Server cooling control method and related equipment

Abstract

This application discloses a server cooling control method and related equipment, relating to the field of computer technology. This application can predict the amount of heat to be removed based on the measured temperature and flow rate of a target device, and implement cooling control by combining the heat difference between the heat to be removed and the average amount of heat to be removed, and the temperature difference between the target temperature and the current measured temperature. The temperature difference between the target temperature and the current measured temperature ensures the accuracy of the cooling control, and the heat difference between the heat to be removed and the average amount of heat to be removed enables the predictability of the cooling control, helping to improve the cooling effect.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a server cooling control method and related equipment. Background Art

[0002] During the operation of the server, some components may overheat. To ensure the normal operation of the server, the server needs to be cooled down by a cooling medium, which can be a liquid or a gas, such as water or air.

[0003] In related technologies, temperature sensors can be installed on server components that may overheat to detect the component's current temperature. The cooling medium flow rate is dynamically adjusted based on the component's current temperature. For example, if the current temperature is above a threshold, the flow rate is increased; if it is below the threshold, the flow rate is decreased. However, this solution still suffers from poor cooling performance. Summary of the Invention

[0004] The present application provides a server cooling control method and related equipment to at least solve the problem of poor cooling effect in the related art.

[0005] This application provides a server cooling control method, comprising:

[0006] Obtaining a measurement result sequence of the target component at the current moment on the server, the measurement result sequence including one or more measurement result information, the measurement result information including: measurement time, measurement temperature, and measurement flow rate of the cooling medium, where the measurement time is the current moment or a historical moment;

[0007] Predicting the amount of heat to be removed required for the target component to reach the target temperature based on the sequence of measurement results at the current moment, and determining an average amount of heat to be removed based on the predicted amounts of heat to be removed at multiple moments;

[0008] Calculating the temperature difference between the target temperature and the measured temperature at the current moment, and the heat difference between the predicted heat to be removed at the current moment and the average heat to be removed;

[0009] Cooling of target components is controlled based on temperature and heat differences.

[0010] The present application also provides a server cooling control device, comprising:

[0011] The measurement result acquisition module is used to obtain the measurement result sequence of the target component on the server at the current moment. The measurement result sequence includes one or more measurement result information. The measurement result information includes: measurement time, measurement temperature and measurement flow rate of the cooling medium. The measurement time is the current moment or a historical moment.

[0012] The prediction module is used to predict the amount of heat to be removed for the target component to reach the target temperature based on the measurement result sequence at the current moment, and to determine the average amount of heat to be removed based on the heat to be removed predicted at multiple moments.

[0013] The calculation module is used to calculate the temperature difference between the target temperature and the measured temperature at the current moment, and the heat difference between the heat to be removed predicted at the current moment and the average heat to be removed.

[0014] The cooling control module is used to control the cooling of the target component according to the temperature difference and the heat difference.

[0015] The present application also provides an electronic device, comprising:

[0016] memory for storing computer programs;

[0017] The processor is configured to implement the steps of the aforementioned server cooling control method when executing a computer program.

[0018] The present application also provides a computer-readable storage medium, in which a computer program is stored. When the computer program is executed by a processor, the steps of the aforementioned server cooling control method are implemented.

[0019] The present application also provides a computer program product, including a computer program, which implements the steps of the aforementioned server cooling control method when executed by a processor.

[0020] Through this application, the amount of heat to be removed can be predicted to achieve cooling control by combining the heat difference between the amount of heat to be removed and the average amount of heat to be removed, and the temperature difference between the target temperature and the current measured temperature. The temperature difference between the target temperature and the current measured temperature can ensure the accuracy of the cooling control, and the heat difference between the amount of heat to be removed and the average amount of heat to be removed can achieve the anticipation of the cooling control, which helps to improve the cooling effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 This is a schematic diagram of an application scenario of an embodiment of the present application;

[0023] Figure 2 This is a flow chart of the steps of a server cooling control method provided by an embodiment of the present application;

[0024] Figure 3 This is a flowchart of another server cooling control method provided by an embodiment of the present application;

[0025] Figure 4 A schematic structural diagram of a server cooling control device provided in an embodiment of the present application;

[0026] Figure 5 A schematic structural diagram of another server cooling control device provided in an embodiment of the present application;

[0027] Figure 6 This is a schematic diagram of the structure of the electronic device provided in this application. DETAILED DESCRIPTION

[0028] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0029] It should be noted that, in the description of this application, the terms "comprises," "includes," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. The terms "first," "second," etc., in this application are used to distinguish similar objects, and are not used to describe a particular order or sequence.

[0030] During server operation, some components may overheat, such as the central processing unit (CPU), graphics processing unit (GPU), and power supply module. Overheating of these components can cause server operation abnormalities and service interruptions. To ensure normal server operation, server cooling technology is required. Server cooling technology achieves cooling by transmitting a cooling medium through the server. The cooling medium can be liquid or gas. The liquid can be water or a specified coolant, and the gas is typically air.

[0031] Figure 1 This is a schematic diagram of an application scenario of an embodiment of the present application. Figure 1 As shown, the server is provided with a cooling device that can output cooling medium to designated components, including: a central processing unit, a graphics processing unit, and a power module.

[0032] In related technologies, the flow rate of the cooling medium can be fixed or dynamically adjusted. For dynamically adjusted flow rate, the flow rate is continuously adjusted based on the actual measured current temperature. For example, if the current temperature is still greater than a preset threshold, or if the current temperature is still greater than the previous temperature, the flow rate of the cooling medium can be increased.

[0033] Of course, compared to a fixed flow rate, dynamically adjusting the flow rate has better energy savings and cooling effects. However, the above dynamic flow rate adjustment can only control the cooling at the next moment based on the current temperature, which will cause cooling delays and still lead to poor cooling effects.

[0034] To address the above technical issues, the present application predicts the amount of heat to be removed based on the measured temperature and flow rate, and implements cooling control by combining the temperature difference between the amount of heat to be removed and the average amount of heat to be removed, and the temperature difference between the target temperature and the current measured temperature. The temperature difference between the target temperature and the current measured temperature ensures the accuracy of the cooling control, and the temperature difference between the amount of heat to be removed and the average amount of heat to be removed enables the predictability of the cooling control, which helps improve the cooling effect.

[0035] In order to enable those skilled in the art to better understand the present application, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0036] Figure 2 This is a flow chart of the steps of a server cooling control method provided by an embodiment of the present application. Figure 2 As shown, the server cooling control method includes:

[0037] S201: Obtain a measurement result sequence of a target component on a server at the current moment, where the measurement result sequence includes one or more measurement result information, including: measurement time, measurement temperature, and measurement flow rate of a cooling medium, where the measurement time is the current moment or a historical moment.

[0038] The target component here may be a key component on the server that overheats and causes abnormal operation of the server, such as a CPU, a GPU, and a power supply.

[0039] For a target component, the measurement result sequence includes measurement result information at one or more measurement moments, with each measurement moment corresponding to a measurement result information. The measurement result information at multiple measurement moments is arranged in order of measurement moment, for example, in ascending order of measurement moment, with the measurement result information at the earliest measurement moment being arranged at the front.

[0040] In embodiments of the present application, a measurement result sequence can be obtained according to a certain time period. The measurement result sequence obtained at each moment includes N+1 measurement result information corresponding to the current moment and the N historical moments before the current moment. For example, for an ascending time sequence: t_N+1, t_N, t_N-1, ..., t_2, t_1, t_0, at time t_0, the measurement result sequence obtained may be R(t_N), R(t_N-1), ..., R(t_1), R(t_0). At time t_1, the measurement result sequence obtained may be R(t_N+1), R(t_N), ..., R(t_2), R(t_1). Here, R(t_0) is the measurement result information at the current time t_0, R(t_N), R(t_N-1), ..., R(t_1) are the measurement result information corresponding to the N historical moments before the current time t_0, R(t_1) is the measurement result information at the current time t_1, and R(t_N+1), R(t_N), ..., R(t_2) are the measurement result information corresponding to the N historical moments before the current time t_1.

[0041] The above measurement results can be obtained through sensors, including temperature sensors and flow rate sensors. The temperature sensor can be an NTC (Negative Temperature Coefficient) thermistor, whose resistance decreases as temperature increases. The flow rate sensor can be a turbine flowmeter. The measured temperature is the temperature of the target component measured by the temperature sensor at the time of measurement. The measured flow rate is the flow rate of the cooling medium measured by the flow rate sensor at the time of measurement.

[0042] In conventional solutions, the number and location of sensors on a target component are fixed. However, this approach can lead to problems with too many or too few sensors. When there are too few sensors, measurements are inaccurate, and when there are too many sensors, sensors are wasted. To address this issue, the present application dynamically adjusts the sensor distribution to ensure measurement accuracy as much as possible while saving the number of sensors. Specifically, the temperature change rate and flow rate change rate of the target component are obtained; then, based on the temperature change rate and flow rate change rate of the target component, the sensor density on the target component is determined; finally, the sensor distribution on the target component is adjusted based on the sensor density.

[0043] In practical applications, it is necessary to first set a default number of sensors to measure the temperature and flow rate of the target component to determine the temperature change rate and flow rate change rate.

[0044] The temperature rate of change is the amount of temperature change per unit time. It can be calculated based on the temperatures measured at multiple measurement moments. For example, the temperature rate of change can be calculated as the ratio of the temperature difference between two measurement moments to the time difference between the two measurement moments. The two measurement moments can be adjacent or non-adjacent. Of course, for greater accuracy, the average of the temperature rates of change obtained across multiple measurement moments can be used as the temperature rate of change.

[0045] The flow rate change rate is the change in measured flow rate per unit time. It can be calculated based on the measured flow rates at multiple measurement moments. For example, the flow rate change rate can be calculated as the ratio of the difference in measured flow rates at two measurement moments to the time difference between the two measurement moments. The two measurement moments can be adjacent or non-adjacent. Of course, for greater accuracy, the flow rate change rate can be calculated as the average of the flow rate change rates obtained from multiple measurement moments.

[0046] Once the temperature change rate and flow rate change rate are determined, the sensor density on the target component can be determined. Sensor density can be the number of sensors per unit area. Sensor density can be positively correlated with the temperature change rate and flow rate change rate. That is, when the temperature change rate is constant, sensor density increases with increasing flow rate change rate and decreases with decreasing flow rate change rate. When the flow rate change rate is constant, sensor density increases with increasing temperature change rate and decreases with decreasing temperature change rate.

[0047] In some possible implementations, determining the sensor density on a target component based on the target component's temperature change rate and flow rate change rate includes: first, converting the target component's temperature change rate into a temperature sensor density based on a temperature sensitivity coefficient; then, converting the target component's flow rate change rate into a flow sensor density based on a flow rate sensitivity coefficient; and finally, determining the sensor density based on the temperature sensor density and the flow sensor density, wherein the sensor density is positively correlated with the temperature sensor density and the flow sensor density. This application can determine the densities of the temperature sensor and the flow sensor separately, thereby accurately estimating the total sensor density.

[0048] The temperature sensor density is positively correlated with the temperature change rate and the temperature sensitivity coefficient. The temperature sensor density is the number of temperature sensors per unit area, and the flow rate sensor density is the number of flow rate sensors per unit area.

[0049] In one example, the temperature sensor density can be the product of the temperature change rate and the temperature sensitivity coefficient, and the flow sensor density can be the product of the flow rate change rate and the flow sensitivity coefficient. The temperature sensitivity coefficient is related to the material of the temperature sensor, and the flow sensitivity coefficient is related to the material of the flow sensor.

[0050] After obtaining the above-mentioned temperature sensor density and flow rate sensor density, the sensor density can be determined. In conventional algorithms, the temperature sensor density and the flow rate sensor density can be added together to form the sensor density.

[0051] In another embodiment, the sensor density can be calculated using the following formula:

[0052] (1)

[0053] Where D is the sensor density, k is the conversion factor, a is the temperature sensitivity coefficient, TR is the temperature change rate, b is the flow rate sensitivity coefficient, and FR is the flow rate change rate. a × TR is the temperature sensor density, and b × FR is the flow sensor density.

[0054] It can be seen from the above formula (1) that the present application can dynamically adjust the sum of the temperature sensor density and the flow rate sensor density as the sensor density.

[0055] When the above-mentioned sensor density is obtained, the number of sensors can be calculated based on the area of ​​the target component and the sensor density, so that the sensors can be evenly arranged on the target device according to the number of sensors. The sensor is connected to the main control platform via an I2C (Inter-Integrated Circuit, serial communication bus protocol) SPI (Serial Peripheral Interface, synchronous serial communication interface) interface to send the collected measurement result sequence to the main control platform so that the main control platform executes the method of the present application. A database can also be set up to temporarily store the collected measurement result sequence so that the main control platform can read the measurement result sequence from the database and execute the method of the present application.

[0056] After obtaining the above measurement result sequence, the measurement result sequence may be preprocessed to improve the quality of the measurement result sequence. The preprocessing may include but is not limited to: outlier removal and data padding.

[0057] Among them, outlier removal is used to remove abnormal measurement result information. Specifically, the average measurement temperature, measurement temperature standard deviation, average measurement flow rate, and measurement flow rate standard deviation of the measurement result sequence can be calculated, and then for each measurement result information, the measurement temperature deviation between the measurement temperature in the measurement result information and the average measurement temperature is calculated, and the ratio of the measurement temperature deviation to the measurement temperature standard deviation is used as the abnormality of the measurement temperature. In addition, the measurement flow rate deviation between the measurement flow rate in the measurement result information and the average measurement flow rate is calculated, and the ratio of the measurement flow rate deviation to the measurement flow rate standard deviation is used as the abnormality of the measurement flow rate. Finally, when the abnormality of the measurement temperature and / or the abnormality of the measurement flow rate is greater than or equal to the preset abnormality, the measurement result information is deleted; when the abnormality of the measurement temperature and the abnormality of the measurement flow rate are both less than the preset abnormality, the measurement result information is retained.

[0058] Data filling is used to fill in missing measured temperatures and / or measured flow rates in the measurement result sequence. This can be achieved specifically through a linear interpolation algorithm. For example, the missing measured temperature is calculated as the average of the two adjacent measured temperatures, and the missing flow rate is calculated as the average of the two adjacent measured flow rates.

[0059] S202: predicting the amount of heat to be removed required for the target component to reach the target temperature based on the sequence of measurement results at the current moment, and determining an average amount of heat to be removed based on the predicted amounts of heat to be removed at multiple moments.

[0060] The target temperature is a temperature set for a target component. The target component is in an optimal operating state when the temperature is less than or equal to the target temperature. The target temperature may be different for different target components.

[0061] If the temperature of the target component is higher than the target temperature, the excess heat on the target component needs to be removed by the cooling medium so that the target component reaches the target temperature. The excess heat to be removed by the cooling medium is the heat to be removed.

[0062] In practical applications, a sequence of measurement results is obtained at each moment, and each sequence of measurement results can be used to predict a heat quantity to be removed. This sequence of measurement results can be input into a pre-trained neural network model to predict the corresponding heat quantity to be removed at that moment. Consequently, multiple measurement result sequences can be obtained at multiple moments, and multiple heat quantities to be removed can be predicted from these sequences. In other words, a heat quantity to be removed can be obtained at each moment, and the average of the heat quantities to be removed from these multiple moments can be used as the average heat quantity to be removed.

[0063] In some possible implementations, predicting the amount of heat required to reach the target temperature of the target component based on the current sequence of measurement results includes calculating statistical information based on the current sequence of measurement results, thereby predicting the amount of heat required to reach the target temperature of the target component based on the current sequence of measurement results and the statistical information. In this manner, combining statistical information with the change in measurement results over a period of time to predict the amount of heat required to reach the target temperature helps improve the accuracy of the prediction.

[0064] The statistical information includes at least one of the following: an average measured temperature, a maximum temperature difference, and a standard deviation of a measured flow rate.

[0065] The above average measured temperature is the average value of the measured temperatures at multiple measurement moments in the measurement result sequence at the current moment. For details, please refer to the following formula:

[0066] (2)

[0067] Where AMT is the average measured temperature, N+1 is the length of the measurement result sequence at the current moment, and MT(i) is the i-th measured temperature.

[0068] The maximum temperature difference is the difference between the maximum and minimum measured temperatures in the current measurement result sequence. Specific reference is made to the following formula:

[0069] (3)

[0070] Where MTM is the maximum temperature difference, max(MT(i)) is the maximum measured temperature, and min(MT(i)) is the minimum measured temperature.

[0071] The above-mentioned measured flow rate standard deviation is the standard deviation of multiple measured flow rates in the measurement result sequence at the current moment, and can be calculated using the following formula:

[0072] (4)

[0073] Where SDFV is the standard deviation of the measured flow rate, N+1 is the length of the measurement result sequence at the current moment, MF(i) is the i-th measured flow rate, and AMF is the average of multiple measured flow rates in the measurement result sequence at the current moment, that is, the average measured flow rate, which can be calculated using the following formula:

[0074] (5)

[0075] In one possible implementation, the current sequence of measurement results and statistical information can be input into a neural network model to obtain the amount of heat required for the target component to reach its target temperature. Different neural network models can be used for different target components, each of which records the target temperature of the target component. Of course, different target components can also share a set of neural network models. In this case, when predicting a specific target component, the target temperature of the target component, along with the current sequence of measurement results and statistical information, needs to be input into the neural network model for prediction to obtain the amount of heat required for the target component to reach its target temperature. In this way, the number of neural network models can be minimized, saving storage space.

[0076] In one embodiment, predicting the amount of heat required to reach the target temperature of a target component based on the current measurement result sequence and statistical information includes concatenating the measurement result sequence and statistical information into a feature vector for the target component, and then inputting the feature vector into a long short-term memory (LSTM) network model to obtain the amount of heat required to reach the target temperature. This allows the LSTM model to accurately extract changes in the measurement result information within the measurement result sequence, facilitating accurate prediction of the amount of heat required to reach the target temperature.

[0077] In some implementations, when multiple target components share the above-mentioned long short-term memory network model, or, when the target temperature of the target component changes, it is also necessary to splice the target temperature into the feature vector. The positions of the target temperature, measurement result sequence and statistical information in the feature vector can be flexibly set, and the embodiment of the present application does not limit the order of their positions. For example, the feature vector includes the measurement result sequence, statistical information and target temperature in sequence. In this way, whether the target temperatures of multiple target components are different, or the target temperature of the same target component changes, there is no need to adjust the long short-term memory network, which reduces the number of training times and costs.

[0078] The prediction process of the heat to be removed can be expressed by the following formula:

[0079] (6)

[0080] Where y(t) is the predicted amount of heat to be removed at time t, Wy is the weight matrix of the output layer of the long short-term memory network model, h(t) is the hidden state of the hidden layer output for the measurement result sequence at time t, By is the bias term of the output layer of the long short-term memory network model, f() is the activation function of the long short-term memory network model, Wh is the weight matrix of the hidden layer, h(t-1) is the hidden state of the hidden layer output for the measurement result sequence at time t-1, X(t) is the eigenvector of the measurement result sequence at time t, and Bh is the bias term of the hidden layer.

[0081] S203: Calculate the temperature difference between the target temperature and the measured temperature at the current moment, and the heat difference between the heat to be eliminated predicted at the current moment and the average heat to be eliminated.

[0082] The temperature difference may be a difference obtained by subtracting the measured temperature from the target temperature, and the heat difference may be a difference obtained by subtracting the average heat to be eliminated from the heat to be eliminated.

[0083] S204: Cooling the target component according to the temperature difference and the heat difference.

[0084] In a possible implementation, when the temperature difference is greater than a preset temperature difference, the target flow rate of the cooling medium may be increased, and when the temperature difference is less than the preset temperature difference, the target flow rate of the cooling medium may be decreased.

[0085] Accordingly, when the heat difference is greater than the preset heat difference, the target flow rate of the cooling medium can be increased, and when the heat difference is less than the preset heat difference, the target flow rate of the cooling medium can be reduced.

[0086] In some possible implementations, cooling control of a target component based on a temperature difference and a heat difference includes determining a target flow rate of a cooling medium based on the temperature difference and the heat difference, and controlling cooling of the target component based on the target flow rate. In this way, cooling control can be accurately achieved based on the target flow rate. In the embodiment of the present application, the main control platform is also connected to a cooling device to control the flow rate of the cooling device to be the target flow rate.

[0087] In some possible implementations, the target flow rate is positively correlated with the temperature difference and the heat difference. That is, when the temperature difference and the heat difference are larger, the target flow rate is larger, and when the temperature difference and the heat difference are smaller, the target flow rate is smaller. Thus, the target flow rate can be dynamically increased or decreased based on this relationship. For example, a mapping table can be set to query the target flow rate corresponding to the temperature difference and the heat difference, or the target flow rate can be calculated based on the functional relationship between the temperature difference, the heat difference, and the target flow rate.

[0088] In one possible implementation, determining the target flow rate of the cooling medium based on the temperature difference and the heat difference specifically includes: first, converting the temperature difference into a first predicted flow rate, and converting the heat difference into a second predicted flow rate, and determining the target flow rate of the cooling medium based on the first predicted flow rate and the second predicted flow rate. Performing separate conversions based on the temperature difference and the heat difference fully considers the independence of the two, while simultaneously determining the target flow rate based on both, thereby improving the accuracy of the target flow rate.

[0089] The target flow rate may be a weighted sum of the first predicted flow rate and the second predicted flow rate. The weighting coefficients of the first predicted flow rate and the second predicted flow rate may be flexibly set to adjust the degree of influence of the first predicted flow rate and the second predicted flow rate on the target flow rate.

[0090] The prediction process for the first predicted flow rate includes: multiplying the first conversion coefficient by the current temperature difference as a first sub-flow rate; and / or determining a second sub-flow rate based on the current temperature difference and the temperature differences at multiple historical moments before the current moment; and then determining the first predicted flow rate based on the first sub-flow rate and / or the second sub-flow rate. It can be seen that the first predicted flow rate can be determined using one or both of these methods to improve its accuracy.

[0091] The above-mentioned second sub-flow rate not only takes into account the temperature difference at the current moment, but also comprehensively considers the temperature difference at historical moments. That is, prediction is made based on the temperature difference over a period of time, which can improve the accuracy of the first predicted flow rate and avoid erroneous predictions caused by occasional fluctuations in temperature differences.

[0092] Specifically, the above-mentioned second sub-flow rate can be determined by the following process: first, the temperature difference integral term at the previous moment is updated by the temperature difference at the current moment to obtain the temperature difference integral term at the current moment; then, based on the temperature difference at the current moment and the temperature difference at the previous moment, the temperature difference change rate at the current moment is determined; finally, based on the temperature difference at the current moment, the temperature difference integral term at the current moment and the temperature difference change rate at the current moment, the second sub-flow rate is determined.

[0093] The above temperature difference integral term is positively correlated with the temperature difference integral term at the previous moment. The temperature difference integral term is used to indicate the accumulation of temperature differences and can also be understood as the cumulative temperature difference. It can be expressed by the following formula:

[0094] (7)

[0095] Where I(t) is the integral term of the temperature difference at the current time t, I(t-1) is the integral term of the temperature difference at the previous time t-1, e(t) is the temperature difference at the current time t, and ITL is the sampling time interval, that is, the time difference between t and t-1.

[0096] The temperature difference change rate is used to indicate the speed of change of the temperature difference. It can be defined as the ratio of the difference between the temperature difference at the current moment and the temperature difference at the previous moment to the sampling time interval. It can be specifically expressed by the following formula:

[0097] (8)

[0098] Where D(t) is the rate of change of the temperature difference at the current time t, and e(t-1) is the temperature difference at the previous time t-1.

[0099] After determining the temperature difference integral term, temperature difference change rate, and temperature difference, a second sub-flow rate can be determined based on the temperature difference integral term, temperature difference change rate, and temperature difference. The second sub-flow rate is positively correlated with the temperature difference integral term, temperature difference change rate, and temperature difference. This allows the second sub-flow rate to comprehensively consider the current temperature difference, the temperature difference at historical times, and the temperature difference change rate, thereby improving the accuracy of the second sub-flow rate.

[0100] In one implementation, the temperature difference integral term, the temperature difference change rate, and the temperature difference may be added together or weighted to form the second sub-flow rate.

[0101] In another implementation, determining the second sub-flow rate based on the current temperature difference, the current temperature difference integral term, and the current temperature difference change rate includes weighting the current temperature difference, the current temperature difference integral term, and the current temperature difference change rate to obtain a control parameter, and mapping the control parameter to the second sub-flow rate based on the maximum flow rate, wherein the second sub-flow rate is positively correlated with the maximum flow rate and the control parameter. In this manner, the second sub-flow rate can be accurately controlled within the maximum flow rate, ensuring the rationality of the second sub-flow rate.

[0102] The above control parameters can be expressed by the following formula:

[0103] (9)

[0104] Among them, U(t) is the control parameter at the current time t, Kp, Ki and Kd are the weights of temperature difference, temperature difference integral term and temperature difference change rate, respectively, which can be flexibly adjusted.

[0105] In one example, the second sub-flow rate may be the product of the maximum flow rate and the control parameter, or the ratio of the product of the maximum flow rate and the control parameter to a flow rate mapping coefficient, for example, as shown in the following formula:

[0106] (10)

[0107] Where V2(t) is the second subflow velocity at the current time t, and Vmax is the maximum flow velocity. B is the flow velocity mapping coefficient, which can be flexibly set according to the value range of U(t) so that U(t) / B is between 0 and 1. For example, when the maximum value of U(t) is 100, B can be set to 100.

[0108] In one embodiment of the present application, the first predicted flow rate is determined based on the first sub-flow rate and / or the second sub-flow rate, including the following three situations:

[0109] In the first case, the first sub-flow rate is used as the first predicted flow rate.

[0110] In the second case, the second sub-flow rate is used as the second predicted flow rate.

[0111] In the third case, the first sub-flow rate and the second sub-flow rate are weighted to obtain the first predicted flow rate. The weighting coefficients of the first and second sub-flow rates can be flexibly set to adjust the degree of their influence on the final first predicted flow rate. This allows for compatibility between the two prediction algorithms, helping to improve the accuracy of the first predicted flow rate.

[0112] In one embodiment of the present application, the heat difference is converted into the second predicted flow rate by multiplying the second conversion coefficient and the heat difference as the second predicted flow rate. In this way, the second conversion coefficient can be pre-set, and the conversion can be achieved through a simple multiplication operation, which reduces the computational complexity.

[0113] It can be seen that the first predicted flow rate may be a weighted combination of the first sub-flow rate and / or the second sub-flow rate, and the second predicted flow rate.

[0114] In summary, this application can perform cooling control based on the temperature difference combined with the predicted heat difference, which can be performed in advance and improve the cooling control effect. In addition, the sensor distribution can be dynamically adjusted to ensure measurement accuracy as much as possible while saving sensors.

[0115] After each cooling control cycle, and before the next cooling control cycle, the first conversion coefficient, the second conversion coefficient, the weights corresponding to the temperature difference, the temperature difference integral term, and the temperature difference change rate used in calculating the control parameters, and the weights of the first sub-flow rate and the second sub-flow rate used in calculating the first predicted flow rate can be dynamically adjusted. This allows the next cooling control cycle to use the adjusted parameters, thereby improving the dynamics of the cooling control.

[0116] Figure 3 This is a flow chart of another server cooling control method provided by an embodiment of the present application. Figure 3 As shown, the server cooling control method includes:

[0117] S301: Converting the temperature change rate of the target component on the server into a temperature sensor density according to the temperature sensitivity coefficient, and converting the flow velocity change rate of the target component into a flow velocity sensor density according to the flow velocity sensitivity coefficient.

[0118] S302: Determine a sensor density according to the temperature sensor density and the flow rate sensor density, and adjust the sensor distribution on the target component according to the sensor density.

[0119] The sensors include: a temperature sensor and a flow rate sensor, the sensor density is positively correlated to the temperature sensor density and the flow rate sensor density, and the sensor density is positively correlated to the temperature change rate and the flow rate change rate.

[0120] S303: Obtain a sequence of measurement results of the target component at the current moment on the server.

[0121] The measurement result sequence includes one or more pieces of measurement result information, and the measurement result information includes: measurement time, measurement temperature, and measurement flow rate of the cooling medium. The measurement time is the current time or a historical time.

[0122] S304: Calculate statistical information based on the measurement result sequence at the current moment, and combine the measurement result sequence and the statistical information into a feature vector of the target component.

[0123] The statistical information includes at least one of the following: an average measured temperature, a maximum temperature difference, and a standard deviation of a measured flow rate.

[0124] S305: Input the feature vector into the long short-term memory network model to obtain the amount of heat to be removed required for the target component to reach the target temperature, and determine the average amount of heat to be removed based on the heat to be removed predicted at multiple moments.

[0125] S306: Calculate the temperature difference between the target temperature and the measured temperature at the current moment, and the heat difference between the heat to be removed predicted at the current moment and the average heat to be removed.

[0126] S307: The product of the first conversion coefficient and the temperature difference at the current moment is used as the first sub-flow rate.

[0127] S308: updating the temperature difference integral term at the previous moment by the temperature difference at the current moment to obtain the temperature difference integral term at the current moment, and determining the temperature difference change rate at the current moment according to the temperature difference at the current moment and the temperature difference at the previous moment.

[0128] S309: Weighting the temperature difference, the temperature difference integral term, and the temperature difference change rate at the current moment to obtain a control parameter, and mapping the control parameter to a second sub-flow rate according to the maximum flow rate.

[0129] The second sub-flow rate is positively correlated with the maximum flow rate and is positively correlated with the control parameter.

[0130] S310: Weighting the first sub-flow rate and the second sub-flow rate to obtain a first predicted flow rate, and taking the product of the second conversion coefficient and the heat difference as the second predicted flow rate.

[0131] S311: Determine a target flow rate of the cooling medium according to the first predicted flow rate and the second predicted flow rate, so as to perform cooling control on the target component based on the target flow rate.

[0132] The steps S301 to S311 can be specifically referred to the aforementioned Figure 2 The server cooling control method shown will not be described in detail here.

[0133] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method.

[0134] Figure 4 This is a schematic diagram of the structure of the server cooling control device provided in the embodiment of the present application. Figure 4 As shown, an embodiment of the present application further provides a server cooling control device 400, comprising:

[0135] The measurement result acquisition module 401 is used to obtain the measurement result sequence of the target component on the server at the current moment. The measurement result sequence includes one or more measurement result information. The measurement result information includes: measurement time, measurement temperature and measurement flow rate of the cooling medium. The measurement time is the current moment or a historical moment.

[0136] The prediction module 402 is configured to predict the amount of heat to be removed required for the target component to reach the target temperature based on the measurement result sequence at the current moment, and determine an average amount of heat to be removed based on the predicted amounts of heat to be removed at multiple moments.

[0137] The calculation module 403 is used to calculate the temperature difference between the target temperature and the measured temperature at the current moment, and the heat difference between the heat to be removed predicted at the current moment and the average heat to be removed.

[0138] The cooling control module 404 is used to control the cooling of the target component according to the temperature difference and the heat difference.

[0139] In some possible implementations, the cooling control module 404 is further configured to:

[0140] A target flow rate of the cooling medium is determined according to the temperature difference and the heat difference; and cooling of the target component is controlled based on the target flow rate.

[0141] In some possible implementations, the cooling control module 404 is further configured to:

[0142] The temperature difference is converted into a first predicted flow rate; the heat difference is converted into a second predicted flow rate; and a target flow rate of the cooling medium is determined based on the first predicted flow rate and the second predicted flow rate.

[0143] In some possible implementations, the cooling control module 404 is further configured to:

[0144] The product of the first conversion coefficient and the temperature difference at the current moment is used as the first sub-flow rate, and / or the second sub-flow rate is determined by the temperature difference at the current moment and the temperature differences at multiple historical moments before the current moment; the first predicted flow rate is determined based on the first sub-flow rate and / or the second sub-flow rate.

[0145] In some possible implementations, the cooling control module 404 is further configured to:

[0146] The temperature difference integral term at the previous moment is updated by the temperature difference at the current moment to obtain the temperature difference integral term at the current moment; the temperature difference change rate at the current moment is determined based on the temperature difference at the current moment and the temperature difference at the previous moment; the second sub-flow rate is determined based on the temperature difference at the current moment, the temperature difference integral term at the current moment and the temperature difference change rate at the current moment.

[0147] In some possible implementations, the cooling control module 404 is further configured to:

[0148] The temperature difference, the temperature difference integral term and the temperature difference change rate at the current moment are weighted to obtain the control parameter; the control parameter is mapped to a second sub-flow rate according to the maximum flow rate, and the second sub-flow rate is positively correlated with the maximum flow rate and the control parameter.

[0149] In some possible implementations, the cooling control module 404 is further configured to:

[0150] The first sub-flow rate and the second sub-flow rate are weighted to obtain a first predicted flow rate.

[0151] In some possible implementations, the cooling control module 404 is further configured to:

[0152] The product of the second conversion coefficient and the heat difference is used as the second predicted flow rate.

[0153] In some possible implementations, the prediction module 402 is further configured to:

[0154] Calculate statistical information based on the current measurement result sequence, where the statistical information includes at least one of the following: an average measured temperature, a maximum temperature difference, and a standard deviation of the measured flow rate; and predict the amount of heat to be removed for the target component to reach the target temperature based on the current measurement result sequence and the statistical information.

[0155] In some possible implementations, the prediction module 402 is further configured to:

[0156] The measurement result sequence and statistical information are spliced ​​into a feature vector of the target component; the feature vector is input into a long short-term memory network model to obtain the heat to be removed for the target component to reach the target temperature.

[0157] In some possible implementations, refer to Figure 5 As shown, the above device also includes:

[0158] The change rate acquisition module 405 is used to acquire the temperature change rate and flow rate change rate of the target component.

[0159] The density determination module 406 is configured to determine the sensor density on the target component according to the temperature change rate and the flow rate change rate of the target component, where the sensor density is positively correlated with the temperature change rate and the flow rate change rate.

[0160] The sensor adjustment module 407 is used to adjust the distribution of sensors on the target component according to the sensor density. The sensors include: temperature sensors and flow rate sensors.

[0161] In some possible implementations, the density determination module 406 is further configured to:

[0162] The temperature change rate of the target component is converted into the temperature sensor density according to the temperature sensitivity coefficient; the flow velocity change rate of the target component is converted into the flow velocity sensor density according to the flow velocity sensitivity coefficient; the sensor density is determined according to the temperature sensor density and the flow velocity sensor density, and the sensor density is positively correlated with the temperature sensor density and the flow velocity sensor density.

[0163] For the description of the features in the embodiment corresponding to the above-mentioned server cooling control device, reference can be made to the relevant description of the embodiment corresponding to the server cooling control method, which will not be repeated here.

[0164] Figure 6 This is a schematic diagram of the structure of the electronic device provided in this application. Figure 6 As shown, the electronic device 50 provided in this embodiment includes: at least one processor 501 and a memory 502. Optionally, the electronic device 50 further includes a communication component 503. The processor 501, the memory 502 and the communication component 503 are connected via a bus.

[0165] During the specific implementation process, at least one processor 501 executes the computer-executable instructions stored in the memory 502 , so that the at least one processor 501 executes the above-mentioned server cooling control method embodiment.

[0166] The specific implementation process of the processor 501 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0167] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the application may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0168] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0169] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0170] An embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored. The computer program is configured to execute the steps of any of the above-mentioned server cooling control method embodiments when running.

[0171] In an exemplary embodiment, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk, or an optical disk.

[0172] An embodiment of the present application further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of any of the above-mentioned server cooling control method embodiments are implemented.

[0173] An embodiment of the present application further provides another computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above-mentioned server cooling control method embodiments are implemented.

[0174] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0175] The above is a detailed introduction to a server cooling control method and related equipment provided by this application. This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. It should be pointed out that for ordinary technicians in this technical field, without departing from the principles of this application, several improvements and modifications can be made to this application, and these improvements and modifications also fall within the scope of protection of the claims of this application.

Claims

1. A server cooling control method, characterized in that: include: Obtaining a measurement result sequence of a target component at a current moment on a server, the measurement result sequence including one or more measurement result information, the measurement result information including: a measurement time, a measured temperature, and a measured flow rate of a cooling medium, the measurement time being the current moment or a historical moment; Predicting the amount of heat to be removed required for the target component to reach a target temperature based on the sequence of measurement results at the current moment, and determining an average amount of heat to be removed based on the amount of heat to be removed predicted at multiple moments; Calculating a temperature difference between the measured temperature at the current moment and the target temperature, and a heat difference between the heat to be removed predicted at the current moment and the average heat to be removed; Cooling control is performed on the target component based on the temperature difference and the heat difference.

2. The method according to claim 1, characterized in that The cooling control of the target component according to the temperature difference and the heat difference includes: determining a target flow rate of the cooling medium according to the temperature difference and the heat difference; Cooling control of the target component is performed based on the target flow rate.

3. The method according to claim 2, characterized in that Determining the target flow rate of the cooling medium according to the temperature difference and the heat difference includes: converting the temperature difference into a first predicted flow rate; converting the heat difference into a second predicted flow rate; A target flow rate of the cooling medium is determined according to the first predicted flow rate and the second predicted flow rate.

4. The method according to claim 3, characterized in that Converting the temperature difference into a first predicted flow rate comprises: multiplying the first conversion coefficient by the temperature difference at the current moment as the first sub-flow rate, and / or determining the second sub-flow rate by the temperature difference at the current moment and the temperature differences at multiple historical moments before the current moment; The first predicted flow rate is determined according to the first sub-flow rate and / or the second sub-flow rate.

5. The method according to claim 4, characterized in that The determining the second sub-flow rate by using the temperature difference at the current moment and the temperature differences at multiple historical moments before the current moment includes: The temperature difference integral term at the previous moment is updated by the temperature difference at the current moment to obtain the temperature difference integral term at the current moment; Determining a rate of change of the temperature difference at the current moment based on the temperature difference at the current moment and the temperature difference at the previous moment; The second sub-flow rate is determined according to the temperature difference at the current moment, the temperature difference integral term at the current moment, and the temperature difference change rate at the current moment.

6. The method according to claim 5, characterized in that The determining the second sub-flow rate according to the temperature difference at the current moment, the temperature difference integral term at the current moment, and the temperature difference change rate at the current moment includes: weighting the temperature difference at the current moment, the temperature difference integral term, and the temperature difference change rate to obtain a control parameter; The control parameter is mapped to the second sub-flow rate according to the maximum flow rate, and the second sub-flow rate is positively correlated with the maximum flow rate and the control parameter.

7. The method according to claim 4, characterized in that Determining the first predicted flow rate according to the first sub-flow rate and / or the second sub-flow rate includes: The first sub-flow rate and the second sub-flow rate are weighted to obtain the first predicted flow rate.

8. The method according to claim 3, characterized in that Converting the heat difference into a second predicted flow rate comprises: The product of the second conversion coefficient and the heat difference is used as the second predicted flow rate.

9. The method according to claim 1, characterized in that The step of predicting the amount of heat to be removed for the target component to reach a target temperature based on the sequence of measurement results at the current moment includes: Calculating statistical information based on the sequence of measurement results at the current moment, the statistical information including at least one of the following: an average measured temperature, a maximum temperature difference, and a standard deviation of the measured flow rate; The amount of heat to be removed required for the target component to reach a target temperature is predicted based on the measurement result sequence at the current moment and the statistical information.

10. The method according to claim 9, characterized in that The predicting, based on the current measurement result sequence and the statistical information, the amount of heat to be removed required for the target component to reach the target temperature includes: splicing the measurement result sequence and the statistical information into a feature vector of the target component; The characteristic vector is input into a long short-term memory network model to obtain the heat to be removed required for the target component to reach the target temperature.

11. The method according to claim 1, wherein The method further comprises: Obtaining a temperature change rate and a flow rate change rate of the target component; determining a sensor density on the target component according to a temperature change rate of the target component and a flow rate change rate, wherein the sensor density is positively correlated with the temperature change rate and the flow rate change rate; The distribution of sensors on the target component is adjusted according to the sensor density, wherein the sensors include: a temperature sensor and a flow rate sensor.

12. The method according to claim 11, characterized in that Determining the sensor density on the target component according to the temperature change rate and the flow rate change rate of the target component includes: converting the temperature change rate of the target component into a temperature sensor density according to a temperature sensitivity coefficient; converting the flow velocity change rate of the target component into a flow velocity sensor density according to a flow velocity sensitivity coefficient; The sensor density is determined according to the temperature sensor density and the flow rate sensor density, and the sensor density is positively correlated to the temperature sensor density and the flow rate sensor density.

13. An electronic device, characterized in that: include: memory for storing computer programs; A processor, configured to implement the steps of the server cooling control method according to any one of claims 1 to 12 when executing the computer program.

14. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the server cooling control method according to any one of claims 1 to 12 are implemented.

15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the server cooling control method according to any one of claims 1 to 12 are implemented.

Images