Machine room inter-row air conditioning refrigeration control method, electronic device and program product

Abstract

The present disclosure provides a machine room inter-row air conditioner refrigeration control method, electronic equipment and program product, and relates to the technical field of machine room air conditioner control. The method comprises: acquiring mutual influence parameters between each inter-row air conditioner in a machine room and correlation weights between each inter-row air conditioner and each cabinet in the machine room; wherein the mutual influence parameters are determined based on the return air temperature change when each inter-row air conditioner is running, and are used to represent the mutual thermal influence degree between each inter-row air conditioner; the correlation weights are used to represent the refrigeration influence degree of each inter-row air conditioner on each cabinet; the real-time running power of each cabinet is acquired, and the refrigeration demand parameters of each cabinet are determined according to the real-time running power; the target running parameters of each inter-row air conditioner are determined according to the mutual influence parameters, the correlation weights and the refrigeration demand parameters, and each inter-row air conditioner is controlled to run according to the target running parameters. According to the embodiment of the present disclosure, the global precise distribution and dynamic optimization of refrigeration capacity can be realized.

Description

Technical Field

[0001] This disclosure relates to the field of computer room air conditioning control technology, and in particular to a computer room in-row air conditioning cooling control method, electronic equipment and program products. Background Technology

[0002] With the rapid development of communication technology, the power consumption and heat dissipation of IT equipment such as servers in data centers and communication equipment rooms continue to rise. In order to achieve efficient cooling and avoid local overheating, the industry generally adopts a cooling architecture in which in-row air conditioners are directly deployed between the rows of server racks, and combined with cold aisle or hot aisle sealing technology, so that cold air can enter the equipment in an orderly manner to complete the heat dissipation.

[0003] However, the existing inter-row air conditioning cooling capacity allocation cannot accurately match the actual heat load demand of each rack, which can easily lead to local hot spots or waste of cooling capacity, resulting in high overall energy consumption of the computer room. Summary of the Invention

[0004] This disclosure provides a method, electronic equipment, and program product for controlling the cooling of air conditioning in a computer room, which can achieve precise global allocation and dynamic optimization of cooling capacity, effectively avoid waste of cooling capacity, and significantly reduce the overall operating energy consumption of the computer room.

[0005] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0006] According to one aspect of this disclosure, a method for controlling the cooling of air conditioning in a computer room is provided, comprising: acquiring mutual influence parameters between air conditioners in each row within the computer room and correlation weights between each air conditioner in each row and each server rack in the computer room; wherein, the mutual influence parameters are determined based on the return air temperature changes during the operation of each air conditioner in each row, and are used to characterize the degree of mutual thermal influence between the air conditioners in each row; the correlation weights are used to characterize the degree of cooling influence of each air conditioner in each row on each server rack; acquiring the real-time operating power of each server rack, and determining the cooling demand parameters of each server rack based on the real-time operating power; determining the target operating parameters of each air conditioner in each row based on the mutual influence parameters, correlation weights, and cooling demand parameters, and controlling each air conditioner in each row to operate according to the target operating parameters.

[0007] In one embodiment of this disclosure, the method further includes: controlling the target air conditioner to operate under a preset load state so that the airflow in the computer room reaches a stable thermal equilibrium state; recording the initial return air temperature and the final return air temperature after stabilization of the affected air conditioner, as well as the initial return air temperature and the final return air temperature after stabilization of the target air conditioner; and calculating the influence factor of the target air conditioner on the affected air conditioner based on the difference between the initial and final return air temperatures of the affected air conditioner and the difference between the initial and final return air temperatures of the target air conditioner, as a mutual influence parameter.

[0008] In one embodiment of this disclosure, the method further includes: for any target inter-row air conditioner, selecting a preset number of candidate influence parameters with the greatest influence from the calculated mutual influence parameters; normalizing the preset number of candidate influence parameters to obtain the mutual influence weights of the target inter-row air conditioner on each affected inter-row air conditioner; and determining the target operating parameters of each inter-row air conditioner based on the mutual influence parameters, the correlation weights, and the cooling demand parameters, including: using the mutual influence weights as the allocation coefficients for the target inter-row air conditioners to share the cooling task, and combining the correlation weights and the cooling demand parameters to determine the target operating parameters of each inter-row air conditioner.

[0009] In one embodiment of this disclosure, the method further includes: determining the corresponding air supply attenuation coefficient according to the channel isolation mode of the computer room, the channel isolation mode including closed cold aisle mode and closed hot aisle mode; obtaining the physical distance between each row air conditioner and each cabinet; and calculating the association weight of each row air conditioner to each cabinet based on the physical distance and the air supply attenuation coefficient.

[0010] In one embodiment of this disclosure, the association weight of each row air conditioner to each cabinet is calculated based on the physical distance and the air supply attenuation coefficient, including: when the aisle isolation mode is the closed cold aisle mode, the row air conditioner closest to the cabinet is taken as the primary associated air conditioner, and the row air conditioners on both sides of the cabinet are taken as secondary associated air conditioners; the primary association weight is calculated based on the distance between the primary associated air conditioner and the cabinet, and the secondary association weight is calculated based on the distance between the secondary associated air conditioner and the cabinet and the preset air supply attenuation coefficient.

[0011] In one embodiment of this disclosure, the cooling demand parameters of each cabinet are determined based on the real-time operating power, including: determining the required cooling capacity of each cabinet based on the real-time operating power and a preset redundancy coefficient; converting the required cooling capacity into the required air volume based on the specific heat capacity of air, air density, and the supply and return air temperature difference between the air conditioners in the rows, and using the required air volume as the cooling demand parameter.

[0012] In one embodiment of this disclosure, the target operating parameters of each row of air conditioners are determined based on mutual influence parameters, correlation weights, and cooling demand parameters. This includes: using the minimum total operating air volume or the minimum total energy consumption of each row of air conditioners as the objective function; using the constraint that the actual received air volume of each cabinet is not less than the cooling demand parameter and the operating air volume of each row of air conditioners does not exceed its maximum rated air volume; and using a linear programming algorithm to solve the objective function by combining the mutual influence parameters and correlation weights to obtain the target air volume or target fan speed of each row of air conditioners as the target operating parameters.

[0013] In one embodiment of this disclosure, the method further includes: obtaining the current temperature of each cabinet and the current operating parameters of each row air conditioner; inputting the current temperature, current operating parameters, cooling demand parameters, correlation weights, and mutual influence parameters into a pre-built temperature prediction model to obtain the predicted temperature of each cabinet in a future preset period; correcting the target operating parameters according to the predicted temperature, and controlling each row air conditioner to operate according to the corrected target operating parameters.

[0014] According to another aspect of this disclosure, an electronic device is provided, comprising: a memory for storing program instructions; and a processor for calling the program instructions stored in the memory to implement the above-described method for controlling the cooling of air conditioning in a computer room.

[0015] According to another aspect of this disclosure, a computer program product is provided, the computer program product including program instructions for implementing the above-described computer room in-row air conditioning cooling control method.

[0016] The technical solutions provided in this disclosure can include the following beneficial effects: This embodiment of the disclosure determines the cooling demand parameters by acquiring the real-time operating power of each cabinet, realizing the transformation from passive temperature feedback to active heat load matching. It can accurately meet the actual cooling needs of each cabinet and effectively avoid the generation of local hot spots. At the same time, by acquiring the mutual influence parameters and correlation weights determined based on the return air temperature change, the actual thermal coupling relationship between air conditioners in each row and between air conditioners and cabinets under complex airflow organization is accurately quantified. Based on this, the target operating parameters of each air conditioner are determined collaboratively, realizing the global accurate allocation and dynamic optimization of cooling capacity, effectively avoiding cooling waste and significantly reducing the overall operating energy consumption of the data center.

[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0018] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0019] Obviously, the accompanying drawings described below are merely some embodiments of this disclosure. Those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0020] Figure 1 This diagram illustrates an inter-row air conditioning layout according to an embodiment of the present disclosure; Figure 2 This diagram illustrates a flowchart of a computer room inter-row air conditioning cooling control method according to an embodiment of the present disclosure; Figure 3 This diagram illustrates the flowchart for determining the mutual influence parameters in an embodiment of this disclosure. Figure 4 This document illustrates a flowchart of the association weight calculation for the inter-row air conditioners to each cabinet in an embodiment of this disclosure. Figure 5 This invention discloses a flowchart of another method for controlling the cooling of an inter-row air conditioning system in a computer room, as shown in an embodiment of the present invention. Figure 6 This diagram illustrates the model prediction flowchart in an embodiment of the present disclosure. Figure 7 This illustration shows another inter-row air conditioning layout according to an embodiment of the present disclosure; Figure 8 A structural block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0021] To facilitate understanding of the technical solutions of this disclosure, the disclosure will be further described below with reference to the accompanying drawings.

[0022] The terms "first" and "second," etc., used in this disclosure are only used to distinguish different objects and not to describe a specific order. The terms "comprising" and "having," and any variations thereof, in the embodiments of this disclosure are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may also include steps or units not listed.

[0023] Throughout this disclosure, the references to "embodiments" do not necessarily refer to the same embodiments, nor are they independent or alternative embodiments mutually exclusive with other embodiments. Those skilled in the art will understand, explicitly and implicitly, that the embodiments described in this disclosure can be combined with other embodiments.

[0024] In the embodiments of this disclosure, "at least one" refers to one or more, "more" refers to two or more, "at least two" refers to two or three or more, and "and / or" is used to describe the relationship between associated objects. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0025] To facilitate understanding, the relevant concepts involved in this disclosure are explained below: In-row air conditioning is a high-efficiency precision air conditioner that uses horizontal air supply and is directly installed between rows of data center racks, specifically designed for hot and cold aisle isolation environments.

[0026] Effective diffusion distance refers to the distance that the airflow can reach when its central axis velocity decays to a certain value.

[0027] As the power consumption of data center communication equipment continues to increase, its heat dissipation also increases significantly. To achieve efficient equipment cooling and avoid localized overheating, the industry commonly uses in-row air conditioners deployed directly between server racks to dissipate heat by being close to the heat source, thus eliminating localized hotspots. Simultaneously, the use of cold / hot aisle isolation or cold / hot aisle sealing technology allows cool air to enter the equipment in an orderly manner to complete the heat dissipation process, effectively preventing the cooling efficiency reduction caused by short-circuiting of cold and hot airflows.

[0028] According to relevant data center design specifications, the air conditioning units between rows in Class A and B server rooms should be configured as an N+X system, meaning that the number of air conditioners should be configured with N units designed in advance, plus X redundant units (X=1~N), and hot and cold aisle isolation should be provided. When the air conditioners are operating, the primary air conditioner should be turned on first. When the primary air conditioner's cooling capacity is insufficient or during maintenance, the backup air conditioner should be turned on. Because the heat output of the server racks varies, the efficiency of the air conditioners between racks in the same row also differs; therefore, the airflow of different primary air conditioners will also be inconsistent.

[0029] Existing data centers commonly use in-row air conditioners of the same size as server racks, with one air conditioner placed between several racks. This method brings the cold end closer to the heat-generating equipment, but it also poses a risk of localized hotspots. For example... Figure 1 As shown, a data center has 22 server racks, each with a power of 8kW, and 6 in-row air conditioners (5 main and 1 standby). Air conditioner e is the standby unit. Each air conditioner has a rated cooling capacity of 35kW and is arranged in a closed cold aisle configuration. The cold air delivered by the in-row air conditioners diffuses and transfers cooling energy to surrounding racks. When racks far from the air conditioners (far end) experience a surge in heat dissipation due to increased traffic, localized hotspots can easily form, as seen in the rack near the standby air conditioner in the diagram. The existing air conditioning system relies on return air temperature for temperature control. It takes 5-10 minutes for hot air to diffuse from the rack back panel to the air conditioner return air vent, resulting in a response delay. Furthermore, since the standby air conditioner is only used for redundancy and does not participate in real-time temperature control, when the main air conditioner detects an abnormal return air temperature, it increases its fan speed, while the standby air conditioner closest to the hotspot remains idle and cannot adjust the temperature in time to eliminate the risk of high temperatures.

[0030] In one solution, after the return air temperature of the air conditioner rises due to local hot spots, the air conditioner increases the fan speed to increase the air volume. When this method was used on site, it was found that there were local hot spots near the standby air conditioner. However, the main air conditioner, which is far from the hot spots, had a poor ability to eliminate the hot spots by increasing the fan speed. Another solution involves regularly sending maintenance personnel to the site to measure the temperature and adjust the air conditioning parameters, or installing separate exhaust devices (electric fans) in the hot aisle to eliminate local hot spots. While this method can be effective to some extent, it still has many drawbacks and cannot meet the requirements of today's intelligent and standardized management of computer rooms.

[0031] Currently, the operation of in-row air conditioning systems is prone to hotspots. Existing technologies either rely on numerous sensors, increasing costs and maintenance complexity, or are overly theoretical and lack practical application. Therefore, this disclosure provides an in-row air conditioning cooling control scheme for computer rooms. This scheme requires no changes to existing computer room conditions and effectively manages the temperature field of the computer room by intelligently processing data from in-row air conditioning systems and server racks, combined with a power and environmental monitoring system platform.

[0032] The deficiencies of the above solutions and the proposed solutions are the result of the inventor's practice and careful research. Therefore, the discovery process of the above problems and the solutions proposed in this disclosure below should be considered as the inventor's contribution to this disclosure.

[0033] The following detailed description of this exemplary implementation method is provided in conjunction with the accompanying drawings and embodiments.

[0034] Figure 2 This diagram illustrates a flowchart of a computer room in-row air conditioning cooling control method according to an embodiment of the present disclosure. This method can be executed by a computer room in-row air conditioning cooling control system. Figure 2 As shown, the computer room inter-row air conditioning cooling control method provided in this embodiment includes S201 to S203.

[0035] In S201, the mutual influence parameters between the air conditioners in each row of the computer room and the association weights between the air conditioners in each row and each rack in the computer room are obtained. The mutual influence parameters are determined based on the return air temperature changes during the operation of the air conditioners in each row and are used to characterize the degree of mutual thermal influence between the air conditioners in each row. The association weights are used to characterize the degree of cooling influence of the air conditioners in each row on each rack.

[0036] In S202, the real-time operating power of each rack is acquired, and the cooling demand parameters of each rack are determined based on the real-time operating power. By acquiring the real-time operating power of each rack, the instantaneous changes in the heat load of IT equipment in the data center can be dynamically sensed.

[0037] In some embodiments, determining the cooling demand parameters for each rack based on real-time operating power can be achieved through various thermodynamic conversions or model calculations. For example, in one implementation, the required cooling capacity for each rack is determined based on real-time operating power and a preset redundancy coefficient; the required cooling capacity is converted into required airflow based on air specific heat capacity, air density, and the supply and return air temperature difference between the air conditioners in each rack, and the required airflow is used as the cooling demand parameter. The preset redundancy coefficient is used to cope with instantaneous peak power consumption of the equipment or environmental thermal disturbances, ensuring cooling margin. In another implementation, the transient required cooling capacity for each rack can also be calculated based on real-time operating power and the difference between the current actual temperature and the target set temperature of each rack, combined with a pre-built rack thermodynamic model, and then converted into required airflow based on air property parameters.

[0038] In S203, the target operating parameters of each row air conditioner are determined based on the mutual influence parameters, correlation weights, and cooling demand parameters, and each row air conditioner is controlled to operate according to the target operating parameters.

[0039] Due to the complex airflow organization within the server room, the air supply from a single in-row air conditioner can simultaneously affect multiple server racks, while a single server rack can receive cool air from multiple air conditioners. Therefore, when determining the target operating parameters, the first step is to construct an air supply distribution matrix for each in-row air conditioner to each server rack using correlation weights. Then, mutual influence parameters are used to correct for cooling capacity coupling and airflow interference between air conditioners, thereby establishing a mapping relationship between the air conditioner operating parameters and the actual airflow received by the server racks. Subsequently, this mapping relationship is combined with the previously determined cooling demand parameters, and an optimization algorithm is used to solve for the globally optimal operating parameters of each air conditioner that satisfy the cooling needs of all server racks.

[0040] In one specific implementation, the target operating parameters for each row of air conditioners are determined based on mutual influence parameters, correlation weights, and cooling demand parameters. These parameters include: using the minimum total operating airflow or the minimum total energy consumption of each row of air conditioners as the objective function; and using the constraint that the actual received airflow of each cabinet is not less than the cooling demand parameter, and that the operating airflow of each row of air conditioners does not exceed its maximum rated airflow. Combining the mutual influence parameters and correlation weights, a linear programming algorithm is used to solve the objective function, obtaining the target airflow or target fan speed for each row of air conditioners, which serves as the target operating parameter. Through this global optimization solution, the overall energy consumption of the air conditioning system can be minimized while ensuring that no local hotspots occur in any cabinet.

[0041] This embodiment of the disclosure achieves a shift from passive temperature response to active heat load matching by acquiring the rack operating power in real time and dynamically calculating cooling demand parameters. By introducing mutually influential parameters and correlation weights, it accurately quantifies the thermal coupling relationship between the air conditioning room and the air conditioner and the rack under complex airflow organization. Based on this, it uses an optimization algorithm to globally solve the target operating parameters. This not only achieves precise on-demand allocation of cooling capacity to racks with high heat loads and completely eliminates the hidden danger of local hot spots, but also effectively avoids excessive cooling supply and ineffective interference between air conditioners, significantly reducing the overall operating energy consumption of the air conditioning system between computer rooms.

[0042] In some embodiments, the above-described inter-row air conditioning cooling control method may further include: Figure 3 S301 to S303 are shown.

[0043] In S301, the target inter-row air conditioner is controlled to operate under a preset load state so that the airflow in the computer room reaches a stable state of thermal balance.

[0044] The above steps involve controlling the target air conditioner to operate under a preset load (such as full load or a specific high load) and actively changing its air supply state, thereby stimulating significant airflow disturbance and heat exchange processes in the computer room, and waiting for the airflow to reach a stable thermal equilibrium state in order to capture the true thermal impact of the air conditioner on the surrounding environment.

[0045] In S302, the initial return air temperature and the final return air temperature after stabilization of the affected inter-row air conditioner are recorded, as well as the initial return air temperature and the final return air temperature after stabilization of the target inter-row air conditioner.

[0046] In S303, the influence factor of the target air conditioner on the affected air conditioner is calculated based on the difference between the initial and final return air temperatures of the affected air conditioner and the difference between the initial and final return air temperatures of the target air conditioner, and is used as a mutual influence parameter.

[0047] This disclosure calculates the impact factor by recording the changes in return air temperature (i.e., the difference between the initial and final return air temperature) of the affected and target in-row air conditioners before and after the disturbance. By using the change in return air temperature rather than the absolute temperature value, the interference caused by the fluctuation of the background temperature in the computer room can be effectively eliminated, thus reflecting the actual heat penetration degree of the target air conditioner's supply air to the return air inlet of the affected air conditioner more purely and accurately.

[0048] In some embodiments, the above-described computer room in-row air conditioning cooling control method may further include, for any target in-row air conditioner, selecting a preset number of candidate influence parameters with the greatest influence from the calculated mutual influence parameters; normalizing the preset number of candidate influence parameters to obtain the mutual influence weights of the target in-row air conditioner on each affected in-row air conditioner. The above-described determination of the target operating parameters for each in-row air conditioner based on the mutual influence parameters, correlation weights, and cooling demand parameters includes: using the mutual influence weights as allocation coefficients for the cooling tasks shared by the target in-row air conditioners, and combining the correlation weights and cooling demand parameters to determine the target operating parameters for each in-row air conditioner.

[0049] In a real-world data center environment, the airflow from one in-row air conditioner typically affects multiple surrounding air conditioners, but the impact on those located further away or physically obstructed is negligible. Therefore, for any target in-row air conditioner, selecting a preset number (e.g., 2 or 4 adjacent units) of candidate influence parameters with the greatest impact from the calculated mutual influence parameters can filter out weak and noise-sensitive secondary influences. Subsequently, these primary candidate influence parameters are normalized to a sum of 1, thus obtaining the mutual influence weight of the target in-row air conditioner on each affected in-row air conditioner. When determining the target operating parameters, this mutual influence weight is used as the allocation coefficient for the cooling task shared by the target in-row air conditioner. This means that when calculating the total cooling capacity required by the target air conditioner, not only is the demand of its directly responsible cabinet considered (through associated weights), but also its thermal interference compensation to adjacent air conditioners is taken into account through the allocation coefficient, thereby achieving decoupling and linearization of multivariate coupling relationships in the mathematical model. By screening and normalizing the main influencing parameters, minor interferences are effectively eliminated, reducing the computational complexity of the system and improving the convergence speed and real-time performance of the algorithm. On the other hand, the mutual influence weights are used as allocation coefficients, so that each row of air conditioners can reasonably share the additional heat load in adjacent areas caused by airflow crosstalk while undertaking its own direct cooling task for the cabinet.

[0050] In some embodiments, the above-described inter-row air conditioning cooling control method may further include: Figure 4 S401 to S403 are shown.

[0051] In S401, the corresponding air supply attenuation coefficient is determined according to the channel isolation mode of the computer room. The channel isolation modes include closed cold aisle mode and closed hot aisle mode.

[0052] The air supply attenuation coefficient characterizes the degree of attenuation of cooling capacity or airflow during the transmission of cold air from the inter-row air conditioner to the rack air inlet, caused by factors such as airflow diffusion, entrainment of surrounding hot air, and channel resistance. Specifically, different channel isolation modes exhibit significant differences in their ability to constrain airflow. For example, the closed cold aisle mode effectively limits the diffusion of cold airflow, resulting in less cooling capacity attenuation and strong directionality; while the closed hot aisle mode allows cold airflow to diffuse rapidly in open spaces, leading to greater attenuation. Therefore, dynamically matching the corresponding air supply attenuation coefficient based on the channel isolation mode can accurately reflect the cooling capacity transmission loss characteristics under specific physical conditions.

[0053] In S402, obtain the physical distance between each row air conditioner and each cabinet.

[0054] In S403, the association weight of each row air conditioner to each cabinet is calculated based on the physical distance and the air supply attenuation coefficient.

[0055] In some embodiments, the association weight of each row air conditioner to each cabinet is calculated based on the physical distance and the air supply attenuation coefficient, including: when the aisle isolation mode is the closed cold aisle mode, the row air conditioner closest to the cabinet is taken as the primary associated air conditioner, and the row air conditioners on both sides of the cabinet are taken as secondary associated air conditioners; the primary association weight is calculated based on the distance between the primary associated air conditioner and the cabinet, and the secondary association weight is calculated based on the distance between the secondary associated air conditioner and the cabinet and the preset air supply attenuation coefficient.

[0056] This disclosed embodiment introduces a channel isolation mode and physical distance, and combines this with an air supply attenuation coefficient to calculate the association weight, fully considering the impact of the actual airflow organization characteristics of the computer room on cold air transmission. Especially for the closed cold aisle mode, by distinguishing between primary and secondary associated air conditioners and employing different distance and attenuation calculation logics, it accurately matches the physical laws of directional cold air transport and lateral infiltration within the cold aisle. This weight calculation method based on real physical environment constraints and airflow attenuation characteristics effectively overcomes the errors caused by traditional methods that rely solely on simple spatial distance to estimate association, enabling high-precision quantification of the actual cooling contribution of each row of air conditioners to each rack. This provides extremely reliable data support for subsequent global accurate allocation of cooling capacity, complete elimination of local hotspots, and significant reduction in overall system energy consumption.

[0057] In some embodiments, the above-described method for controlling the cooling of air conditioning between computer rooms may further include obtaining the current temperature of each cabinet and the current operating parameters of each air conditioner between the rows; inputting the current temperature, current operating parameters, cooling demand parameters, correlation weights, and mutual influence parameters into a pre-built temperature prediction model to obtain the predicted temperature of each cabinet in a future preset period; correcting the target operating parameters based on the predicted temperature, and controlling each air conditioner between the rows to operate according to the corrected target operating parameters.

[0058] The pre-built temperature prediction model can be constructed based on machine learning algorithms such as Long Short-Term Memory (LSTM) networks and Transformer. To further improve the accuracy and environmental adaptability of the prediction, the temperature prediction model also deeply integrates time and seasonal parameters at the feature input end. The time parameters can include specific time periods within a day (such as peak business hours and nighttime off-peak hours) and weekday / holiday attributes, used to accurately capture the high-frequency fluctuation patterns of IT equipment workload within short periods. The seasonal parameters are used to characterize macroscopic operating condition changes over long periods, specifically including the current season (such as spring, summer, autumn, winter) or monthly characteristics. By inputting the above multi-dimensional time features along with the thermodynamic state parameters and spatial coupling parameters within the data center into the model, the temperature prediction model can comprehensively capture the complex dynamic patterns of the data center's thermal environment under the dual effects of "short-term high-frequency fluctuations" and "long-term macroscopic evolution," thereby accurately outputting the temperature change trend of each cabinet within a preset future time period. After obtaining the predicted temperature, the system assesses whether the predicted temperature exceeds the safety threshold. If there is a risk of exceeding the limit, algorithms such as Model Predictive Control (MPC) are used to pre-correct the target operating parameters, and the air conditioners in each row are controlled to operate according to the corrected parameters. This predictive control strategy not only ensures the absolute safety and reliability of IT equipment operation, but also avoids energy waste caused by over-cooling or frequent adjustments, further improving the overall energy efficiency, stability, and intelligence level of the computer room's in-row air conditioning system.

[0059] The following is in conjunction with the appendix Figure 5 This disclosure provides a detailed description of the air conditioning cooling control method for computer room rows according to embodiments of the present invention.

[0060] like Figure 5 As shown, first, each air conditioner in the row is adjusted to full load (i.e., operating at maximum speed under determined supply and return air temperatures). The stable return air temperature of the other air conditioners is recorded. That is, when the air conditioner is running at full load, the return air temperature of the other air conditioners changes due to the influence until it stabilizes. When the return air temperature of all air conditioners no longer changes, this moment and the corresponding data are recorded. The influence factor of air conditioner a on the other air conditioners is calculated with reference to the following formula:

[0061] in, It is the affected air conditioner x initial return air temperature. It is the final return air temperature after it stabilizes; It is the initial return air temperature of air conditioner A. It is the final return air temperature after it stabilizes.

[0062] Calculate in sequence , The larger the value, the greater the influence of air conditioner a on air conditioner x.

[0063] After obtaining X sets of data as described above, record the first four influencing factors and their corresponding air conditioners for each set. Calculate the weights of these four influencing factors as the basis for allocating cooling tasks. For example, if a set of data yields four values ​​i, j, k, and g, then the weight of i would be... .

[0064] In a closed cold aisle configuration, primary airflow is jet-driven, supplemented by diffused airflow. The nearest air conditioner opposite the server rack and the air conditioners on both sides are then linked. If the server rack is located in a corner, only one side and the opposite air conditioner are linked. Since jet-driven airflow is more efficient than diffused airflow, the influence weights are assigned based on the distance between the server rack and the air conditioners and the airflow pattern. The nearest air conditioner opposite the server rack has the highest influence weight. calculate:

[0065] The diffused airflow from both sides of the air conditioner is assigned different weights of influence based on distance. :

[0066] If it is a closed hot aisle mode, the main function is diffused air supply, which affects the weighting. :

[0067] in, This is an adjustable coefficient used to distinguish between jet air supply and diffused air supply effects. The initial value can be 0.6, and it can be dynamically adjusted according to the actual operation of the computer room to better adapt to different site conditions; X is the distance between a certain cabinet and the air conditioner opposite it. , The distance between the air conditioners on both sides of a certain server rack.

[0068] Due to the different locations of the server racks, some racks can be connected to a total of 3 air conditioners on the opposite side and both sides, while some racks can only be connected to the air conditioners on the opposite side and one side. It may be necessary to calculate the impact weight based on the situation of each rack.

[0069] After considering the influence between air conditioners and between air conditioners and server racks, the power of each server rack can be obtained through an environmental monitoring platform. In principle, electrical energy in electronic devices is basically converted into heat energy. Therefore, the cooling capacity of the air conditioner should not be less than the server rack power P. Considering various uncertain heat losses, this disclosed solution configures the required cooling capacity according to 1.2 times the server rack power. ,therefore:

[0070] Based on the above steps, the required cooling capacity of a certain air conditioner's associated cabinet can be determined. The cooling capacity calculation formula is as follows:

[0071] Where Q is the air volume; ρ is the air density, taken as 1.2 kg / m³; The specific heat capacity of air at constant pressure is taken as 1.005 kJ / (kg·℃); The supply and return air temperature difference of the air conditioner is a constant.

[0072] Therefore, the required air volume Q can be calculated based on the cabinet power P, thereby controlling the speed of the associated air conditioning fan to deliver the required air volume.

[0073] Based on the influencing factors between air conditioners calculated through the above steps, and knowing the airflow of each air conditioner to meet cooling requirements, the rotation speed of each row of air conditioners can be dynamically adjusted using a simple linear programming method. This forms a control strategy for the row-to-row air conditioners adapted to the server racks in the data center, minimizing the occurrence of local hotspots. Furthermore, based on the above steps, this disclosure proposes a rack-level temperature prediction system. It introduces a convolutional network model that integrates time and seasonal parameters. By learning and analyzing the daily power changes of each rack and the air conditioner operating parameters, it predicts future rack power changes under the current season and controls the rack ambient temperature in a timely manner, or identifies the possibility and occurrence points of potential local hotspots. Based on this, such as... Figure 6 As shown, the system combines the influencing factors between spaces and the correlation weights between server racks and air conditioners to dynamically generate hourly, scenario-adaptive air conditioning operation strategies. Through a closed-loop feedback mechanism, it continuously collects the strategy execution effects, optimizes strategy configuration, and dynamically configures backup air conditioner parameters based on the results. It can detect insufficient cooling capacity based on server rack availability, providing early warnings or meeting energy-saving requirements while ensuring cooling supply safety.

[0074] In this embodiment, the influence between air conditioner operations is represented by an influence factor based on the actual operating conditions of the air conditioner. This effectively avoids the inaccurate wind pressure calculation caused by complex conditions in the computer room environment, such as cabinet obstruction, air supply / return short circuit, and airflow turbulence. Subsequently, a fan speed and air supply distance model is established based on the air conditioner parameters, which is associated with each cabinet and its cooling air conditioner. The power supply parameters of the cabinets are obtained through the environmental monitoring system platform, and the power of each cabinet is calculated, thereby accurately and scientifically providing the required cooling capacity for the cabinets. Finally, a time-series neural network model is introduced, and the total cooling capacity is predicted in advance according to the season and day based on historical power consumption and cooling data, so as to achieve the effect of pre-heat suppression and elimination of hot spots.

[0075] To better understand the present disclosure, a specific embodiment will be described below.

[0076] The existing data center has a total of 22 server racks. At a certain time, the power consumption of each server rack is as follows: Figure 7As shown in the previous text, the air conditioners a to f were operated at full load in sequence, and their influence factors were measured as a series of 6 sets, as detailed in Table 1 below.

[0077] Table 1

[0078] The first four influencing factors in each data set are assigned weights, and factors with a weight of 0 are removed. See Table 2 below for details.

[0079] Table 2

[0080] Based on the steps outlined above, associate each cabinet with the air conditioner. For example, a manufacturer's in-row air conditioner with a rated cooling capacity of 35kW has a width × depth × height of 300 × 1200 × 2000 mm and a standard cabinet size of 19 inches. The cabinets within its effective air supply distance are calculated based on relevant parameters, as shown in Table 3 below.

[0081] Table 3

[0082] The required air volume for a single cabinet is shown in Table 4 below.

[0083] Table 4

[0084] The required air volume for each air conditioner is shown in Table 5 below.

[0085] Table 5

[0086] Therefore, the air speed can be adjusted according to the ratio of air volume to air speed between air conditioners to match the cooling demand.

[0087] Subsequently, linear programming was used to determine the constraints and objective function, ensuring that the effective air volume in all areas met the requirements. The final optimized air conditioning air volume is shown in Table 6 below.

[0088] Table 6

[0089] After optimization, the air volume was reduced to 83% of the original.

[0090] It is understandable that dynamic regulation that changes with the seasons requires more operational and time parameters, and this disclosure will not provide examples of each.

[0091] This disclosure verifies the mutual influence between each air conditioner through actual verification, avoiding unmeasurable airflow obstruction factors, and cleverly utilizes the relationship between each air conditioner to quantify the air supply impact. It abandons the previous approach of matching total rack power with total cooling capacity, instead associating each air conditioner with a single rack, quantifying the relationship between air conditioning cooling and the rack, and clearly defining the required cooling capacity and air conditioner distribution for each rack, enabling more precise and scientific control of air conditioning operation. Based on the actual rack power, it dynamically optimizes the air conditioning operation strategy through the influence relationship between air conditioners, the cooling capacity distribution relationship between racks and air conditioners, and the actual operating conditions. After analyzing the mutual influence and constraints of various factors, it provides personalized adaptation and optimization solutions, optimizing air conditioning energy consumption while meeting cooling needs, and achieving energy-saving goals. It introduces a convolutional network model that integrates time and seasonal parameters to personalize the air conditioning operation strategy for each data center, each season, and each time period. It can also predict future rack power changes and identify the possibility and occurrence point of potential local hotspots, providing early warnings, achieving intelligent and safe operation and maintenance while complying with energy-saving policies.

[0092] Based on the same inventive concept, embodiments of this disclosure also provide an electronic device. For example... Figure 8 As shown, the electronic device provided in this embodiment includes a processor 801 and a memory 802; the memory 802 is used to store program instructions; the processor 801 is used to call the program instructions stored in the memory 802 to implement the computer room inter-row air conditioning cooling control method described in the above method embodiment.

[0093] In some embodiments, processor 801 is a device with data processing capabilities, including but not limited to a central processing unit (CPU), a field-programmable gate array (FPGA), and other programmable logic devices. Processor 801 may include one or more processing cores.

[0094] In some embodiments, the memory 802 is a device with data storage capability, including but not limited to random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), and flash memory.

[0095] In some embodiments, the processor 801 and memory 802 can be configured separately or integrated together. As an example, the processor 801 and memory 802 can be integrated on a single board or a system on a chip (SOC).

[0096] like Figure 8 As shown, the electronic device provided in this embodiment may further include an interface 803. The interface 803 is connected between the processor 801 and the memory 802, enabling information exchange between the processor 801 and the memory 802. In some embodiments, the processor 801, the memory 802, and the interface 803 are interconnected via a bus 804.

[0097] In some embodiments, there may be one or more processors 801.

[0098] In some embodiments, there may be one or more memories 802.

[0099] In some embodiments, there may be one or more interfaces 803.

[0100] In some embodiments, the memory 802 stores one or more program instructions. When the one or more program instructions are executed by the one or more processors 801, the one or more processors 801 implement the computer room inter-row air conditioning cooling control method described in the above method embodiments and achieve the corresponding technical effects.

[0101] In some embodiments, the electronic device may also communicate with one or more external devices, such as a keyboard, pointing device, Bluetooth device, etc. In some embodiments, the electronic device may also communicate with one or more devices that enable a user to interact with the electronic device. In some embodiments, the electronic device may also communicate with devices that enable the electronic device to communicate with one or more other computing devices. In some embodiments, the above communication may be performed through interface 803.

[0102] Understandable Figure 8 The illustrated structure does not constitute a specific limitation on the electronic device. In some embodiments, the electronic device may include... Figure 8 The number of components shown may be more or fewer. In some embodiments, the electronic device has... Figure 8 Based on the diagram, some components can be combined, some components can be separated, or different component arrangements can be made.

[0103] In some embodiments, Figure 8 The components shown can be implemented in hardware, software, or a combination of both.

[0104] Based on the same inventive concept, this disclosure also provides a computer-readable storage medium storing program instructions. When the program instructions are executed by a computer, they implement the computer room inter-row air conditioning cooling control method described in the above method embodiments, thereby achieving the corresponding technical effects.

[0105] In some embodiments, the computer-readable storage medium can be any available medium capable of storing program instructions or a data storage device such as a data center containing one or more available media. In some embodiments, the available medium can be a magnetic medium, an optical medium, or a semiconductor medium, etc. As an example, a magnetic medium can be a floppy disk, a hard disk, a magnetic tape, etc. As an example, an optical medium can be a high-density digital video optical disc. As an example, a semiconductor medium can be a solid-state drive.

[0106] Based on the same inventive concept, this disclosure also provides a computer program product, which includes program instructions for implementing the computer room in-row air conditioning cooling control method described in the above method embodiments. When executed by a computer, the program instructions cause the computer to implement the computer room in-row air conditioning cooling control method described in the above method embodiments, achieving the corresponding technical effects. In specific implementations, the program instructions can be written using any combination of one or more programming languages. The program instructions can be executed entirely on the user device, partially on the user device, as a standalone software package, partially on the user device and partially on a remote device, or entirely on a remote device or server.

[0107] Those skilled in the art will understand that all or part of the steps of the above embodiments can be specifically implemented in the following forms: a completely hardware implementation, a completely software implementation, or an implementation combining hardware and software aspects.

[0108] The above description is merely a specific embodiment of this disclosure, but the scope of protection of this disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for controlling the cooling of an air conditioning system in a computer room, characterized in that, include: The interaction parameters between air conditioners in each row within the computer room and the association weights between each air conditioner in each row and each server rack in the computer room are obtained. The interaction parameters are determined based on the return air temperature changes during the operation of each air conditioner in each row and are used to characterize the degree of mutual thermal influence between the air conditioners in each row. The association weights are used to characterize the degree of cooling influence of each air conditioner in each row on each server rack. Obtain the real-time operating power of each cabinet, and determine the cooling demand parameters of each cabinet based on the real-time operating power; Based on the mutual influence parameters, the correlation weights, and the cooling demand parameters, the target operating parameters for each row of air conditioners are determined, and each row of air conditioners is controlled to operate according to the target operating parameters.

2. The method according to claim 1, characterized in that, The method further includes: The target inter-row air conditioner is controlled to operate under a preset load state so that the airflow in the computer room reaches a stable state of thermal balance. Record the initial return air temperature and the final return air temperature after stabilization of the affected inter-row air conditioner, as well as the initial return air temperature and the final return air temperature after stabilization of the target inter-row air conditioner; Based on the difference between the initial and final return air temperatures of the affected in-row air conditioners and the difference between the initial and final return air temperatures of the target in-row air conditioner, the influence factor of the target in-row air conditioner on the affected in-row air conditioner is calculated and used as the mutual influence parameter.

3. The method according to claim 2, characterized in that, The method further includes: For any target inter-row air conditioner, select a preset number of candidate influence parameters with the greatest influence from the calculated mutual influence parameters; The preset number of candidate influence parameters are normalized to obtain the mutual influence weights of the target inter-row air conditioner on each affected inter-row air conditioner. The step of determining the target operating parameters for each row of air conditioners based on the mutual influence parameters, the correlation weights, and the cooling demand parameters includes: The mutual influence weights are used as the allocation coefficients for the cooling tasks of the target inter-row air conditioners. Combined with the correlation weights and the cooling demand parameters, the target operating parameters of each inter-row air conditioner are determined.

4. The method according to claim 1, characterized in that, The method further includes: The corresponding air supply attenuation coefficient is determined based on the channel isolation mode of the computer room, which includes closed cold aisle mode and closed hot aisle mode; Obtain the physical distance between each row of air conditioners and each server rack; Based on the physical distance and the air supply attenuation coefficient, the association weight of each row air conditioner to each cabinet is calculated.

5. The method according to claim 4, characterized in that, The step of calculating the association weight of each row of air conditioners to each cabinet based on the physical distance and the air supply attenuation coefficient includes: When the aisle isolation mode is the closed cold aisle mode, the air conditioner in the row opposite the cabinet that is closest to it is used as the main associated air conditioner, and the air conditioners in the rows on both sides of the cabinet are used as auxiliary associated air conditioners. The primary associated weight is calculated based on the distance between the primary associated air conditioner and the cabinet, and the secondary associated weight is calculated based on the distance between the secondary associated air conditioner and the cabinet and the preset air supply attenuation coefficient.

6. The method according to claim 1, characterized in that, The step of determining the cooling demand parameters for each cabinet based on the real-time operating power includes: The required cooling capacity for each cabinet is determined based on the real-time operating power and the preset redundancy coefficient. Based on the specific heat capacity of air, air density, and the supply and return air temperature difference between the air conditioners, the required cooling capacity is converted into the required air volume, and the required air volume is used as the cooling demand parameter.

7. The method according to claim 1, characterized in that, The step of determining the target operating parameters for each row of air conditioners based on the mutual influence parameters, the correlation weights, and the cooling demand parameters includes: The objective function is to minimize the total operating air volume or the total energy consumption of the air conditioners in each row. The constraints are that the actual air volume received by each cabinet is not less than the cooling demand parameters, and the operating air volume of each row air conditioner does not exceed its maximum rated air volume. By combining the mutual influence parameters and the correlation weights, the objective function is solved using a linear programming algorithm to obtain the target air volume or target fan speed of each air conditioner, which is then used as the target operating parameter.

8. The method according to claim 1, characterized in that, The method further includes: Obtain the current temperature of each cabinet and the current operating parameters of the air conditioners between each row; The current temperature, the current operating parameters, the cooling demand parameters, the correlation weights, and the mutual influence parameters are input into a pre-built temperature prediction model to obtain the predicted temperature of each cabinet in the future preset time period. Based on the predicted temperature, the target operating parameters are corrected, and the air conditioners between each row are controlled to operate according to the corrected target operating parameters.

9. An electronic device, characterized in that, include: Memory, used to store program instructions; A processor is configured to call the program instructions stored in the memory to implement the computer room inter-row air conditioning cooling control method as described in any one of claims 1 to 8.

10. A computer program product, characterized in that, It includes a computer program or instructions, which, when executed by a processor, implement the computer room inter-row air conditioning cooling control method as described in any one of claims 1 to 8.

Images