A server cabinet cooling unit fault early warning system and method

Abstract

The application discloses a kind of server cabinet cooling unit failure early warning system and method, it is related to the technical field of intelligent early warning, including the following steps, setting variable and invariant, calculate historical temperature change rate, establish first mapping relationship, second mapping relationship and third mapping relationship, calculate temperature margin, construct temperature interval, judge first cooling state, send temperature early warning signal.The application sets up temperature monitoring point and identifies cabinet model to match rated temperature, improves the accuracy of fault detection, uses historical data for first screening and second screening, reduces false positives and false negatives, establishes mapping relationship, so that the early warning system is more flexible and adaptable, by constructing temperature interval to judge cooling state, potential cooling problems are found in advance and early warning signal is sent, by making full use of historical data and current data, improve the utilization efficiency of data, provide data support for the maintenance and optimization of cabinet cooling unit.

Description

Technical Field

[0001] This invention relates to the technical field of intelligent early warning, and in particular to a server rack cooling unit fault early warning system and method. Background Technology

[0002] In recent years, the server rack industry has been addressing the high heat density challenges of data centers by adopting high-density designs, hot aisle containment, and intelligent temperature control systems. Server racks will place greater emphasis on intelligence and modularity, with intelligence being reflected in the integration of IoT and AI technologies to achieve real-time monitoring and early warning of rack status.

[0003] Currently, Chinese invention patent CN115774650A discloses an overheating fault identification device, method, apparatus, medium, and server rack. This method obtains the sound signal energy value by extracting and analyzing the sound signal features. When the sound signal energy value is within the preset energy range of the overheating fault, the location and temperature of the overheating fault are obtained by combining the installation position of the sound wave sensor and the sound signal energy value to identify the overheating fault. However, the related technology does not construct the influence of various external parameters and the rack's own parameters on the cooling function based on historical data, thus failing to provide accurate temperature warnings. This can easily lead to incorrect judgments and waste of time resources. Furthermore, the warning analysis process is not simplified based on the actual measured temperature, which is not conducive to the rapid response of the warning analysis and has certain limitations. Summary of the Invention

[0004] The technical problem solved by this invention is that related technologies do not construct the impact of various external parameters and the cabinet's own parameters on the cooling function based on historical data, so as to accurately warn of temperature. This can easily lead to incorrect judgments and waste of time resources. Furthermore, the process of early warning analysis is not simplified based on the actual measured temperature, which is not conducive to the rapid response of early warning analysis and has certain limitations.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: Firstly, a method for early warning of server rack cooling unit failures, comprising the following steps:

[0006] Step S100: Set temperature monitoring points, identify the rack model, and match the rated temperature according to the rack model;

[0007] Step S200: Obtain historical data, perform a first screening on the historical data based on the rated temperature to obtain first data, set variables and invariants according to the controlled variable method, and perform a second screening on the first data based on the variables and invariants to obtain second data, and calculate the historical temperature change rate based on the second data;

[0008] The second data filtering methods include:

[0009] Historical data, including historical server power, historical ambient temperature, and historical wind pressure, are set as parameters to be set. Any parameter to be set is set as a variable, and the remaining parameters to be set are set as invariants. The variables and their corresponding invariants are set as reference groups, which include a first reference group, a second reference group, and a third reference group. The rules for the reference groups are then obtained.

[0010] Create a classification model, classify the first data according to the classification model to obtain the second data, the second data including the second data of the first reference group, the second data of the second reference group, and the second data of the third reference group;

[0011] The method for constructing the classification model includes:

[0012] Construct a data table and boundary rules, wherein the boundary rules include variable boundary rules and invariant boundary rules, wherein the variable boundary rules represent that not all variables are equal, and the invariant boundary rules represent that all invariants are equal;

[0013] Step S300: Establish a first mapping relationship, a second mapping relationship, and a third mapping relationship based on the historical temperature change rate;

[0014] Obtain the historical average surface temperature of the rack corresponding to the second data. The historical average surface temperature of the rack corresponding to the second data is represented as a time series value. The time interval between adjacent historical average surface temperatures of the rack is recorded as the first time interval. The calculation method for the historical temperature change rate includes:

[0015] Calculate the difference between the average surface temperatures of adjacent historical cabinets in time sequence, and record it as the second difference. Calculate the ratio of the second difference to the first time period, and record it as the first ratio. Set the first ratio as the historical temperature change rate.

[0016] The first mapping relationship represents the mapping relationship between historical server power and historical temperature change rate; the second mapping relationship represents the mapping relationship between historical ambient temperature and historical temperature change rate; and the third mapping relationship represents the mapping relationship between historical wind pressure and historical temperature change rate.

[0017] The methods for constructing the first, second, and third mapping relationships include:

[0018] Obtain the historical temperature change rate corresponding to the first mapping relationship, the second mapping relationship, or the third mapping relationship; obtain the value of the variable corresponding to the historical temperature change rate; and construct the mapping relationship between the value of the variable and the historical temperature change rate.

[0019] When the variable corresponding to the first ratio is the historical server power, the corresponding mapping relationship is recorded as the first mapping relationship; when the variable corresponding to the first ratio is the historical ambient temperature, the corresponding mapping relationship is recorded as the second mapping relationship; when the variable corresponding to the first ratio is the historical wind pressure, the corresponding mapping relationship is recorded as the third mapping relationship.

[0020] Step S400: Obtain current data, including current server power, current ambient temperature and current wind pressure; calculate temperature margin based on current data; construct temperature range based on temperature margin; determine first cooling state based on temperature range; and send temperature warning signal based on first cooling state.

[0021] As a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the method for setting the temperature monitoring point includes:

[0022] Obtain the 3D model of the server rack, and draw the mapping image of the server surface in the 3D model onto the server rack surface, denoted as the first image. The first image represents the mapping image of each server surface in the 3D model onto each surface of the server rack.

[0023] Obtain the intersection of the angle bisectors of the first image and the geometric center of the corresponding server, and calculate the distance from the intersection of the angle bisectors of the first image to the geometric center of the corresponding server, denoted as the first distance;

[0024] Using the intersection of the angle bisectors of the first image as the center and the first distance as the radius, a first circle is drawn. The first value is set as the interval distance, which represents the length of the interval arc. Temperature sensors are set at the intersection of the interval distances and at the corresponding center of the circle. The temperature sensors are set on the outer surface of the cabinet.

[0025] As a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the method includes: taking an image of the upper surface of the rack with an industrial camera, wherein the image of the upper surface of the rack is attached with a rack QR code, and scanning the rack QR code to identify and obtain the rack model.

[0026] Retrieve the rack database, input the rack model into the rack database, and match the rated temperature corresponding to the rack model;

[0027] Obtain historical data, perform a first filter on the historical data based on the rated temperature, and obtain the first data;

[0028] The first data filtering method includes: retrieving the operating database, inputting the rack model into the operating database, and matching the historical data corresponding to the rack model. The historical data includes historical server power, historical ambient temperature, historical wind pressure, and historical average rack surface temperature.

[0029] The historical average surface temperature of the cabinet is compared with the rated temperature. When the historical average surface temperature of the cabinet is greater than the rated temperature, the corresponding historical data is deleted. When the historical average surface temperature of the cabinet is less than or equal to the rated temperature, the corresponding historical data is set as the first data.

[0030] The rated temperature is expressed as the maximum average temperature of the cabinet surface.

[0031] As a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the second data filtering method further includes:

[0032] The reference group rules include:

[0033] The first reference group sets historical server power as a variable and historical ambient temperature and historical wind pressure as invariants;

[0034] The second reference group sets historical ambient temperature as a variable and historical server power and historical wind pressure as invariants;

[0035] The third reference group sets historical wind pressure as a variable and historical server power and historical ambient temperature as invariants;

[0036] The classification model includes a first reference group classification model, a second reference group classification model, and a third reference group classification model. The second data of the first reference group, the second data of the second reference group, and the second data of the third reference group are obtained through the first reference group classification model, the second reference group classification model, and the third reference group classification model, respectively.

[0037] As a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the method for constructing the classification model further includes:

[0038] A classification model is constructed by setting the first column of the data table to the corresponding variable boundary rule and the second column of the data table to the corresponding invariant boundary rule using SQL statements.

[0039] Activate the query function and categorize the queried data into the corresponding first reference group second data, second reference group second data, or third reference group second data.

[0040] In a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the expression of the variable boundary rule is:

[0041] l1:x2-x1+x3-x2+...+x m -x m-1 ≠0

[0042] The expression for the invariant boundary rule is:

[0043]

[0044] Where l1 and l2 are the variable boundary rule and the corresponding invariant boundary rule, respectively, x m Let y be a variable. m and k m These are the corresponding invariants;

[0045] When x m When it belongs to the first reference group, x m Represented as historical server power, y m and k m This is expressed as historical ambient temperature and historical wind pressure;

[0046] When x m When it belongs to the second reference group, x m Historical ambient temperature, y m and k m This is represented as historical server power and historical wind pressure;

[0047] When x m When it belongs to the third reference group, x m Historical wind pressure, y m and k m This is represented by historical server power and historical ambient temperature.

[0048] In a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the following steps are taken: a second time period is set as the monitoring cycle, current data is acquired, including current server power, current ambient temperature, and current air pressure, and the current average surface temperature of the rack is calculated. The method for calculating the current average surface temperature of the rack includes:

[0049] Get the temperature at any time point of each monitoring point, count the number of monitoring points, calculate the sum of the temperatures at each monitoring point, calculate the ratio of the sum of the temperatures at each monitoring point to the number of monitoring points, and record it as the third ratio. Set the third ratio as the current average surface temperature of the cabinet.

[0050] The current average surface temperature of the cabinet is compared with the rated temperature to obtain a comparison result. The comparison result includes the current average surface temperature of the cabinet being greater than the rated temperature and the current average surface temperature of the cabinet being less than or equal to the rated temperature. Based on the comparison result, a first operation is performed, which includes sending an early warning signal and analyzing the current data.

[0051] If the average surface temperature of the current cabinet is higher than the rated temperature, a temperature warning signal will be sent; otherwise, the current data will be analyzed.

[0052] As a preferred embodiment of the server rack cooling unit fault early warning method of the present invention, the method for calculating the temperature margin includes:

[0053] When the first operation is to analyze the current data, the current server power, current ambient temperature and current wind pressure are obtained. The current server power, current ambient temperature and current wind pressure are input into the first mapping relationship, the second mapping relationship and the third mapping relationship respectively to obtain the corresponding historical temperature change rate, which are recorded as the first current temperature change rate, the second temperature change rate and the third temperature change rate respectively.

[0054] Obtain the current average surface temperature of the rack, calculate the product of the current average surface temperature of the rack with the first current temperature change rate, the second temperature change rate, and the third temperature change rate, and denot them as the first product, the second product, and the third product. Set the first product, the second product, and the third product as the temperature margin, and select the maximum value of the temperature margin.

[0055] Calculate the sum of the maximum values ​​of the rated temperature and the temperature margin, and denote it as the first sum;

[0056] Calculate the difference between the rated temperature and the maximum value of the temperature margin, and denote it as the first difference. Set the first sum as the upper limit of the temperature range and the first difference as the lower limit of the temperature range to construct the temperature range.

[0057] The current average surface temperature of the rack is compared with the temperature range to determine the first cooling state, which includes normal state and fault state. A temperature warning signal is sent according to the first cooling state.

[0058] A temperature warning signal is sent when the first cooling state is in a fault state; no temperature warning signal is sent when the first cooling state is in a normal state.

[0059] Secondly, a server rack cooling unit fault early warning system includes a data acquisition module, an analysis module, and an early warning module;

[0060] The acquisition module is used to set temperature monitoring points, identify the cabinet model, match the rated temperature according to the cabinet model, acquire historical data, perform a first filter on the historical data according to the rated temperature to obtain first data, set variables and invariants according to the control variable method, and perform a second filter on the first data according to the variables and invariants to obtain second data.

[0061] The second data filtering methods include:

[0062] Historical data, including historical server power, historical ambient temperature, and historical wind pressure, are set as parameters to be set. Any parameter to be set is set as a variable, and the remaining parameters to be set are set as invariants. The variables and their corresponding invariants are set as reference groups, which include a first reference group, a second reference group, and a third reference group. The rules for the reference groups are then obtained.

[0063] Create a classification model, classify the first data according to the classification model to obtain the second data, the second data including the second data of the first reference group, the second data of the second reference group, and the second data of the third reference group;

[0064] The method for constructing the classification model includes:

[0065] Construct a data table and boundary rules, wherein the boundary rules include variable boundary rules and invariant boundary rules, wherein the variable boundary rules represent that not all variables are equal, and the invariant boundary rules represent that all invariants are equal;

[0066] The analysis module calculates the historical temperature change rate based on the second data, and establishes a first mapping relationship, a second mapping relationship, and a third mapping relationship based on the historical temperature change rate.

[0067] Obtain the historical average surface temperature of the rack corresponding to the second data. The historical average surface temperature of the rack corresponding to the second data is represented as a time series value. The time interval between adjacent historical average surface temperatures of the rack is recorded as the first time interval. The calculation method for the historical temperature change rate includes:

[0068] Calculate the difference between the average surface temperatures of adjacent historical cabinets in time sequence, and record it as the second difference. Calculate the ratio of the second difference to the first time period, and record it as the first ratio. Set the first ratio as the historical temperature change rate.

[0069] The first mapping relationship represents the mapping relationship between historical server power and historical temperature change rate; the second mapping relationship represents the mapping relationship between historical ambient temperature and historical temperature change rate; and the third mapping relationship represents the mapping relationship between historical wind pressure and historical temperature change rate.

[0070] The methods for constructing the first, second, and third mapping relationships include:

[0071] Obtain the historical temperature change rate corresponding to the first mapping relationship, the second mapping relationship, or the third mapping relationship; obtain the value of the variable corresponding to the historical temperature change rate; and construct the mapping relationship between the value of the variable and the historical temperature change rate.

[0072] When the variable corresponding to the first ratio is the historical server power, the corresponding mapping relationship is recorded as the first mapping relationship; when the variable corresponding to the first ratio is the historical ambient temperature, the corresponding mapping relationship is recorded as the second mapping relationship; when the variable corresponding to the first ratio is the historical wind pressure, the corresponding mapping relationship is recorded as the third mapping relationship.

[0073] The early warning module acquires current data, including current server power, current ambient temperature, and current wind pressure. It calculates temperature margin based on the current data, constructs a temperature range based on the temperature margin, determines a first cooling state based on the temperature range, and sends a temperature early warning signal based on the first cooling state.

[0074] The beneficial effects of this invention are as follows: By setting temperature monitoring points and identifying the rack model to match the rated temperature, the actual working status of the rack can be monitored more accurately, thereby improving the accuracy of fault detection. Historical data is used for first and second screening to ensure the representativeness and reliability of the selected data. This method effectively eliminates irrelevant data, reducing false alarms and missed alarms. By calculating historical temperature change rates and establishing mapping relationships, early warning parameters are dynamically adjusted according to the actual operating conditions of the rack, making the early warning system more flexible and adaptable. By acquiring current data in real time and calculating temperature margins, temperature ranges are constructed to determine the cooling status, potential cooling problems are detected in advance, and early warning signals are sent, thereby reducing the risk of fault occurrence. By fully utilizing historical and current data, this method improves data utilization efficiency and provides data support for the maintenance and optimization of rack cooling units. Attached Figure Description

[0075] Figure 1 This is a basic flowchart illustrating a server rack cooling unit fault early warning method according to an embodiment of the present invention. Detailed Implementation

[0076] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0077] Example, refer to Figure 1 As an embodiment of the present invention, a method for early warning of server rack cooling unit failure is provided, comprising the following steps:

[0078] Step S100: Set temperature monitoring points, identify the rack model, and match the rated temperature according to the rack model;

[0079] Step S200: Obtain historical data, perform a first screening on the historical data based on the rated temperature to obtain first data, set variables and invariants according to the controlled variable method, and perform a second screening on the first data based on the variables and invariants to obtain second data, and calculate the historical temperature change rate based on the second data;

[0080] The second data filtering methods include:

[0081] Historical data, including historical server power, historical ambient temperature, and historical wind pressure, are set as parameters to be set. Any parameter to be set is set as a variable, and the remaining parameters to be set are set as invariants. The variables and their corresponding invariants are set as reference groups, which include a first reference group, a second reference group, and a third reference group. The rules for the reference groups are then obtained.

[0082] Create a classification model, classify the first data according to the classification model to obtain the second data, the second data including the second data of the first reference group, the second data of the second reference group, and the second data of the third reference group;

[0083] The method for constructing the classification model includes:

[0084] Construct a data table and boundary rules, wherein the boundary rules include variable boundary rules and invariant boundary rules, wherein the variable boundary rules represent that not all variables are equal, and the invariant boundary rules represent that all invariants are equal;

[0085] Step S300: Establish a first mapping relationship, a second mapping relationship, and a third mapping relationship based on the historical temperature change rate;

[0086] Obtain the historical average surface temperature of the rack corresponding to the second data. The historical average surface temperature of the rack corresponding to the second data is represented as a time series value. The time interval between adjacent historical average surface temperatures of the rack is recorded as the first time interval. The calculation method for the historical temperature change rate includes:

[0087] Calculate the difference between the average surface temperatures of adjacent historical cabinets in time sequence, and record it as the second difference. Calculate the ratio of the second difference to the first time period, and record it as the first ratio. Set the first ratio as the historical temperature change rate.

[0088] The first mapping relationship represents the mapping relationship between historical server power and historical temperature change rate; the second mapping relationship represents the mapping relationship between historical ambient temperature and historical temperature change rate; and the third mapping relationship represents the mapping relationship between historical wind pressure and historical temperature change rate.

[0089] The methods for constructing the first, second, and third mapping relationships include:

[0090] Obtain the historical temperature change rate corresponding to the first mapping relationship, the second mapping relationship, or the third mapping relationship; obtain the value of the variable corresponding to the historical temperature change rate; and construct the mapping relationship between the value of the variable and the historical temperature change rate.

[0091] When the variable corresponding to the first ratio is the historical server power, the corresponding mapping relationship is recorded as the first mapping relationship; when the variable corresponding to the first ratio is the historical ambient temperature, the corresponding mapping relationship is recorded as the second mapping relationship; when the variable corresponding to the first ratio is the historical wind pressure, the corresponding mapping relationship is recorded as the third mapping relationship.

[0092] Step S400: Obtain current data, including current server power, current ambient temperature and current wind pressure; calculate temperature margin based on current data; construct temperature range based on temperature margin; determine first cooling state based on temperature range; and send temperature warning signal based on first cooling state.

[0093] This invention matches rated temperatures by setting temperature monitoring points and identifying cabinet models, enabling more accurate monitoring of the actual operating status of the cabinet and thus improving the accuracy of fault detection. Historical data is used for first and second screening to ensure the representativeness and reliability of the selected data. This method effectively eliminates irrelevant data, reducing false alarms and missed alarms. By calculating historical temperature change rates and establishing mapping relationships, warning parameters are dynamically adjusted based on the actual operating conditions of the cabinet, making the warning system more flexible and adaptable. By acquiring current data in real time and calculating temperature margins, temperature ranges are constructed to determine the cooling status, allowing for the early detection of potential cooling problems and the sending of warning signals, thereby reducing the risk of faults. By fully utilizing historical and current data, this method improves data utilization efficiency and provides data support for the maintenance and optimization of cabinet cooling units.

[0094] The method for setting up the temperature monitoring points includes:

[0095] Obtain the 3D model of the server rack, and draw the mapping image of the server surface in the 3D model onto the server rack surface, denoted as the first image. The first image represents the mapping image of each server surface in the 3D model onto each surface of the server rack.

[0096] Obtain the intersection of the angle bisectors of the first image and the geometric center of the corresponding server, and calculate the distance from the intersection of the angle bisectors of the first image to the geometric center of the corresponding server, denoted as the first distance;

[0097] Using the intersection of the angle bisectors of the first image as the center and the first distance as the radius, a first circle is drawn. The first value is set as the interval distance, which represents the length of the interval arc. Temperature sensors are set at the intersection of the interval distances and at the corresponding center of the circle. The temperature sensors are set on the outer surface of the cabinet.

[0098] In practice, by acquiring a 3D model of the server rack and mapping the server surface onto the rack surface, the location of temperature monitoring points is precisely determined. This ensures that the layout of the temperature sensors is closely related to the server's thermal distribution, improving the accuracy of temperature monitoring. By calculating the distance from the intersection of the angle bisectors of the first image to the geometric center of the corresponding server, and using this distance as the radius to draw a first circle, the layout of the temperature sensors is optimized. This ensures that the sensors can cover key hot areas, thereby more effectively monitoring temperature changes. Temperature sensors are placed at the boundaries of the intervals and at the corresponding center of the circle to ensure the continuity and uniformity of temperature monitoring. This method avoids monitoring blind spots and improves the efficiency of temperature monitoring.

[0099] An image of the upper surface of the cabinet is captured by an industrial camera. The image of the upper surface of the cabinet is accompanied by a QR code of the cabinet. The cabinet model is identified and obtained by scanning the QR code.

[0100] Retrieve the rack database, input the rack model into the rack database, and match the rated temperature corresponding to the rack model;

[0101] Obtain historical data, perform a first filter on the historical data based on the rated temperature, and obtain the first data;

[0102] The first data filtering method includes: retrieving the operating database, inputting the rack model into the operating database, and matching the historical data corresponding to the rack model. The historical data includes historical server power, historical ambient temperature, historical wind pressure, and historical average rack surface temperature.

[0103] The historical average surface temperature of the cabinet is compared with the rated temperature. When the historical average surface temperature of the cabinet is greater than the rated temperature, the corresponding historical data is deleted. When the historical average surface temperature of the cabinet is less than or equal to the rated temperature, the corresponding historical data is set as the first data.

[0104] The rated temperature is expressed as the maximum average temperature of the cabinet surface.

[0105] In practice, industrial cameras are used to capture images of the cabinet's upper surface and scan QR codes to automatically identify the cabinet model. This method reduces errors and time consumption associated with manual identification. By accessing the cabinet database and inputting the cabinet model to match the corresponding rated temperature, the accuracy of temperature matching is ensured. This precise matching helps improve the accuracy of temperature monitoring. Historical data is acquired and initially filtered based on the rated temperature, deleting historical data that does not meet the criteria and retaining valid data. This filtering method improves data availability and accuracy. By comparing the historical average cabinet surface temperature with the rated temperature, temperature anomalies are detected earlier, thereby enhancing fault early warning performance and facilitating timely measures to prevent failures.

[0106] The second data filtering method also includes:

[0107] The reference group rules include:

[0108] The first reference group sets historical server power as a variable and historical ambient temperature and historical wind pressure as invariants;

[0109] The second reference group sets historical ambient temperature as a variable and historical server power and historical wind pressure as invariants;

[0110] The third reference group sets historical wind pressure as a variable and historical server power and historical ambient temperature as invariants;

[0111] The classification model includes a first reference group classification model, a second reference group classification model, and a third reference group classification model. The second data of the first reference group, the second data of the second reference group, and the second data of the third reference group are obtained through the first reference group classification model, the second reference group classification model, and the third reference group classification model, respectively.

[0112] In practice, by calculating historical temperature change rates, temperature data is analyzed in more detail to identify trends and patterns of temperature changes. Server power, ambient temperature, and wind pressure in the historical data are set as variables and invariants, respectively, forming different reference groups. This method helps to isolate the influence of single variables and more accurately assess the contribution of each parameter to temperature changes. Classification models are created for different reference groups, and the influence of different variables is specifically analyzed and classified, thereby improving the model's relevance and accuracy. The first data is classified by the classification model to obtain the second data. This method makes more effective use of historical data, providing more accurate data support for fault early warning. Through effective data filtering and fault early warning, the reliability of the server rack cooling system is improved, and system failures caused by temperature issues are reduced.

[0113] The method for constructing the classification model also includes:

[0114] A classification model is constructed by setting the first column of the data table to the corresponding variable boundary rule and the second column of the data table to the corresponding invariant boundary rule using SQL statements.

[0115] Activate the query function and categorize the queried data into the corresponding first reference group second data, second reference group second data, or third reference group second data.

[0116] The expression for the variable boundary rule is:

[0117] l1:x2-x1+x3-x2+...+x m -x m-1 ≠0

[0118] The expression for the invariant boundary rule is:

[0119]

[0120] Where l1 and l2 are the variable boundary rule and the corresponding invariant boundary rule, respectively, x m Let y be a variable. m and k m These are the corresponding invariants;

[0121] When x m When it belongs to the first reference group, x m Represented as historical server power, y m and k m This is expressed as historical ambient temperature and historical wind pressure;

[0122] When x m When it belongs to the second reference group, x m Historical ambient temperature, y m and k m This is represented as historical server power and historical wind pressure;

[0123] When x m When it belongs to the third reference group, x m Historical wind pressure, y m and k m This is represented by historical server power and historical ambient temperature.

[0124] In practice, by constructing data tables and boundary rules, and clearly distinguishing between variable boundary rules and invariant boundary rules, it is helpful to accurately identify and classify data during data processing. Using SQL statements, the first column of the data table is set as the corresponding variable boundary rule, and the second column as the corresponding invariant boundary rule, efficiently building a classification model. The powerful functionality of SQL queries quickly processes and classifies large amounts of data. By initiating the query function, the queried data is precisely categorized into the corresponding first reference group second data, second reference group second data, or third reference group second data. This precise classification facilitates subsequent data analysis and fault prediction.

[0125] In practice, the historical temperature change rate is accurately calculated by measuring the difference (second difference) between the average surface temperatures of adjacent historical server racks in chronological order and their ratio to the first time period (first ratio). This method ensures the accuracy of the temperature change rate, providing a reliable data foundation for subsequent analysis. It also establishes mapping relationships (first, second, and third mapping relationships) between historical server power, historical ambient temperature, historical wind pressure, and the historical temperature change rate, which helps analyze the impact of different variables on temperature changes. This mapping relationship construction method improves the accuracy of fault early warning. By clarifying objectives, introducing advanced monitoring technologies, establishing a data analysis platform, and setting early warning thresholds and rules, the intelligence level and predictive accuracy of the early warning mechanism are continuously improved.

[0126] The second time period is set as the monitoring cycle, and current data is acquired, including current server power, current ambient temperature, and current wind pressure. The current average surface temperature of the server rack is calculated, and the calculation method for the current average surface temperature of the server rack includes:

[0127] Get the temperature at any time point of each monitoring point, count the number of monitoring points, calculate the sum of the temperatures at each monitoring point, calculate the ratio of the sum of the temperatures at each monitoring point to the number of monitoring points, and record it as the third ratio. Set the third ratio as the current average surface temperature of the cabinet.

[0128] The current average surface temperature of the cabinet is compared with the rated temperature to obtain a comparison result. The comparison result includes the current average surface temperature of the cabinet being greater than the rated temperature and the current average surface temperature of the cabinet being less than or equal to the rated temperature. Based on the comparison result, a first operation is performed, which includes sending an early warning signal and analyzing the current data.

[0129] If the average surface temperature of the current cabinet is higher than the rated temperature, a temperature warning signal will be sent; otherwise, the current data will be analyzed.

[0130] The method for calculating the temperature margin includes:

[0131] When the first operation is to analyze the current data, the current server power, current ambient temperature and current wind pressure are obtained. The current server power, current ambient temperature and current wind pressure are input into the first mapping relationship, the second mapping relationship and the third mapping relationship respectively to obtain the corresponding historical temperature change rate, which are recorded as the first current temperature change rate, the second temperature change rate and the third temperature change rate respectively.

[0132] Obtain the current average surface temperature of the rack, calculate the product of the current average surface temperature of the rack with the first current temperature change rate, the second temperature change rate, and the third temperature change rate, and denot them as the first product, the second product, and the third product. Set the first product, the second product, and the third product as the temperature margin, and select the maximum value of the temperature margin.

[0133] Calculate the sum of the maximum values ​​of the rated temperature and the temperature margin, and denote it as the first sum;

[0134] Calculate the difference between the rated temperature and the maximum value of the temperature margin, and denote it as the first difference. Set the first sum as the upper limit of the temperature range and the first difference as the lower limit of the temperature range to construct the temperature range.

[0135] The current average surface temperature of the rack is compared with the temperature range to determine the first cooling state, which includes normal state and fault state. A temperature warning signal is sent according to the first cooling state.

[0136] A temperature warning signal is sent when the first cooling state is in a fault state; no temperature warning signal is sent when the first cooling state is in a normal state.

[0137] In practice, by setting a monitoring period (second time period) and acquiring current data (current server power, current ambient temperature, and current wind pressure) in real time, the surface temperature of the server rack is monitored and analyzed in real time. By acquiring the temperature at each monitoring point, counting the number of monitoring points, calculating the sum of the temperatures and their ratio to the number of monitoring points (third ratio), the average surface temperature of the current server rack is accurately calculated. This method ensures the accuracy of temperature calculation. By comparing the current average surface temperature of the server rack with the rated temperature and performing the first operation (sending an early warning signal or analyzing the current data) based on the comparison result, abnormal temperatures can be detected promptly and measures can be taken to help prevent malfunctions.

[0138] This invention matches rated temperatures by setting temperature monitoring points and identifying cabinet models, enabling more accurate monitoring of the actual operating status of the cabinet and thus improving the accuracy of fault detection. Historical data is used for first and second screening to ensure the representativeness and reliability of the selected data. This method effectively eliminates irrelevant data, reducing false alarms and missed alarms. By calculating historical temperature change rates and establishing mapping relationships, warning parameters are dynamically adjusted based on the actual operating conditions of the cabinet, making the warning system more flexible and adaptable. By acquiring current data in real time and calculating temperature margins, temperature ranges are constructed to determine the cooling status, allowing for the early detection of potential cooling problems and the sending of warning signals, thereby reducing the risk of faults. By fully utilizing historical and current data, this method improves data utilization efficiency and provides data support for the maintenance and optimization of cabinet cooling units.

[0139] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium is implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention should all be covered within the scope of the claims of the present invention.

Claims

1. A method for early warning of server rack cooling unit failure, characterized in that, Includes the following steps: Step S100: Set temperature monitoring points, identify the rack model, and match the rated temperature according to the rack model; Step S200: Obtain historical data, perform a first screening on the historical data based on the rated temperature to obtain first data, set variables and invariants according to the controlled variable method, and perform a second screening on the first data based on the variables and invariants to obtain second data, and calculate the historical temperature change rate based on the second data; The second data filtering methods include: Historical data, including historical server power, historical ambient temperature, and historical wind pressure, are set as parameters to be set. Any parameter to be set is set as a variable, and the remaining parameters to be set are set as invariants. The variables and their corresponding invariants are set as reference groups, which include a first reference group, a second reference group, and a third reference group. The rules for the reference groups are then obtained. Create a classification model, classify the first data according to the classification model to obtain the second data, the second data including the second data of the first reference group, the second data of the second reference group, and the second data of the third reference group; The method for constructing the classification model includes: Construct a data table and boundary rules, wherein the boundary rules include variable boundary rules and invariant boundary rules, wherein the variable boundary rules represent that not all variables are equal, and the invariant boundary rules represent that all invariants are equal; Step S300: Establish a first mapping relationship, a second mapping relationship, and a third mapping relationship based on the historical temperature change rate; Obtain the historical average surface temperature of the rack corresponding to the second data. The historical average surface temperature of the rack corresponding to the second data is represented as a time series value. The time interval between adjacent historical average surface temperatures of the rack is recorded as the first time interval. The calculation method for the historical temperature change rate includes: Calculate the difference between the average surface temperatures of adjacent historical cabinets in time sequence, and record it as the second difference. Calculate the ratio of the second difference to the first time period, and record it as the first ratio. Set the first ratio as the historical temperature change rate. The first mapping relationship represents the mapping relationship between historical server power and historical temperature change rate; the second mapping relationship represents the mapping relationship between historical ambient temperature and historical temperature change rate; and the third mapping relationship represents the mapping relationship between historical wind pressure and historical temperature change rate. The methods for constructing the first, second, and third mapping relationships include: Obtain the historical temperature change rate corresponding to the first mapping relationship, the second mapping relationship, or the third mapping relationship; obtain the value of the variable corresponding to the historical temperature change rate; and construct the mapping relationship between the value of the variable and the historical temperature change rate. When the variable corresponding to the first ratio is the historical server power, the corresponding mapping relationship is recorded as the first mapping relationship; when the variable corresponding to the first ratio is the historical ambient temperature, the corresponding mapping relationship is recorded as the second mapping relationship; when the variable corresponding to the first ratio is the historical wind pressure, the corresponding mapping relationship is recorded as the third mapping relationship. Step S400: Obtain current data, including current server power, current ambient temperature and current wind pressure; calculate temperature margin based on current data; construct temperature range based on temperature margin; determine first cooling state based on temperature range; and send temperature warning signal based on first cooling state.

2. The server rack cooling unit fault early warning method as described in claim 1, characterized in that: The method for setting up the temperature monitoring points includes: Obtain the 3D model of the server rack, and draw the mapping image of the server surface in the 3D model onto the server rack surface, denoted as the first image. The first image represents the mapping image of each server surface in the 3D model onto each surface of the server rack. Obtain the intersection of the angle bisectors of the first image and the geometric center of the corresponding server, and calculate the distance from the intersection of the angle bisectors of the first image to the geometric center of the corresponding server, denoted as the first distance; Using the intersection of the angle bisectors of the first image as the center and the first distance as the radius, a first circle is drawn. The first value is set as the interval distance, which represents the length of the interval arc. Temperature sensors are set at the intersection of the interval distances and at the corresponding center of the circle. The temperature sensors are set on the outer surface of the cabinet.

3. The server rack cooling unit fault early warning method as described in claim 1, characterized in that: An image of the upper surface of the cabinet is captured by an industrial camera. The image of the upper surface of the cabinet is accompanied by a QR code of the cabinet. The cabinet model is identified and obtained by scanning the QR code. Retrieve the rack database, input the rack model into the rack database, and match the rated temperature corresponding to the rack model; Obtain historical data, perform a first filter on the historical data based on the rated temperature, and obtain the first data; The first data filtering method includes: retrieving the operating database, inputting the rack model into the operating database, and matching the historical data corresponding to the rack model. The historical data includes historical server power, historical ambient temperature, historical wind pressure, and historical average rack surface temperature. The historical average surface temperature of the cabinet is compared with the rated temperature. When the historical average surface temperature of the cabinet is greater than the rated temperature, the corresponding historical data is deleted. When the historical average surface temperature of the cabinet is less than or equal to the rated temperature, the corresponding historical data is set as the first data. The rated temperature is expressed as the maximum average temperature of the cabinet surface.

4. The server rack cooling unit fault early warning method as described in claim 1, characterized in that: The second data filtering method also includes: The reference group rules include: The first reference group sets historical server power as a variable and historical ambient temperature and historical wind pressure as invariants; The second reference group sets historical ambient temperature as a variable and historical server power and historical wind pressure as invariants; The third reference group sets historical wind pressure as a variable and historical server power and historical ambient temperature as invariants; The classification model includes a first reference group classification model, a second reference group classification model, and a third reference group classification model. The second data of the first reference group, the second data of the second reference group, and the second data of the third reference group are obtained through the first reference group classification model, the second reference group classification model, and the third reference group classification model, respectively.

5. The server rack cooling unit fault early warning method as described in claim 4, characterized in that: The method for constructing the classification model also includes: A classification model is constructed by setting the first column of the data table to the corresponding variable boundary rule and the second column of the data table to the corresponding invariant boundary rule using SQL statements. Activate the query function and categorize the queried data into the corresponding first reference group second data, second reference group second data, or third reference group second data.

6. The server rack cooling unit fault early warning method as described in claim 5, characterized in that: The expression for the variable boundary rule is: l1:x2-x1+x3-x2+...+x m -x m-1 ≠0 The expression for the invariant boundary rule is: Where l1 and l2 are the variable boundary rule and the corresponding invariant boundary rule, respectively, x m Let y be a variable. m and k m These are the corresponding invariants; When x m When it belongs to the first reference group, x m Represented as historical server power, y m and k m This is expressed as historical ambient temperature and historical wind pressure; When x m When it belongs to the second reference group, x m Historical ambient temperature, y m and k m This is represented as historical server power and historical wind pressure; When x m When it belongs to the third reference group, x m Historical wind pressure, y m and k m This is represented by historical server power and historical ambient temperature.

7. The server rack cooling unit fault early warning method as described in claim 1, characterized in that: The second time period is set as the monitoring cycle, and current data is acquired, including current server power, current ambient temperature, and current wind pressure. The current average surface temperature of the server rack is calculated, and the calculation method for the current average surface temperature of the server rack includes: Get the temperature at any time point of each monitoring point, count the number of monitoring points, calculate the sum of the temperatures at each monitoring point, calculate the ratio of the sum of the temperatures at each monitoring point to the number of monitoring points, and record it as the third ratio. Set the third ratio as the current average surface temperature of the cabinet. The current average surface temperature of the cabinet is compared with the rated temperature to obtain a comparison result. The comparison result includes the current average surface temperature of the cabinet being greater than the rated temperature and the current average surface temperature of the cabinet being less than or equal to the rated temperature. Based on the comparison result, a first operation is performed, which includes sending an early warning signal and analyzing the current data. If the average surface temperature of the current cabinet is higher than the rated temperature, a temperature warning signal will be sent; otherwise, the current data will be analyzed.

8. The server rack cooling unit fault early warning method as described in claim 1, characterized in that: The method for calculating the temperature margin includes: When the first operation is to analyze the current data, the current server power, current ambient temperature and current wind pressure are obtained. The current server power, current ambient temperature and current wind pressure are input into the first mapping relationship, the second mapping relationship and the third mapping relationship respectively to obtain the corresponding historical temperature change rate, which are recorded as the first current temperature change rate, the second temperature change rate and the third temperature change rate respectively. Obtain the current average surface temperature of the rack, calculate the product of the current average surface temperature of the rack with the first current temperature change rate, the second temperature change rate, and the third temperature change rate, and denot them as the first product, the second product, and the third product. Set the first product, the second product, and the third product as the temperature margin, and select the maximum value of the temperature margin. Calculate the sum of the maximum values ​​of the rated temperature and the temperature margin, and denote it as the first sum; Calculate the difference between the rated temperature and the maximum value of the temperature margin, and denote it as the first difference. Set the first sum as the upper limit of the temperature range and the first difference as the lower limit of the temperature range to construct the temperature range. The current average surface temperature of the rack is compared with the temperature range to determine the first cooling state, which includes normal state and fault state. A temperature warning signal is sent according to the first cooling state. A temperature warning signal is sent when the first cooling state is in a fault state; no temperature warning signal is sent when the first cooling state is in a normal state.

9. A server rack cooling unit fault early warning system, characterized in that, It includes a data acquisition module, an analysis module, and an early warning module; The acquisition module is used to set temperature monitoring points, identify the cabinet model, match the rated temperature according to the cabinet model, acquire historical data, perform a first filter on the historical data according to the rated temperature to obtain first data, set variables and invariants according to the control variable method, and perform a second filter on the first data according to the variables and invariants to obtain second data. The second data filtering methods include: Historical data, including historical server power, historical ambient temperature, and historical wind pressure, are set as parameters to be set. Any parameter to be set is set as a variable, and the remaining parameters to be set are set as invariants. The variables and their corresponding invariants are set as reference groups, which include a first reference group, a second reference group, and a third reference group. The rules for the reference groups are then obtained. Create a classification model, classify the first data according to the classification model to obtain the second data, the second data including the second data of the first reference group, the second data of the second reference group, and the second data of the third reference group; The method for constructing the classification model includes: Construct a data table and boundary rules, wherein the boundary rules include variable boundary rules and invariant boundary rules, wherein the variable boundary rules represent that not all variables are equal, and the invariant boundary rules represent that all invariants are equal; The analysis module calculates the historical temperature change rate based on the second data, and establishes a first mapping relationship, a second mapping relationship, and a third mapping relationship based on the historical temperature change rate. Obtain the historical average surface temperature of the rack corresponding to the second data. The historical average surface temperature of the rack corresponding to the second data is represented as a time series value. The time interval between adjacent historical average surface temperatures of the rack is recorded as the first time interval. The calculation method for the historical temperature change rate includes: Calculate the difference between the average surface temperatures of adjacent historical cabinets in time sequence, and record it as the second difference. Calculate the ratio of the second difference to the first time period, and record it as the first ratio. Set the first ratio as the historical temperature change rate. The first mapping relationship represents the mapping relationship between historical server power and historical temperature change rate; the second mapping relationship represents the mapping relationship between historical ambient temperature and historical temperature change rate; and the third mapping relationship represents the mapping relationship between historical wind pressure and historical temperature change rate. The methods for constructing the first, second, and third mapping relationships include: Obtain the historical temperature change rate corresponding to the first mapping relationship, the second mapping relationship, or the third mapping relationship; obtain the value of the variable corresponding to the historical temperature change rate; and construct the mapping relationship between the value of the variable and the historical temperature change rate. When the variable corresponding to the first ratio is the historical server power, the corresponding mapping relationship is recorded as the first mapping relationship; when the variable corresponding to the first ratio is the historical ambient temperature, the corresponding mapping relationship is recorded as the second mapping relationship; when the variable corresponding to the first ratio is the historical wind pressure, the corresponding mapping relationship is recorded as the third mapping relationship. The early warning module acquires current data, including current server power, current ambient temperature, and current wind pressure. It calculates temperature margin based on the current data, constructs a temperature range based on the temperature margin, determines a first cooling state based on the temperature range, and sends a temperature early warning signal based on the first cooling state.

Images