based on the ratio of the real-time heat load of a single cabinet to the rated maximum heat load of the cabinet, the cabinet is dynamically divided into high-heat zone, medium-heat zone and low-heat zone. The cooling capacity allocation results for each zone are output according to the heat load ratio of each zone.

[0038] Specifically, based on the operating mode, select the corresponding formula to calculate the real-time heat load Q of a single cabinet: In active airflow mode, the server rack has its own built-in cooling fan. The heat load is determined by the rack's active power and heat exchange efficiency, and the heat output is: in, This is the heat dissipation efficiency coefficient of the server rack, a fixed value calibrated at the factory. This represents the real-time active power of the server rack.

[0039] In the no-exhaust mode, the heat load is determined by convection heat transfer. Considering whether the KONE unit has an attached air supply or not, the heat load of a single cabinet in the no-exhaust mode is calculated as follows: in, The convective heat transfer coefficient characterizes the heat exchange capacity between the surface of the air-cooled cabinet and the environment under natural convection (without Coanda air supply). Forced convection (with air supply attached to the Coanda) W / (㎡·℃), For the Coanda nozzle air velocity; The effective heat dissipation area of ​​the cabinet is the surface area of ​​the cabinet without exhaust fan that participates in convection heat exchange, minus the ineffective area in contact with the ground and walls; The temperature at the center of the main heat dissipation surface of the cabinet is collected in real time by a wireless temperature sensor. The reference temperature for the equipment room environment is collected by a wireless temperature sensor.

[0040] Furthermore, each server rack is divided into high / medium / low heat zones according to the following rules: High fever zone A: ; Central heating zone B: ; Low-heat zone C: ; Specifically, the total heat load is aggregated by zone, using the following formula: in, The rated maximum heat load of the cabinet is the maximum theoretical heat output of a single cabinet under rated operating conditions. This refers to the internal temperature of the server rack, which is the real-time operating temperature of the central area inside the server rack. This represents the number of server racks within the partition.

[0041] Total heat load of the entire computer room: ,in For the total heat load of high-heat zone A, For the total heat load of heat-generating zone B, The total heat load for low-heat zone C is [value]; the cooling capacity allocation ratio for each zone is [value]. ,satisfy ,and Actual cooling capacity of each zone: , , .

[0042] Temperature control core constraints: This ensures that the cooling capacity fully covers the heat load of the corresponding zone.

[0043] Based on the heat load, the cooling supply is proportionally allocated. High heat generation zones are given priority and sufficient cooling to dissipate heat accumulation, while medium / low heat generation zones are given cooling as needed, avoiding waste of cooling capacity and local overheating, and achieving precise matching between cooling capacity and heat load.

[0044] In this embodiment, different heat load calculation formulas are used for active air supply mode and no air supply mode: in active air supply mode, the calculation is based on active power and heat dissipation efficiency coefficient; in no air supply mode, the calculation is based on convection heat transfer, effective heat dissipation area, and temperature difference. The cabinet is dynamically divided into high, medium, and low heat generation zones according to the ratio of the heat load to the rated maximum heat load, and then the cooling capacity is allocated proportionally. Compared with existing technologies that use fixed power thresholds or uniform cooling, this application uses heat load calculation driven by the physical mechanism of different modes and proportional zoned cooling to accurately match the cooling capacity with the actual heat generation of the cabinet, avoiding over-cooling or under-cooling.

[0045] Furthermore, after outputting the cooling capacity allocation results for each zone, it also includes: Historical zone heat load data from multiple past control cycles is obtained, and the historical zone heat load data is input into a time-series prediction model to predict the zone heat load at future times. When the predicted value is higher than the preset ratio of the current zone heat load, the upper limit of the reference output of the refrigeration unit is pre-adjusted to be higher. When the predicted value is lower than the preset ratio of the current zone heat load, the baseline output lower limit of the refrigeration unit will be lowered.

[0046] Furthermore, the time-series prediction model is an LSTM model; before acquiring historical zonal heat load data from multiple past control cycles and predicting future zonal heat loads using the time-series prediction model, the following steps are also included: Collect historical heat load data for multiple consecutive days in different zones and divide it into training and validation sets according to a preset ratio; The historical zone heat load data in the training set are input into the LSTM model in time sequence to obtain the heat load prediction value; The LSTM model is trained using the mean square error between the predicted heat load and the actual collected zone heat load data as the loss function, and the Adam optimization algorithm is employed.

[0047] Specifically, heat load forecasting is performed. Based on heat load data from the past 18 cycles, the zone heat load for the next 15 minutes is calculated using the following formula to generate a pre-cooling strategy: in, For the future That is, the future predicted by the system The target zone total heat load at any time, in W (watts), is the core input of the precooling strategy. It is used to adjust the output of the refrigeration unit and phase change cold storage in advance to avoid temperature control lag. The current control moment is the system's current time node, corresponding to the timestamp of the main control cycle. It serves as the reference moment for all historical heat load data, ensuring timing consistency. It is provided in real time by the system time synchronization module. The heat load forecast duration is the time interval from the current time t to the forecast time, corresponding to the look-ahead window of the pre-cooling strategy, and adapting to the typical time scale of changes in the heat load of the computer room. This is the LSTM bias term, a time-series prediction bias, used to correct systematic biases in predicted values, compensate for model training errors and static heat loads in the computer room (such as lighting, wall heat dissipation, and other fixed heat loads), avoid systematically overestimating / underestimating predicted values, and improve prediction accuracy. It is dimensionless and related to the weighting coefficients. Synchronous training and synchronous fixation occur during LSTM model training, and the data is updated periodically in sync with the weights during operation. Preferably, The model was trained using an LSTM model, optimized through backpropagation, and then solidified. The detailed process is as follows: Collect seven consecutive days of historical heat load data for the target computer room, with the sampling period consistent with the system's main control cycle. / Second-rate, (Generally taken as 0-1s), a total of 7×24×3600 ÷ The dataset consists of a set of valid data, which is divided into a training set and a validation set in an 8:2 ratio. Construct a simplified LSTM network adapted for time-series heat load prediction, with an input layer dimension of 18 (corresponding to heat load data from 18 historical periods), a hidden layer of 1 (32 neurons), and an output layer dimension of 1 (corresponding to future heat load data). (Predicted heat load at any given time) The training data is input into the LSTM network in chronological order. The mean squared error (MSE) between the predicted and measured values ​​is used as the loss function. The Adam optimization algorithm is employed, with a learning rate of 0.001 and 50 iterations. The network weights are continuously adjusted through backpropagation. The core optimization object is the 18 temporal weight coefficients ω1∼ω 18 The optimization process follows the principle of "recent historical data has a higher weight than long-term data" to match the temporal variation of heat load. The validation set data is input into the trained LSTM network to verify the prediction accuracy. When the prediction error is ≤3%, the weight optimization is deemed qualified. If the error exceeds the standard, the number of iterations is increased (up to 80 rounds) or 2 days of historical data are added for retraining until the accuracy requirement is met. Finally, the specific values ​​of the 18 weight coefficients are determined. Optimize the qualified ω1∼ω 18 The weight coefficients are fixed in the storage module of the AI ​​main control unit. During the operation phase, the latest hot load data is collected for one day each month to fine-tune the training of the LSTM network, update the weight coefficients, and adapt to the long-term load change characteristics of the data center. In this embodiment, the time-series prediction model employs an LSTM model. Historical heat load data from multiple consecutive days is collected and divided into training and validation sets according to a preset ratio. The mean square error between the predicted and actual heat load values ​​is used as the loss function, and the Adam optimization algorithm is employed to train the model. During the operational phase, the latest data is periodically collected for fine-tuning. Compared to existing technologies that use fixed-parameter models or manually set thresholds, this application, through data-driven LSTM model training and periodic fine-tuning updates, enables the prediction model to adapt to the long-term changing characteristics of the data center load, maintaining prediction accuracy.

[0048] In this embodiment, after outputting the cooling capacity allocation result, historical zone heat load data from multiple control cycles is acquired. A time-series prediction model is used to predict the zone heat load at future times. Based on the predicted value relative to the current load, the upper or lower limit of the refrigeration unit's baseline output is adjusted in advance. Compared to existing technologies that only respond retrospectively based on the current load, this application implements a pre-cooling strategy through heat load prediction, adjusting the refrigeration unit's output in advance and avoiding temperature control delays caused by lag in the refrigeration system's response.

[0049] Step S104: Control the temperature control execution unit and the dual-cavity ceiling heat and cold isolation unit to supply cooling to each zone according to the cooling capacity allocation result.

[0050] The temperature control unit and the dual-cavity ceiling thermal insulation unit provide cooling to each zone according to the cooling capacity allocation results, including: The dual-cavity ceiling heat and cold isolation unit controls the distribution of cold air supplied by the cold source to the top of each zone cabinet through the positive pressure static pressure chamber. When in active air supply mode, the directional air supply nozzle of the temperature control unit is opened to deliver cold air to the air inlet of the active air supply cabinet. When in no-airflow mode, the Coanda attachment nozzle of the temperature control execution unit is opened to blow cold air onto the surface of the airless sealed cabinet; the inner side of the Coanda attachment nozzle is provided with an arc-shaped wall-mounting guide plate to allow the cold air to adhere to the surface of the cabinet and form an attachment laminar flow cold film. The dual-cavity ceiling heat and cold isolation unit actively captures and discharges the hot airflow generated by each cabinet through the exhaust negative pressure heat collection cavity.

[0051] In this embodiment, cold air is evenly distributed above each zone of the server racks via a positive pressure static pressure chamber. The corresponding nozzle is selected according to the first operating mode: in active airflow mode, the directional airflow nozzle delivers cold air to the rack's air inlet; in no-airflow mode, the Coanda attached nozzle blows cold air onto the rack surface. The exhaust negative pressure heat capture chamber actively captures the exhausted hot airflow. Compared to existing technologies with single-outlet uniform airflow and disordered mixing of hot and cold air, this application, through a physical isolation structure of lower-level airflow and upper-level heat exhaust, combined with operating mode-specific nozzle airflow, forms a complete directional airflow path from the cold source to the rack and then to the heat exhaust, eliminating the mixing of hot and cold air.

[0052] Furthermore, after cooling each zone, the process also includes: Collect real-time temperature inside the cabinet; The real-time temperature is compared with the target temperature to obtain the temperature deviation; When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments according to the correction signal.

[0053] Furthermore, when the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments according to the correction signal, including: When the absolute value of the temperature deviation is greater than the first temperature threshold and not greater than the second temperature threshold, the nozzle air volume corresponding to the first working mode is adjusted by the first adjustment step; wherein, when the first working mode is the active air outlet mode, the directional air outlet is adjusted, and when the first working mode is the no air outlet mode, the Coanda attached nozzle is adjusted. When the absolute value of the temperature deviation is greater than the second temperature threshold, the PID control parameters, the output power of the refrigeration unit, and the wind speed of the Coanda attached nozzle are adjusted according to the second adjustment step size. Wherein, the first adjustment step size is smaller than the second adjustment step size.

[0054] Preferably, discretized PID calculations are performed, and integral saturation and overshoot in temperature / voltage control are completely avoided through integral limiting constraints. This is used for system air supply voltage stabilization and precise cabinet temperature control. Discretized calculations output control commands such as fan speed and nozzle opening. Integral limiting can prevent output saturation caused by long-term integration. Specifically: Furthermore, the integral limiting constraint is as follows: in, The proportionality coefficient represents the deviation. For output The proportional effect is that the larger the coefficient, the faster the response speed. It is dimensionless and adapts to the system's rapid response requirements. It is obtained through adaptive optimization: the initial value is factory calibrated (temperature control). =5.0, regulated =3.5), during operation, the AI ​​main control unit makes real-time fine adjustments based on load fluctuations, and the optimal value is fixed once a month; The integral coefficient represents the effect of the accumulated deviation on the output u(k), used to eliminate steady-state deviations (such as long-term voltage regulation deviations and temperature control static errors), and is dimensionless; and... Synchronous adaptive optimization, initial value factory calibration (temperature control) =0.8, regulated voltage =0.5), subject to integral limit constraint to avoid integral saturation; For the first The deviation value for each control cycle, with units consistent with the control target (temperature in °C, pressure in Pa), represents the difference between the current measured value and the target value, and determines the direction and magnitude of adjustment; =Target value - Measured value; The target value is preset and fixed, while the measured value is collected in real time by the corresponding sensor (temperature / pressure sensor); The PID discretization sampling period is consistent with the system's main control period and synchronized with the periods of the global thermal parameter acquisition unit and the AI ​​main control unit, and is embedded in the PID submodule. This is the lower limit of the integral term, the minimum constraint value for the integral term, used to prevent negative integral saturation (output too low), avoid the actuator from malfunctioning, and control temperature. = (-10)-0℃·s, regulated voltage = (-20)-0 Pa·s; This is the upper limit of the integral term, the maximum constraint value for the integral term, used to prevent positive integral saturation (excessive output) and avoid overshoot in temperature control / voltage regulation. =0-10℃·s, voltage stabilization =0-20Pa·s.

[0055] Furthermore, the real-time temperature inside the cabinet was collected. Used to calculate temperature deviation , Set the target control temperature for the server rack; determine if the temperature control accuracy meets the standard (the required temperature control accuracy is ±0.3℃, i.e.) Based on temperature deviation and rack thermal inertia, the PID parameters are adaptively adjusted. Corrections are graded according to the magnitude of the deviation: fine adjustment level ( Adjust the nozzle airflow in 1%-5% increments, without adjusting the PID core parameters to avoid temperature control oscillations; coarse adjustment stage ( ): Adaptively adjust PID parameters, refrigeration unit output power, and KONE nozzle air velocity in steps of 2%-8%.

[0056] In this embodiment, when the absolute value of the temperature deviation is greater than a first temperature threshold but not greater than a second temperature threshold, only the nozzle airflow is adjusted. When the absolute value of the deviation is greater than the second temperature threshold, the PID parameters, the power of the cooling unit, and the Coanda nozzle airflow are adjusted simultaneously, with the first adjustment step size being smaller than the second adjustment step size. Compared with the existing single fixed gain adjustment scheme, this application performs correction actions of different magnitudes according to the degree of deviation, with fine-tuning for small deviations and coarse-tuning for large deviations, avoiding temperature control oscillations and improving the temperature control stability and response rationality of the system.

[0057] In this embodiment, after cooling each zone, the real-time temperature inside the cabinet is collected and compared with the target temperature. When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit, enabling the main control unit to make corresponding adjustments based on the correction signal. Compared with existing technologies that rely solely on ambient temperature or outlet air temperature for open-loop or semi-open-loop control, this application directly uses the cabinet's internal temperature as feedback, forming a closed-loop control chain of "cooling—collection—comparison—correction." This allows the main control unit to perceive changes in the thermal state inside the cabinet in real time and proactively intervene to adjust when the deviation exceeds the allowable range. This prevents the cabinet temperature from deviating from the target value due to sudden load changes, environmental disturbances, or other factors, thereby improving the accuracy of temperature control and the system's responsiveness, and providing a triggering prerequisite for subsequent graded corrections.

[0058] In this embodiment, a dual-cavity ceiling-mounted thermal isolation unit is used to achieve physical isolation of hot and cold airflows. A condition identification unit distinguishes between active air-discharge cabinets and closed cabinets without airflow. A main control unit dynamically divides heat-generating zones and allocates cooling capacity based on the real-time heat load of each cabinet. A temperature control execution unit performs condition-specific air delivery. Compared to existing whole-house centralized air conditioning systems with unified air delivery and disordered cold air diffusion, this application, through the synergistic effect of thermal isolation, condition identification, and dynamic zoned cooling, enables the cooling capacity to be directionally delivered to different heat-generating zones as needed. This fundamentally avoids the mixing and loss of cold and hot airflows, thus improving cooling capacity utilization.

[0059] Furthermore, the aforementioned air conditioning cooling control method further includes: Obtain the actual static pressure of the positive pressure static pressure chamber for air supply, and calculate the deviation between the actual static pressure and the target static pressure for air supply; When the absolute value of the deviation is greater than the first preset threshold, the speed of the air supply pressure regulating fan is adjusted to reduce the deviation; Obtain the actual negative pressure of the exhaust negative pressure heat collection chamber, and calculate the deviation between the actual negative pressure and the target negative pressure of the exhaust. When the absolute value of the deviation is greater than the second preset threshold, the speed of the exhaust pressure regulating fan is adjusted to reduce the deviation.

[0060] The chamber pressure stabilization and regulation logic is as follows: Calculate the supply air pressure deviation: ,like ( The value range is 0-10Pa. Adjust the speed of the air supply pressure regulating fan. Increasing the rotational speed increases the static pressure, Δ When the static pressure is less than 0, the rotational speed is reduced to decrease the static pressure. The actual static pressure of the positive pressure chamber is directly collected by a high-precision pressure sensor installed in the center of the positive pressure chamber. To achieve the target of stable air supply pressure control, the system is preset to a fixed pressure of 80Pa, with an adjustable range of 50-100Pa.

[0061] Calculate the exhaust pressure deviation: ,like ( The value range is 0-10Pa. Adjust the speed of the exhaust pressure regulating fan. Increasing the rotational speed increases the absolute value of the negative pressure. Reducing the rotation speed decreases the absolute value of the negative pressure, where, The actual negative pressure of the exhaust negative pressure chamber is directly collected by a high-precision pressure sensor installed in the center of the exhaust negative pressure chamber; The system is preset to a negative pressure for exhaust ventilation and a heat capture and pressure stabilization control target of -25Pa, with an adjustable range of -30 to -20Pa; otherwise, the current fan speed is kept constant.

[0062] In this embodiment, the actual static pressure of the positive pressure static pressure chamber is acquired and its deviation from the target static pressure is calculated. When the deviation exceeds a threshold, the speed of the supply air pressure regulating fan is adjusted to reduce the deviation. Simultaneously, the actual negative pressure of the negative pressure heat capture chamber is acquired and its deviation from the target negative pressure is calculated. When the deviation exceeds a threshold, the speed of the exhaust air pressure regulating fan is adjusted to reduce the deviation. Compared with existing static pressure chambers or negative pressure chambers without pressure closed-loop control, this application uses independent pressure closed-loop control to compensate for pressure fluctuations caused by changes in air resistance, filter blockage, etc., in real time, ensuring the uniformity of cooling in the positive pressure chamber and the heat plume capture effect in the negative pressure chamber.

[0063] Another embodiment of this application provides a cooling control system, including: Refrigeration unit 100, double-cavity ceiling heat and cold insulation unit 200, main control unit 300, temperature control execution unit 400 and operating condition identification unit 500; The dual-cavity ceiling thermal insulation unit 200 includes a positive pressure static pressure chamber for air supply, a negative pressure heat collection chamber for exhaust, an air supply pressure regulating fan, and an exhaust pressure regulating fan. The temperature control execution unit 400 includes a directional air supply nozzle and a Coanda attached nozzle. The air supply pressure regulating fan is connected to the positive pressure static pressure chamber for air supply, and the exhaust pressure regulating fan is connected to the negative pressure heat collection chamber for exhaust. The air supply pressure regulating fan is connected to an outdoor refrigerant pipeline, and the exhaust pressure regulating fan is connected to an outdoor exhaust pipeline. The directional air supply nozzle and the Coanda attached nozzle are both located on the lower panel of the positive pressure static pressure chamber for air supply. The refrigeration unit 100, the dual-cavity ceiling thermal insulation unit 200, the temperature control execution unit 400, and the operating condition identification unit 500 are all connected to the main control unit 300. The main control unit 300 is used to execute the air conditioning cooling control method described in this embodiment.

[0064] In this embodiment, the system includes a cooling unit 100, a dual-cavity ceiling-mounted thermal insulation unit 200, an operating condition identification unit 500, a main control unit 300, and a temperature control execution unit 400, with each unit connected to the main control unit 300. This application integrates physical isolation of hot and cold airflow, automatic identification of cabinet operating conditions, dynamic zoned cooling based on heat load, precise air delivery based on operating conditions, and adaptive closed-loop feedback into a complete hardware system architecture through multi-unit collaboration and modular connection. This provides clear structural support and connection relationships for the aforementioned methods, making the system as a whole feasible and deployable.

[0065] It is understood that the above-described device embodiments correspond to the method embodiments of this application, and can implement any of the above-described method embodiments of this application to provide an air conditioning cooling control method.

[0066] In this embodiment, the system includes a refrigeration unit, a dual-cavity ceiling-mounted thermal isolation unit, an operating condition identification unit, a main control unit, and a temperature control execution unit, with each unit connected to the main control unit. This application integrates physical isolation of hot and cold airflow, automatic identification of cabinet operating conditions, dynamic zoned cooling based on heat load, precise air delivery based on operating conditions, and adaptive closed-loop feedback into a complete hardware system architecture through multi-unit collaboration and modular connection. This provides clear structural support and connection relationships for the aforementioned methods, making the system as a whole feasible and deployable.

[0067] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An air conditioning temperature reduction control method, characterized by, This is applicable to an air conditioning cooling control system, which includes a refrigeration unit, a dual-cavity ceiling heat and cold insulation unit, a main control unit, a temperature control execution unit, and an operating condition identification unit. The dual-cavity ceiling thermal insulation unit includes a positive pressure static pressure chamber for air supply, a negative pressure heat collection chamber for exhaust, an air supply pressure regulating fan, and an exhaust pressure regulating fan. The temperature control unit includes a directional air supply nozzle and a Coanda attached nozzle. The air supply pressure regulating fan is connected to the positive pressure static pressure chamber for air supply, and the exhaust pressure regulating fan is connected to the negative pressure heat collection chamber for exhaust. The air supply pressure regulating fan is connected to an outdoor refrigerant pipeline, and the exhaust pressure regulating fan is connected to an outdoor exhaust pipeline. The directional air supply nozzle and the Coanda attached nozzle are both located on the lower panel of the positive pressure static pressure chamber for air supply. The refrigeration unit, the dual-cavity ceiling thermal insulation unit, the temperature control unit, and the operating condition identification unit are all connected to the main control unit. The air conditioning cooling control method includes: Collect the operating condition data of the cabinet; the operating condition data includes the air outlet wind speed, the natural wind speed in the space, and the temperature value required for heat exchange temperature difference; The operating condition data is transmitted to the operating condition identification unit, so that the operating condition identification unit determines the first operating condition mode of the current cabinet based on the wind speed difference and heat exchange temperature difference, and transmits the first operating condition mode to the main control unit; the first operating condition mode includes active air outlet mode or no air outlet mode; the wind speed difference is the difference between the air outlet wind speed of the cabinet and the natural wind speed in the space. According to the first operating mode, the real-time active power and temperature data of the cabinet are obtained, and the real-time heat load of a single cabinet is calculated. Based on the real-time heat load of the single cabinet, different heat-generating zones are dynamically divided and the cooling capacity allocation results of each zone are output. The temperature control unit and the dual-cavity ceiling heat and cold isolation unit are controlled to provide cooling to each zone according to the cooling capacity allocation result.

2. The air conditioning temperature reduction control method according to claim 1, characterized by, Based on the first operating mode, real-time active power and temperature data of the rack are obtained, and the real-time heat load of a single rack is calculated. Based on the real-time heat load of the single rack, different heat-generating zones are dynamically divided, and the cooling capacity allocation results for each zone are output, including: When the first operating mode is active air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the heat dissipation efficiency coefficient by the real-time active power of the cabinet; when the first operating mode is no air outlet mode, the real-time heat load of a single cabinet is calculated by multiplying the convective heat transfer coefficient by the effective heat dissipation area of ​​the cabinet and then by the difference between the center temperature of the main heat dissipation surface of the cabinet and the ambient reference temperature. Based on the ratio of the real-time heat load of a single cabinet to the rated maximum heat load of the cabinet, the cabinet is dynamically divided into high-heat zone, medium-heat zone and low-heat zone. The cooling capacity allocation results for each zone are output according to the heat load ratio of each zone.

3. The air conditioning cooling control method according to claim 1, characterized in that, To enable the operating condition identification unit to determine the first operating condition mode of the current cabinet based on the wind speed difference and heat exchange temperature difference, including: Obtain the second operating mode; the second operating mode is the current first operating mode; When the second operating mode is the no-airflow mode, the difference between the rack outlet air temperature and the rack inlet air temperature is used as the heat exchange temperature difference; if the wind speed difference is greater than or equal to the first wind speed threshold, the heat exchange temperature difference is greater than or equal to the first temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the first preset number of cycles, then the second operating mode is determined to be the active airflow mode; otherwise, the first operating mode remains unchanged. When the second operating mode is the active air outlet mode, the difference between the cabinet surface temperature and the ambient reference temperature is used as the heat exchange temperature difference; if the wind speed difference is less than or equal to the second wind speed threshold, the heat exchange temperature difference is less than or equal to the second temperature difference threshold, and the above conditions are maintained continuously for a period of time reaching the second preset number of cycles, then the second operating mode is determined to be the no air outlet mode; otherwise, the first operating mode remains unchanged. Wherein, the first wind speed threshold is greater than the second wind speed threshold, and the first temperature difference threshold is greater than the second temperature difference threshold.

4. The air conditioning cooling control method according to claim 1, characterized in that, The temperature control unit and the dual-cavity ceiling thermal insulation unit provide cooling to each zone according to the cooling capacity allocation results, including: The dual-cavity ceiling heat and cold isolation unit controls the distribution of cold air supplied by the cold source to the top of each zone cabinet through the positive pressure static pressure chamber. When in active air supply mode, the directional air supply nozzle of the temperature control unit is opened to deliver cold air to the air inlet of the active air supply cabinet. When in no-airflow mode, the Coanda attachment nozzle of the temperature control execution unit is opened to blow cold air onto the surface of the airless sealed cabinet; the inner side of the Coanda attachment nozzle is provided with an arc-shaped wall-mounting guide plate to allow the cold air to adhere to the surface of the cabinet and form an attachment laminar flow cold film. The dual-cavity ceiling heat and cold isolation unit actively captures and discharges the hot airflow generated by each cabinet through the exhaust negative pressure heat collection cavity.

5. The air conditioning cooling control method according to claim 1, characterized in that, After cooling each zone, the process also includes: Collect real-time temperature data inside the server rack; The real-time temperature is compared with the target temperature to obtain the temperature deviation; When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments according to the correction signal.

6. The air conditioning cooling control method according to claim 5, characterized in that, When the temperature deviation exceeds a preset threshold, a correction signal is sent to the main control unit so that the main control unit can make adjustments based on the correction signal, including: When the absolute value of the temperature deviation is greater than the first temperature threshold and not greater than the second temperature threshold, the nozzle air volume corresponding to the first working mode is adjusted by the first adjustment step; wherein, when the first working mode is the active air outlet mode, the directional air outlet is adjusted, and when the first working mode is the no air outlet mode, the Coanda attached nozzle is adjusted. When the absolute value of the temperature deviation is greater than the second temperature threshold, the PID control parameters, the output power of the refrigeration unit, and the wind speed of the Coanda attached nozzle are adjusted according to the second adjustment step size. Wherein, the first adjustment step size is smaller than the second adjustment step size.

7. The air conditioning cooling control method according to claim 1, characterized in that, After outputting the cooling capacity allocation results for each zone, the following is also included: Historical zone heat load data from multiple past control cycles is obtained, and the historical zone heat load data is input into a time-series prediction model to predict the zone heat load at future times. When the predicted value is higher than the preset ratio of the current zone heat load, the upper limit of the reference output of the refrigeration unit is pre-adjusted to be higher. When the predicted value is lower than the preset ratio of the current zone heat load, the baseline output lower limit of the refrigeration unit will be lowered.

8. The air conditioning cooling control method according to claim 7, characterized in that, The time-series prediction model is an LSTM model; before acquiring historical zonal heat load data from multiple past control cycles and predicting future zonal heat loads using the time-series prediction model, the following steps are also included: Collect historical heat load data for multiple consecutive days in different zones and divide it into training and validation sets according to a preset ratio; The historical zone heat load data in the training set are input into the LSTM model in time sequence to obtain the heat load prediction value; The LSTM model is trained using the mean square error between the predicted heat load and the actual collected zone heat load data as the loss function, and the Adam optimization algorithm is employed.

9. The air conditioning cooling control method according to claim 1, characterized in that, Also includes: Obtain the actual static pressure of the positive pressure static pressure chamber for air supply, and calculate the deviation between the actual static pressure and the target static pressure for air supply; When the absolute value of the deviation is greater than the first preset threshold, the speed of the air supply pressure regulating fan is adjusted to reduce the deviation; Obtain the actual negative pressure of the exhaust negative pressure heat collection chamber, and calculate the deviation between the actual negative pressure and the target negative pressure of the exhaust. When the absolute value of the deviation is greater than the second preset threshold, the speed of the exhaust pressure regulating fan is adjusted to reduce the deviation.

10. An air conditioning cooling control system, characterized in that, It includes a refrigeration unit, a double-cavity ceiling heat and cold insulation unit, a main control unit, a temperature control execution unit, and an operating condition identification unit; The dual-cavity ceiling thermal insulation unit includes a positive pressure static pressure chamber for air supply, a negative pressure heat collection chamber for exhaust, an air supply pressure regulating fan, and an exhaust pressure regulating fan. The temperature control unit includes a directional air supply nozzle and a Coanda attached nozzle. The air supply pressure regulating fan is connected to the positive pressure static pressure chamber for air supply, and the exhaust pressure regulating fan is connected to the negative pressure heat collection chamber for exhaust. The air supply pressure regulating fan is connected to an outdoor refrigerant pipeline, and the exhaust pressure regulating fan is connected to an outdoor exhaust pipeline. The directional air supply nozzle and the Coanda attached nozzle are both located on the lower panel of the positive pressure static pressure chamber for air supply. The refrigeration unit, the dual-cavity ceiling thermal insulation unit, the temperature control unit, and the operating condition identification unit are all connected to the main control unit. The main control unit is used to execute the air conditioning cooling control method according to any one of claims 1 to 9.

Images