Water cooling system intelligent operation and maintenance management system based on cloud platform

Abstract

The invention relates to the technical field of system management, in particular to a water cooling system intelligent operation and maintenance management system based on a cloud platform, the system comprises a processor and a memory, and the processor executes a computer program stored in the memory to realize the following steps: acquiring a coefficient adjustment factor fluctuating to-be-analyzed data and a first coefficient of the to-be-analyzed data, according to the first coefficient of each piece of to-be-analyzed data and the coefficient adjustment factor of the fluctuation to-be-analyzed data, obtaining a second coefficient of each piece of to-be-analyzed data in the to-be-analyzed data sequence, and according to the second coefficient of the to-be-analyzed data and the prediction model, obtaining predicted cooling liquid temperature data of the water cooling equipment at a future monitoring moment; and adjusting the working power of the water cooling equipment in the water cooling system according to the predicted cooling liquid temperature data. According to the invention, the reliability or effectiveness of subsequent adjustment of the working power of the water-cooling equipment can be ensured, that is, intelligent operation and maintenance management can be reliably and effectively carried out on the water-cooling system.

Description

Technical Field

[0001] The present invention relates to the field of system management technology, and particularly to an intelligent operation and maintenance management system for a water-cooling system based on a cloud platform. Background Art

[0002] The water-cooling system achieves the purpose of cooling or heat dissipation by circulating liquid to absorb heat. And since intelligent operation and maintenance management of the water-cooling system is crucial for improving equipment performance, reducing energy consumption, reducing faults and maintenance costs, and improving operation and maintenance efficiency, etc., during the working process of the water-cooling system, intelligent operation and maintenance management of the water-cooling system is required.

[0003] However, in the prior art, generally, based on the water-cooling liquid temperature sequence collected in the water-cooling equipment, future data is predicted, and the working power of the water-cooling equipment is adjusted according to the predicted future data and the water-cooling liquid temperature collected in real time currently. And the water-cooling liquid temperature collected in the water-cooling equipment will be transmitted to the cloud platform, and the water-cooling system will also be operation and maintenance managed based on the cloud platform. And adjusting the working power of the water-cooling equipment according to the predicted future data and the water-cooling liquid temperature collected in real time currently is one aspect of the operation and maintenance management of the water-cooling system. And a set of water-cooling equipment consists of a cooling tower, a cooling water circulation pump, and a centrifugal chiller. However, when obtaining the water-cooling liquid temperature sequence, local uneven flow in the coolant pipeline will cause some data with large fluctuations to appear in the obtained water-cooling liquid temperature sequence. And when there are data with large fluctuations caused by local uneven flow in the coolant pipeline in the obtained water-cooling liquid temperature sequence, it will cause the problem of low reliability of the prediction result when using the water-cooling liquid temperature sequence to predict the water-cooling liquid temperature at the future monitoring moment. And when the reliability of the prediction result is low, when adjusting the working power of the water-cooling equipment based on the prediction result subsequently, problems such as unreliable adjustment or poor adjustment effect will also occur, that is, the problem of being unable to reliably or effectively perform intelligent operation and maintenance management on the water-cooling system will occur, such as possibly increasing energy consumption. Therefore, how to improve the reliability of predicting the water-cooling liquid temperature under future monitoring to ensure reliable and effective intelligent operation and maintenance management of the water-cooling system subsequently has become an urgent problem to be solved, or how to improve the reliability of predicting the water-cooling liquid temperature under future monitoring to ensure the reliability or effectiveness when adjusting the working power of the water-cooling equipment has become an urgent problem to be solved. Summary of the Invention

[0004] In order to solve the above problems, the present invention provides an intelligent operation and maintenance management system for a water-cooling system based on a cloud platform, and the specific technical solution adopted is as follows:

[0005] An embodiment of the present invention provides an intelligent operation and maintenance management system for a water-cooling system based on a cloud platform, including a processor and a memory. The processor executes the computer program stored in the memory to implement the following steps:

[0006] In the current monitoring time period, obtain the temperature data sequence corresponding to each water-cooling device in the water-cooling system, and the temperature data in the temperature data sequence is the water-coolant temperature data;

[0007] According to the time series decomposition result of the temperature data sequence, obtain the suspected fluctuating temperature data in the temperature data sequence, and according to the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data, obtain the target fluctuation characterization value of the suspected fluctuating temperature data. According to the target fluctuation characterization value, obtain the data sequence to be analyzed of the temperature data sequence and the data to be analyzed for fluctuation in the data sequence to be analyzed;

[0008] According to the order of each data to be analyzed in the data sequence to be analyzed and the data to be analyzed for fluctuation in the data sequence to be analyzed, obtain the coefficient adjustment factor of the data to be analyzed for fluctuation;

[0009] Obtain the first coefficient of each data to be analyzed in the data sequence to be analyzed, and according to the first coefficient of each data to be analyzed and the coefficient adjustment factor of the data to be analyzed for fluctuation, obtain the second coefficient of each data to be analyzed in the data sequence to be analyzed; According to the second coefficient of the data to be analyzed and the prediction model, obtain the predicted coolant temperature data of the water-cooling device at the future monitoring moment;

[0010] Adjust the working power of the water-cooling device in the water-cooling system according to the predicted coolant temperature data.

[0011] Beneficial effects: The present invention first obtains the temperature data sequence corresponding to each water-cooling device in the water-cooling system; then, according to the time-series decomposition result of the temperature data sequence, it obtains the suspected fluctuating temperature data in the temperature data sequence, and based on the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data, it obtains the target fluctuation characterization value of the suspected fluctuating temperature data. According to the target fluctuation characterization value, it obtains the data sequence to be analyzed in the temperature data sequence and the fluctuating data to be analyzed in the data sequence to be analyzed; then, according to the order of each data to be analyzed in the data sequence to be analyzed and the fluctuating data to be analyzed in the data sequence to be analyzed, it obtains the coefficient adjustment factor of the fluctuating data to be analyzed; immediately afterwards, it obtains the first coefficient of each data to be analyzed in the data sequence to be analyzed, and based on the first coefficient of each data to be analyzed and the coefficient adjustment factor of the fluctuating data to be analyzed, it obtains the second coefficient of each data to be analyzed in the data sequence to be analyzed. According to the second coefficient of the data to be analyzed and the prediction model, it obtains the predicted coolant temperature data of the water-cooling device at the future monitoring moment; finally, according to the predicted coolant temperature data, it adjusts the working power of the water-cooling device in the water-cooling system. Moreover, the present invention adjusts the first coefficient according to the coefficient adjustment factor, and then based on the adjusted second coefficient, it can improve the reliability of the prediction result, and thus can ensure the reliability or effectiveness when adjusting the working power of the water-cooling device subsequently, that is, it can reliably and effectively perform intelligent operation and maintenance management on the water-cooling system. Brief Description of the Drawings

[0012] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0013] Figure 1 It is a flowchart of an intelligent operation and maintenance management method for a water-cooling system based on a cloud platform according to the present invention. Detailed Embodiments

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art belong to the scope protected by the embodiments of the present invention.

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art belonging to the present invention.

[0016] This embodiment provides an intelligent operation and maintenance management system for a water-cooled system based on a cloud platform, including a processor and a memory. The processor executes the computer program stored in the memory to implement a monitoring method for the intestinal soothing pet food process treatment based on artificial intelligence, as Figure 1 shown. The intelligent operation and maintenance management method for the water-cooled system based on the cloud platform includes the following steps:

[0017] Step S001, in the current monitoring time period, obtain the temperature data sequence of each water-cooled device in the water-cooled system.

[0018] This embodiment mainly aims at the water-cooled system applied to large-scale demand scenarios, such as factory workshops, data centers, etc. In order to ensure the working effect of the water-cooled system applied to large-scale demand scenarios, it is generally required to design multiple sets of water-cooled devices in the water-cooled system applied to large-scale demand scenarios, that is, the water-cooled system applied to large-scale demand scenarios includes multiple sets of water-cooled devices. One set of water-cooled devices consists of a cooling tower, a cooling water circulation pump, and a centrifugal chiller; and for the convenience of analysis in this embodiment, subsequent analysis will be carried out taking the water-cooled system applied to any large-scale demand scenario as an example. For example, subsequent analysis can be carried out taking the water-cooled system applied to the data center as an example, that is, all the water-cooled systems mentioned in this embodiment are the same water-cooled system, and the water-cooled devices mentioned are the water-cooled devices in this water-cooled system; in addition, the number of water-cooled devices in the water-cooled system in this embodiment needs to be determined by the implementer according to the applied scenario. If the applied scenario or place is large, water-cooled devices with a larger number of sets can be selected.

[0019] Since data prediction needs to be carried out in this embodiment later, this embodiment first needs to obtain the temperature data sequence corresponding to each set of water-cooled devices in the water-cooled system in the current monitoring time period. The temperature data in the temperature data sequence is the temperature data of the water-cooling liquid. The temperature data sequence is mainly used for data prediction, and the specific acquisition method of the temperature data sequence corresponding to each set of water-cooled devices is as follows:

[0020] At each monitoring moment in the current monitoring time period, obtain the temperature of the water-cooling liquid in each set of water-cooled devices, and construct the time series sequence composed of all the temperatures of the water-cooling liquid in the same set of water-cooled devices obtained in the current monitoring time period, which is recorded as the temperature data sequence corresponding to the corresponding water-cooled device. That is, the temperature data sequence corresponding to any water-cooled device is composed of all the temperatures of the water-cooling liquid in this water-cooled device obtained in the current monitoring time period, and the temperature of the water-cooling liquid in the water-cooled device refers to the temperature of the water-cooling liquid in the centrifugal chiller in the corresponding water-cooled device.

[0021] In addition, in specific applications, the implementer needs to set the time interval between the current monitoring time period and adjacent monitoring moments according to the actual situation. For example, in this embodiment, the current monitoring time period can be set to one hour before the current monitoring moment, and it is required that the current monitoring time period includes the current monitoring moment. The time interval between adjacent monitoring moments can be set to 1 second or 0.1 second, etc. It should also be noted that in this embodiment, each set of water cooling equipment in the water cooling system can be adjusted independently. Moreover, in this embodiment, a temperature sensor is used to collect the temperature of the water cooling liquid in the centrifugal chiller, and the collected data will be remotely transmitted to the water cooling system operation and maintenance management cloud platform. After receiving the data, the cloud platform will store it in the data sequence corresponding to each water cooling equipment.

[0022] Therefore, through the above process, this embodiment obtains the temperature data sequence corresponding to each set of water cooling equipment in the water cooling system.

[0023] Step S002: According to the result of the time series decomposition of the temperature data sequence, obtain the suspected fluctuating temperature data in the temperature data sequence. According to the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data, obtain the target fluctuation characterization value of the suspected fluctuating temperature data. According to the target fluctuation characterization value, obtain the data sequence to be analyzed of the temperature data sequence and the fluctuating data to be analyzed in the data sequence to be analyzed.

[0024] When obtaining the temperature data sequence corresponding to the water-cooling device, local flow unevenness in the coolant pipeline will cause some data with large fluctuations to appear in the obtained temperature data sequence. When there are data with large fluctuations caused by local flow unevenness in the coolant pipeline in the obtained temperature data sequence, it will lead to a problem of low reliability of the prediction result when using the temperature data sequence to predict the water-cooling liquid temperature at future monitoring moments. When the reliability of the prediction result is low, there will also be problems of unreliable adjustment or poor adjustment effect when adjusting the working power of the water-cooling device based on the prediction result later. In order to ensure the reliability or effectiveness of adjusting the working power of the water-cooling device in this embodiment, it is necessary to improve the reliability of the prediction. That is, the main purpose of this embodiment is to improve the reliability of predicting the water-cooling liquid temperature at future monitoring moments. And it is known that the data with large fluctuations caused by local flow unevenness in the coolant pipeline is the main reason for the unreliable prediction result. Therefore, in this embodiment, the influence degree of this type of data on the prediction result will be reduced later. That is, in this embodiment, the data with large fluctuations that appear in the temperature data sequence caused by local flow unevenness in the coolant pipeline will be determined first, that is, the data to be analyzed for fluctuations. Then, by analyzing the possibility that the data to be analyzed for fluctuations is caused by local flow unevenness in the coolant pipeline or a sudden temperature rise of the heat source itself, the coefficient of the data to be analyzed for fluctuations will be adjusted accordingly, so as to improve the reliability of the prediction.

[0025] Therefore, based on the above analysis, it can be seen that this embodiment needs to obtain the data to be analyzed for fluctuations first. Before obtaining the data to be analyzed for fluctuations, it is necessary to perform time series decomposition on the temperature data sequence corresponding to each set of water-cooling devices obtained above. Then, according to the time series decomposition result of the temperature data sequence corresponding to each set of water-cooling devices, the suspected fluctuating temperature data in the temperature data sequence corresponding to each set of water-cooling devices is obtained. Subsequently, based on the known suspected fluctuating temperature data, further analysis is carried out to determine the data to be analyzed for fluctuations, that is, the noise temperature data in the suspected fluctuating temperature data is screened to further determine the data to be analyzed for fluctuations. Therefore, in this embodiment, the specific process of obtaining the suspected fluctuating temperature data in the temperature data sequence corresponding to each set of water-cooling devices is as follows:

[0026] First, perform STL time series decomposition on the temperature data series corresponding to each set of water cooling equipment respectively, and obtain the residuals of each temperature data in the temperature data series corresponding to each set of water cooling equipment, that is, the residual terms; and the process of performing time series decomposition on the time series to obtain the trend term, the periodic term, and the residual term is a well-known technology, so it will not be described in detail; then construct a first mapping coordinate system, and the horizontal axis of the first mapping coordinate system represents time, and the vertical axis represents the residuals of the temperature data; then map the acquisition time of each temperature data and the residuals of the temperature data in the temperature data series corresponding to each set of water cooling equipment into the constructed first mapping coordinate system, and obtain the residual points corresponding to each temperature data in the temperature data series corresponding to each set of water cooling equipment, and the vertical coordinate value of the residual point corresponding to the temperature data is the residual of the corresponding temperature data, and the horizontal coordinate value is the acquisition time of the corresponding temperature data.

[0027] Immediately afterwards, in the first mapping coordinate system, obtain the straight line with the vertical coordinate value of 0, and denote it as the first straight line, that is, the residuals of all points on the first straight line are 0; then, according to the residual points corresponding to each temperature data in the temperature data series corresponding to each set of water cooling equipment and the first straight line, obtain the initial fluctuation index values of each temperature data in the temperature data series corresponding to each set of water cooling equipment; then, according to the initial fluctuation index values of each temperature data in the temperature data series, obtain the suspected fluctuation temperature data in the temperature data series corresponding to each set of water cooling equipment.

[0028] In this embodiment, for the convenience of understanding, this embodiment takes the specific acquisition process of the initial fluctuation index value of any temperature data a in the temperature data series A corresponding to any set of water cooling equipment as an example for description, that is, the specific acquisition process of the initial fluctuation index value of temperature data a is as follows:

[0029] First, obtain the set composed of the residual points corresponding to all temperature data in the temperature data series A except temperature data a, and denote it as the first set; then, in the first set, obtain the residual point closest to the residual point corresponding to temperature data a, and denote it as the nearest neighbor point corresponding to temperature data a; then, obtain the Euclidean distance between the residual point corresponding to temperature data a and the nearest neighbor point corresponding to temperature data a, and denote it as the first distance, and obtain the distance between the residual point corresponding to temperature data a and the first straight line, and denote it as the second distance; finally, obtain the normalized value of the result obtained by adding the first distance and the second distance, and denote it as the initial fluctuation index value of temperature data a.

[0030] In addition, the specific expression for obtaining the initial fluctuation index value of temperature data a is:

[0031] B1 a =norm(D1+D2)

[0032] where B1a is the initial fluctuation index value of the temperature data a, D1 is the first distance, D2 is the second distance, and norm() is the normalization function; moreover, when D1 is larger or D2 is larger, it indicates that the temperature data a is less in line with trendiness and periodicity, and also indicates that the degree of outlier of the temperature data a is more obvious. When the degree of outlier is more obvious, it indicates that the temperature data a is more likely to be caused by local uneven flow in the coolant pipeline, external noise, and sudden temperature rise of the heat source itself. When D1 is larger and D2 is larger, B1 a has a larger value. Therefore, when B1 a has a larger value, it indicates that the temperature data a is more likely to be caused by local uneven flow in the coolant pipeline, external noise, and sudden temperature rise of the heat source itself. That is, when the temperature data a is noise temperature data, B1 a also has a relatively large value. So when B1 a has a larger value, it only indicates that the temperature data a is more likely to be suspected of being fluctuating temperature data. On the contrary, when B1 a has a larger value, it indicates that the temperature data a is less likely to be suspected of being fluctuating temperature data. That is, the initial fluctuation index value cannot specifically determine the cause of the large data fluctuation.

[0033] In this embodiment, the specific process for obtaining the suspected fluctuating temperature data in the temperature data sequence is as follows: for the temperature data a, if the initial fluctuation characterization value of the temperature data a is not less than the preset first fluctuation threshold, then it is determined that the temperature data a is suspected of being fluctuating temperature data, that is, the temperature data a is recorded as suspected of being fluctuating temperature data; and in specific applications, the implementer needs to set the preset first fluctuation threshold according to the actual situation and mathematical experimental statistics. For example, in this embodiment, the preset first fluctuation threshold can be set to 0.7.

[0034] Therefore, through the above process, the present embodiment can obtain the suspected fluctuating temperature data in the temperature data sequence corresponding to each set of water-cooling equipment. Since the suspected fluctuating temperature data may not only be caused by local uneven flow in the coolant pipeline, but also may be noise data or data caused by a sudden temperature rise in the heat source itself. Therefore, after obtaining the suspected fluctuating temperature data, the present embodiment also needs to obtain the target fluctuation characterization value of each suspected fluctuating temperature data in the temperature data sequence according to the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data. And the target fluctuation characterization value is an important parameter for screening out noise temperature data to determine the data to be analyzed for fluctuations, and is determined in the present embodiment; in addition, for the convenience of understanding, the present embodiment will hereinafter describe the process of obtaining the target fluctuation characterization value of any suspected fluctuating temperature data b in the temperature data sequence B corresponding to any water-cooling equipment as an example, that is, the process of obtaining the target fluctuation characterization value of the suspected fluctuating temperature data b is as follows:

[0035] First, in the temperature data sequence B, obtain the sequence composed of a continuous preset number of temperature data adjacent to the suspected fluctuating temperature data b and located on the left side of the suspected fluctuating temperature data b, and denote it as the left adjacent sequence of the suspected fluctuating temperature data b. Obtain the sequence composed of a continuous preset number of temperature data adjacent to the suspected fluctuating temperature data b and located on the right side of the suspected fluctuating temperature data b, and denote it as the right adjacent sequence of the suspected fluctuating temperature data b; then obtain the difference sequence of the left adjacent sequence of the suspected fluctuating temperature data b, and denote the variance of the difference sequence of the left adjacent sequence as the first variance of the suspected fluctuating temperature data b. Obtain the difference sequence of the right adjacent sequence of the suspected fluctuating temperature data b, and denote the variance of the difference sequence of the right adjacent sequence as the second variance of the suspected fluctuating temperature data b; and in specific applications, the implementer needs to set the value of the preset number according to the actual situation and mathematical experimental statistics. For example, in the present embodiment, the value of the preset number can be set to 10.

[0036] In addition, in this embodiment, there may be a situation where the number of data on the left or right side of the suspected fluctuating temperature data b is less than the preset number. When the number of data on the left side of the suspected fluctuating temperature data b is less than the preset number, the sequence composed of all the data on the left side of the suspected fluctuating temperature data b is recorded as the left adjacent sequence of the suspected fluctuating temperature data b. When the number of data on the right side of the suspected fluctuating temperature data b is less than the preset number, the sequence composed of all the data on the right side of the suspected fluctuating temperature data b is recorded as the right adjacent sequence of the suspected fluctuating temperature data b. It should also be noted that although the first and last data in the temperature data sequence B are almost impossible to be suspected fluctuating temperature data, in order to prevent special situations, it is still necessary to explain the specific values of the first variance and the second variance of the suspected fluctuating temperature data b when the suspected fluctuating temperature data b is the first or last data in the temperature data sequence B. That is, when the suspected fluctuating temperature data b is the first data in the temperature data sequence B, there is no data on the left side of the suspected fluctuating temperature data b. At this time, the first variance of the suspected fluctuating temperature data b is recorded as 0. When the suspected fluctuating temperature data b is the last data in the temperature data sequence B, there is no data on the right side of the suspected fluctuating temperature data b. At this time, the second variance of the suspected fluctuating temperature data b is recorded as 0.

[0037] After that, a second mapping coordinate system is constructed. The horizontal axis of the second mapping coordinate system represents time, and the vertical axis represents temperature data. Each temperature data in the temperature data sequence B and the acquisition time of each temperature data are mapped into the constructed second mapping coordinate system. The data points obtained by mapping are recorded as the temperature data points corresponding to each temperature data in the temperature data sequence B. The abscissa value of the temperature data point corresponding to the temperature data is the acquisition time of the corresponding temperature data, and the ordinate value is the corresponding temperature data.

[0038] After obtaining the temperature data points corresponding to the temperature data, the near-neighbor suspected data points and the feature set of the suspected fluctuating temperature data b are obtained according to the temperature data points corresponding to each temperature data in the temperature data sequence B. Then, according to the first variance, the second variance, the near-neighbor suspected data points and the feature set of the suspected fluctuating temperature data b, the target fluctuation characterization value of the suspected fluctuating temperature data b is obtained.

[0039] In this embodiment, the specific process of obtaining the near-neighbor suspected data points and the feature set of the suspected fluctuating temperature data b according to the temperature data points corresponding to each temperature data in the temperature data sequence B is as follows:

[0040] First, in the temperature data sequence B, mark all the temperature data points corresponding to the suspected fluctuating temperature data except the suspected fluctuating temperature data b as suspected data points, and mark the temperature data point corresponding to the suspected fluctuating temperature data b as the first data point. Then, in the second mapping coordinate system, obtain the suspected data point that is located on the left side of the first data point and is closest to the first data point, and mark it as the left neighboring suspected data point of the suspected fluctuating temperature data b. Obtain the suspected data point that is located on the right side of the first data point and is closest to the first data point, and mark it as the right neighboring suspected data point of the suspected fluctuating temperature data b. That is, the neighboring suspected data points include the left neighboring suspected data point and the right neighboring suspected data point.

[0041] After that, obtain the interval formed by the abscissa value of the left neighboring suspected data point of the suspected fluctuating temperature data b and the abscissa value of the first data point, and mark it as the left interval. Obtain the interval formed by the abscissa value of the right neighboring suspected data point of the suspected fluctuating temperature data b and the abscissa value of the first data point, and mark it as the right interval. And the left interval contains the abscissa value of the left neighboring suspected data point and does not contain the abscissa value of the first data point. The right interval contains the abscissa value of the right neighboring suspected data point and does not contain the abscissa value of the first data point. Then, in the second mapping coordinate system, obtain all the temperature data points whose abscissa values belong to the left interval, and mark the set formed by all the temperature data points whose abscissa values belong to the left interval as the first feature set. Obtain all the temperature data points whose abscissa values belong to the right interval, and mark the set formed by all the temperature data points whose abscissa values belong to the right interval as the second feature set. That is, the feature set includes the first feature set and the second feature set.

[0042] In this embodiment, the specific process of obtaining the target fluctuation characterization value of the suspected fluctuating temperature data b according to the first variance, the second variance, the neighboring suspected data points, and the feature set of the suspected fluctuating temperature data b is as follows:

[0043] First, obtain the sum of the first variance of the suspected fluctuating temperature data b and the second variance of the suspected fluctuating temperature data b, and perform a negative correlation mapping on the sum of the first variance of the suspected fluctuating temperature data b and the second variance of the suspected fluctuating temperature data b, and mark the mapping result as the first index value of the suspected fluctuating temperature data b.

[0044] Then, obtain the Euclidean distance between the left neighboring suspected data point of the suspected fluctuating temperature data b and the first data point, and denote it as the left neighboring distance of the suspected fluctuating temperature data b. Obtain the Euclidean distance between the right neighboring suspected data point of the suspected fluctuating temperature data b and the first data point, and denote it as the right neighboring distance of the suspected fluctuating temperature data b. Then, obtain the sum of the left neighboring distance and the right neighboring distance of the suspected fluctuating temperature data b, and perform a negative correlation mapping on the sum of the left neighboring distance and the right neighboring distance of the suspected fluctuating temperature data b, and denote the mapping result as the second index value. It should be noted that there may be a situation where the suspected fluctuating temperature data b does not have a left neighboring suspected data point or a right neighboring suspected data point. When the suspected fluctuating temperature data b does not have a left neighboring suspected data point, directly denote the value obtained by performing a negative correlation mapping on the right neighboring distance of the suspected fluctuating temperature data b as the second index value. When the suspected fluctuating temperature data b does not have a right neighboring suspected data point, directly denote the value obtained by performing a negative correlation mapping on the left neighboring distance of the suspected fluctuating temperature data b as the second index value of the suspected fluctuating temperature data b.

[0045] Immediately afterwards, obtain the mean of the ordinate values of all data points in the first feature set of the suspected fluctuating temperature data b, and denote it as the first mean. Calculate the absolute value of the difference between the first mean and the ordinate value of the first data point, and denote it as the first difference value of the suspected fluctuating temperature data b. Obtain the mean of the ordinate values of all data points in the second feature set of the suspected fluctuating temperature data b, and denote it as the second mean. Obtain the absolute value of the difference between the second mean and the ordinate value of the first data point, and denote it as the second difference value of the suspected fluctuating temperature data b. Then, obtain the sum of the first difference value and the second difference value of the suspected fluctuating temperature data b, and then perform a negative correlation mapping on the sum of the first difference value and the second difference value of the suspected fluctuating temperature data b, and denote the mapping result as the third index value of the suspected fluctuating temperature data b. It should be noted that there may be a situation where the suspected fluctuating temperature data b does not have a first feature set or a second feature set. When the suspected fluctuating temperature data b does not have a first feature set, directly denote the result of performing a negative correlation mapping on the second difference value of the suspected fluctuating temperature data b as the third index value of the suspected fluctuating temperature data b. When the suspected fluctuating temperature data b does not have a second feature set, directly denote the result of performing a negative correlation mapping on the first difference value of the suspected fluctuating temperature data b as the third index value of the suspected fluctuating temperature data b.

[0046] Finally, obtain the sum of the first index value, the second index value, and the third index value of the suspected fluctuating temperature data b, perform normalization processing on the sum of the first index value, the second index value, and the third index value of the suspected fluctuating temperature data b, and then record the result obtained from the normalization processing as the target fluctuation characterization value of the suspected fluctuating temperature data b.

[0047] In addition, in this embodiment, the specific expression for obtaining the target fluctuation characterization value of the suspected fluctuating temperature data b is:

[0048]

[0049] where B2 b is the target fluctuation characterization value of the suspected fluctuating temperature data b, exp() is the exponential function with the constant e as the base, W1 is the first index value of the suspected fluctuating temperature data b, W2 is the second index value of the suspected fluctuating temperature data b, W3 is the third index value of the suspected fluctuating temperature data b, C1 is a preset first constant, and C1 is for normalizing the result of exp(-W1)+exp(-W2)+exp(-W3). Therefore, in this embodiment, it is required that the value of C1 is greater than 3, that is, in this embodiment, C1 can be set to 3.

[0050] In addition, when W1, W2, and W3 are larger, it indicates that the suspected fluctuating temperature data b has more prominent or random characteristics within its neighborhood range, and also has less process change characteristics within the neighborhood range where the suspected fluctuating temperature data b is located. The process change characteristic means that the data within the neighborhood range where the suspected fluctuating temperature data b is located generally shows an increasing or decreasing phenomenon. And when the suspected fluctuating temperature data b has more prominent or random characteristics within its neighborhood range and the neighborhood range where the suspected fluctuating temperature data b is located also has less process change characteristics, it indicates that the suspected fluctuating temperature data b is more likely to be noise temperature data. On the contrary, when W1, W2, and W3 are smaller, it indicates that the suspected fluctuating temperature data b is less likely to be noise temperature data; and since when W1, W2, and W3 are larger, B2 b has a smaller value, so when B2 b has a smaller value, it more indicates that the suspected fluctuating temperature data b is more likely to be noise temperature data. On the contrary, when B2 b has a larger value, it more indicates that the suspected fluctuating temperature data b is less likely to be noise temperature data.

[0051] Therefore, according to the above process of obtaining the target fluctuation characterization value of the suspected fluctuating temperature data b, the target fluctuation characterization value of each suspected fluctuating temperature data can be obtained. And based on the above analysis, it can be known that the target fluctuation characterization value is to focus on screening out the noise temperature data in the suspected fluctuating temperature data, and it cannot distinguish whether it is caused by a sudden temperature rise of the heat source itself or local uneven flow in the coolant pipeline. However, the existence of noise temperature data will also affect the prediction result. Therefore, it is necessary to first screen out the noise temperature data in the suspected fluctuating temperature data based on the target fluctuation characterization value, then smooth out the noise temperature data, and determine the data to be analyzed for fluctuations after smoothing. Therefore, in the following, this embodiment will obtain the data sequence to be analyzed of the temperature data sequence corresponding to each water-cooled device and the data to be analyzed for fluctuations in the data sequence to be analyzed according to the target fluctuation characterization value of the suspected fluctuating temperature data. That is, the specific acquisition process of the data sequence to be analyzed of the temperature data sequence and the data to be analyzed for fluctuations in the data sequence to be analyzed is as follows:

[0052] For any suspected fluctuating temperature data b, if the target fluctuation characterization value of the suspected fluctuating temperature data b is not less than the preset second fluctuation threshold, then it is determined that the suspected fluctuating temperature data b is fluctuating temperature data; otherwise, it is determined that the suspected fluctuating temperature data b is noise temperature data. And in specific applications, the implementer needs to set the preset second fluctuation threshold according to the actual situation and mathematical experimental statistics. For example, in this embodiment, the preset second fluctuation threshold can be set to 0.6.

[0053] Therefore, the noise temperature data in the suspected fluctuating temperature data b can be determined through the above method. After determining the noise temperature data, the replacement value of each noise temperature data is obtained. And the replacement value of the noise temperature data is the average value of the adjacent temperature data pairs of the noise temperature data. And the adjacent temperature data pair of any noise temperature data is composed of the temperature data adjacent to and on the left side of the noise temperature data and the temperature data adjacent to and on the right side of the noise temperature data. That is, the noise temperature data and the data in the adjacent temperature data pair of the noise temperature data belong to the same sequence. Then, each noise temperature data in the temperature data sequence is replaced with the replacement value of the corresponding noise temperature data, and the new sequence obtained after the replacement is recorded as the data sequence to be analyzed of the corresponding temperature data sequence. And if the f-th temperature data in any temperature data sequence is fluctuating temperature data, then the f-th data to be analyzed in the data sequence to be analyzed of this temperature data sequence is recorded as the data to be analyzed for fluctuations.

[0054] Therefore, through the above process, this embodiment can obtain the data sequence to be analyzed of each temperature data sequence and the data to be analyzed for fluctuations in the data sequence to be analyzed.

[0055] Step S003: Obtain a coefficient adjustment factor for the data to be analyzed for fluctuations based on the order of each piece of data to be analyzed in the data sequence to be analyzed and the data to be analyzed for fluctuations in the data sequence to be analyzed.

[0056] When there is local uneven flow in the coolant pipeline of the water-cooling equipment, it will cause a phenomenon that the coolant does not circulate temporarily, that is, the water-cooling system will have a temporary failure. However, this situation usually relieves itself within the coolant. Therefore, when this phenomenon occurs, it is not necessary to adjust the power of the water-cooling equipment. However, when performing data prediction, the data to be analyzed for fluctuations caused by this phenomenon will affect the prediction result, that is, the data to be analyzed for fluctuations caused by this phenomenon may cause an incorrect determination when determining the method of adjusting the power of the water-cooling equipment based on the prediction result, resulting in relatively low reliability or effectiveness of the adjustment. Therefore, when performing prediction, this embodiment needs to reduce the coefficient of the data to be analyzed for fluctuations caused by this phenomenon, that is, reduce the influence of the data to be analyzed for fluctuations caused by this phenomenon on the prediction result. However, since the obtained data to be analyzed for fluctuations not only includes the data to be analyzed for fluctuations caused by local uneven flow in the coolant pipeline, but also includes the data to be analyzed for fluctuations caused by a sudden temperature rise of the heat source itself. When there is a trend of a sudden temperature rise of the heat source itself, it is necessary to adjust the power of the water-cooling equipment in a timely manner. Therefore, it is necessary to retain the importance of the data to be analyzed for fluctuations caused by a sudden temperature rise of the heat source itself as much as possible, that is, not to adjust the coefficient of the data to be analyzed for fluctuations caused by a sudden temperature rise of the heat source itself as much as possible, and the heat source in this embodiment is the main factor affecting the temperature in the scenario where the water-cooling system is applied.

[0057] Moreover, since when a sudden rise occurs in the heat source starting from a certain moment, the water-cooling liquid temperatures in multiple sets of water-cooling equipment in the water-cooling system will all have large fluctuations. However, when there is local uneven flow in the coolant pipeline of a certain water-cooling equipment, it will probably only cause large fluctuations in the water-cooling liquid temperature in that water-cooling equipment. Therefore, this embodiment obtains the coefficient adjustment factor for the data to be analyzed for fluctuations based on the above-described characteristics, that is, this embodiment will next obtain the coefficient adjustment factor for each piece of data to be analyzed for fluctuations in the data sequence to be analyzed based on the order of each piece of data to be analyzed in the data sequence to be analyzed and the data to be analyzed for fluctuations in the data sequence to be analyzed. That is, the specific method for obtaining the coefficient adjustment factor for the data to be analyzed for fluctuations is as follows:

[0058] First, according to the order of each data to be analyzed in the data sequence to be analyzed of each temperature data sequence, the characteristic marker values of each data to be analyzed in the data sequence to be analyzed of each temperature data sequence are obtained, and the characteristic marker value of the g-th data to be analyzed in each data sequence to be analyzed is g; then, in all data sequences to be analyzed, all data to be analyzed with the same characteristic marker value are divided into the same set, and each set obtained by the division is denoted as a characteristic subset, and the characteristic marker value of the data to be analyzed in each characteristic subset is used as the identification value of the corresponding characteristic subset, that is, the characteristic marker values of the data to be analyzed in the same characteristic subset are consistent; then, the K-means clustering algorithm is used to cluster the data to be analyzed in each characteristic subset respectively to obtain each clustering cluster corresponding to each characteristic subset. And in specific applications, the K value during clustering, that is, the number of clustering center points, is determined by the elbow method; since the process of clustering using the k-means clustering algorithm and the process of using the elbow method to determine the number of clustering center points are well-known technologies, they will not be described in detail in this embodiment.

[0059] After clustering each characteristic subset is completed, the coefficient adjustment factor of each fluctuating data to be analyzed is obtained based on the obtained clustering result; and for the convenience of understanding in this embodiment, subsequently, the coefficient adjustment factor of any fluctuating data to be analyzed h in any data sequence to be analyzed, that is, the specific obtaining method of the coefficient adjustment factor of the fluctuating data to be analyzed h is as follows:

[0060] First, obtain the feature subset whose recognition value is the same as the feature marker value of the data h to be analyzed for fluctuations, and denote it as the target subset corresponding to the data h to be analyzed for fluctuations. Then, in each clustering cluster corresponding to the target subset, obtain the clustering cluster that contains the data h to be analyzed for fluctuations, and denote it as the clustering cluster to be analyzed. After that, obtain the ratio of the total number of data in the clustering cluster to be analyzed to the total number of data in the target subset corresponding to the data h to be analyzed for fluctuations, and denote it as the first ratio. Immediately afterwards, obtain the reciprocal of the number of clustering clusters corresponding to the target first subset, and denote it as the second ratio. Then, in the clustering cluster to be analyzed, obtain the nearest neighbor data to be analyzed for each data to be analyzed in the clustering cluster to be analyzed, and denote it as the nearest neighbor data within the cluster corresponding to the data to be analyzed. That is, at this time, the nearest neighbor data within the cluster corresponding to the data to be analyzed is the closest to the data to be analyzed in the clustering cluster to be analyzed. After that, obtain the absolute value of the difference between each data to be analyzed in the clustering cluster to be analyzed and its corresponding nearest neighbor data within the cluster, and denote it as the feature difference of the corresponding data to be analyzed. Then, obtain the mean value of the feature differences of all the data to be analyzed in the clustering cluster to be analyzed, and denote it as the first mean value. And denote the negative correlation mapping value of the first mean value as the first mapping value. The first mapping value is exp(-R), where R is the first mean value. Finally, obtain the normalized value of the result obtained by adding the first ratio, the second ratio, and the first mapping value, and denote it as the coefficient adjustment factor of the data h to be analyzed for fluctuations. And here, the normalization function norm() can be used to obtain the normalized value.

[0061] Moreover, when the first ratio, the second ratio, and the first mapping value are larger, it indicates that the possibility that the data h to be analyzed for fluctuations is caused by a sudden rise in the heat source is greater. Also, since the product of the first coefficient and the coefficient adjustment factor is used as the second coefficient later, that is, the larger the coefficient adjustment factor, the less adjustment is made, and the second coefficient is the coefficient used in subsequent predictions. Therefore, when the first ratio, the second ratio, and the first mapping value are larger, it more indicates that the coefficient adjustment factor of the data h to be analyzed for fluctuations is larger. On the contrary, when the first ratio, the second ratio, and the first mapping value are smaller, it more indicates that the coefficient adjustment factor of the data h to be analyzed for fluctuations is smaller.

[0062] Therefore, through the above process, this embodiment can obtain the coefficient adjustment factor of each data to be analyzed for fluctuations.

[0063] Step S004: Obtain the first coefficient of each data to be analyzed in the data sequence to be analyzed, and obtain the second coefficient of each data to be analyzed in the data sequence to be analyzed according to the first coefficient of each data to be analyzed and the coefficient adjustment factor of the data to be analyzed for fluctuations. According to the second coefficient of the data to be analyzed and the prediction model, obtain the predicted coolant temperature data of the water-cooled device at the future monitoring moment.

[0064] In this embodiment, the first coefficients of each piece of data to be analyzed in the data sequence to be analyzed are first obtained. Since an ARIMA prediction model is used for prediction in the subsequent steps of this embodiment, the first coefficients of the data to be analyzed include the first autoregressive coefficient and the first moving average coefficient. The method for obtaining the first coefficients is as follows: First, an ARIMA prediction model is constructed, and the data sequence to be analyzed is imported into the constructed ARIMA prediction model. The coefficients of each piece of data to be analyzed in the data sequence to be analyzed are obtained, and the coefficients of the load data to be analyzed obtained at this time are recorded as the first coefficients of the corresponding load data to be analyzed. The initial coefficients of the load data to be analyzed are determined by the traditional coefficient acquisition method in the ARIMA prediction model. For example, the maximum likelihood (MLE) method can be used to determine the initial coefficients of the load data to be analyzed. Also, since the coefficients of the load data to be analyzed include the autoregressive coefficient in the AR model and the moving average coefficient in the MA model, the initial coefficients of each piece of load data to be analyzed include the first autoregressive coefficient and the first moving average coefficient. In addition, it should be noted that the method for obtaining the first coefficients of the data to be analyzed is a well-known technology, so it will not be described in detail in this embodiment.

[0065] Then, according to the first coefficients of each piece of data to be analyzed and the coefficient adjustment factor of the fluctuating data to be analyzed, the second coefficients of each piece of data to be analyzed in the data sequence to be analyzed are obtained. That is, the specific method for obtaining the second coefficients of the data to be analyzed is as follows: For any piece of data to be analyzed in any data sequence to be analyzed, if it is determined that the data to be analyzed does not belong to the fluctuating data to be analyzed, the first coefficient of the data to be analyzed is used as the second coefficient of the data to be analyzed. If it is determined that the data to be analyzed belongs to the fluctuating data to be analyzed, the product of the first coefficient of the data to be analyzed and the coefficient adjustment factor of the data to be analyzed is used as the second coefficient of the data to be analyzed.

[0066] Next, according to the data sequence to be analyzed of the temperature data sequence corresponding to each set of water-cooling equipment, the second coefficients of each piece of data to be analyzed in the data sequence to be analyzed of the temperature data sequence corresponding to each set of water-cooling equipment, and the ARIMA prediction model, the predicted coolant temperature data of each set of water-cooling equipment at the future monitoring time are obtained. The predicted coolant temperature data of each set of water-cooling equipment at the future monitoring time refers to the predicted coolant temperature data of each set of water-cooling equipment at the next monitoring time. When using the ARIMA prediction model to obtain the predicted coolant temperature data at the future monitoring time, only the first coefficients of the data to be analyzed are replaced with the second coefficients of the data to be analyzed, and other steps or processes remain unchanged.

[0067] Step S005: Adjust the working power of the water-cooling equipment in the water-cooling system according to the predicted coolant temperature data.

[0068] In the following, the working power of each water cooling device will be adjusted according to the predicted coolant temperature data of each water cooling device at the future monitoring moment. Specifically:

[0069] Obtain the coolant temperature in each water cooling device collected at the current monitoring moment, and record it as the actual coolant temperature data of the water cooling device at the current monitoring moment. For any water cooling device, if it is determined that the actual coolant temperature data of the water cooling device at the current monitoring moment is greater than the predicted coolant temperature data of the water cooling device at the future monitoring moment, it indicates that the coolant temperature in the water cooling device has a downward trend. Then, at this time, the working power of the water cooling device should be reduced to save the energy consumption of the water cooling system. If it is determined that the actual coolant temperature data of the water cooling device at the current monitoring moment is less than the predicted coolant temperature data of the water cooling device at the future monitoring moment, it indicates that the coolant temperature in the water cooling device has an upward trend. Then, at this time, the working power of the water cooling device should be increased so that the coolant temperature in the water cooling device can be cooled faster to ensure the working efficiency of the water cooling system. If it is determined that the actual coolant temperature data of the water cooling device at the current monitoring moment is equal to the predicted coolant temperature data of the water cooling device at the future monitoring moment, the working power of the water cooling device will not be adjusted. And in specific applications, the implementer needs to determine the specific adjustment power value according to the actual situation, but it needs to meet that the adjustment degree is proportional to the difference between the actual and the prediction. The difference between the actual and the prediction refers to the absolute value of the difference between the actual coolant temperature data of a certain water cooling device at the current monitoring moment and the predicted coolant temperature data of the water cooling device at the future monitoring moment. For example, when the difference between the actual and the prediction is large, and the actual coolant temperature data of the water cooling device at the current monitoring moment is less than the predicted coolant temperature data of the water cooling device at the future monitoring moment, then at this time, the working power of the water cooling device should be increased, and the adjustment degree should also be large.

[0070] And based on the above-described adjustment principle, in the following, the adjustment process of the working power of any water cooling device V will be taken as an example for specific description, that is, the specific adjustment process of the working power of the water cooling device V is as follows:

[0071] First, it is determined whether the actual coolant temperature data of the water-cooling device V at the current monitoring moment is greater than the predicted coolant temperature data of the water-cooling device V at the future monitoring moment. If so, the working power of the water-cooling device V is adjusted to the first adjustment power S1, that is, starting from the current monitoring moment, the working power of the water-cooling device V is adjusted from S0 to S1. Otherwise, it is continued to determine whether the actual coolant temperature data of the water-cooling device V at the current monitoring moment is less than the predicted coolant temperature data of the water-cooling device V at the future monitoring moment. If so, the working power of the water-cooling device V is adjusted to the second adjustment power S2, that is, starting from the current monitoring moment, the working power of the water-cooling device V is adjusted from S0 to S2; and S1 = S0 × θ × (c2 - norm(|T1 - T0|)), S2 = S0 × θ × (c2 + norm(|T1 - T0|)), where S1 is the first adjustment power, S2 is the second adjustment power, S0 is the working power of the water-cooling device V at the current monitoring moment, θ is the preset adjustment control coefficient, c2 is the preset second constant, norm() is the normalization function, T1 is the predicted coolant temperature data of the water-cooling device V at the future monitoring moment, and T0 is the actual coolant temperature data of the water-cooling device V at the current monitoring moment; and if the actual coolant temperature data of the water-cooling device V at the current monitoring moment is equal to the predicted coolant temperature data of the water-cooling device V at the future monitoring moment, there is no need to adjust the working power of the water-cooling device V, that is, the working power of the water-cooling device V is maintained at S0; in addition, in specific applications, the implementer needs to set the values of the preset adjustment control coefficient and the preset second constant according to the actual situation and mathematical experimental statistics. For example, in this embodiment, the preset adjustment control coefficient can be set to 0.2, the value of the preset second constant can be set to 1, and the preset adjustment control coefficient is used to control the adjustment ratio.

[0072] Therefore, according to the specific adjustment method of the working power of the water-cooling device V in this embodiment, the adjustment of the working power of other water-cooling devices can be completed, that is, the adjustment method of the working power of other water-cooling devices is the same as the adjustment method of the working power of the water-cooling device V described above.

[0073] In summary, in this embodiment, the temperature data sequence corresponding to each water-cooling device in the water-cooling system is first obtained; then, according to the time-series decomposition result of the temperature data sequence, the suspected fluctuating temperature data in the temperature data sequence is obtained, and based on the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data, the target fluctuation characterization value of the suspected fluctuating temperature data is obtained. According to the target fluctuation characterization value, the data sequence to be analyzed in the temperature data sequence and the fluctuating data to be analyzed in the data sequence to be analyzed are obtained; then, according to the order of each data to be analyzed in the data sequence to be analyzed and the fluctuating data to be analyzed in the data sequence to be analyzed, the coefficient adjustment factor of the fluctuating data to be analyzed is obtained; immediately afterwards, the first coefficient of each data to be analyzed in the data sequence to be analyzed is obtained, and according to the first coefficient of each data to be analyzed and the coefficient adjustment factor of the fluctuating data to be analyzed, the second coefficient of each data to be analyzed in the data sequence to be analyzed is obtained. According to the second coefficient of the data to be analyzed and the prediction model, the predicted coolant temperature data of the water-cooling device at the future monitoring moment is obtained; finally, according to the predicted coolant temperature data, the working power of the water-cooling device in the water-cooling system is adjusted. Moreover, in this embodiment, the first coefficient is adjusted according to the coefficient adjustment factor, and then the reliability of the prediction result can be improved based on the adjusted second coefficient, thereby ensuring the reliability or effectiveness when adjusting the working power of the water-cooling device subsequently, that is, the intelligent operation and maintenance management of the water-cooling system can be carried out reliably and effectively.

[0074] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An intelligent operation and maintenance management system for a water-cooling system based on a cloud platform, comprising a processor and a memory, characterized in that, The processor executes the computer program stored in the memory to implement the following steps: In the current monitoring time period, obtain the temperature data sequence corresponding to each set of water-cooling equipment in the water-cooling system, and the temperature data in the temperature data sequence is the water-coolant temperature data; According to the time series decomposition result of the temperature data sequence, obtain the suspected fluctuating temperature data in the temperature data sequence, and according to the suspected fluctuating temperature data adjacent to the suspected fluctuating temperature data and the temperature data within the neighborhood range of the suspected fluctuating temperature data, obtain the target fluctuation characterization value of the suspected fluctuating temperature data, and according to the target fluctuation characterization value, obtain the data sequence to be analyzed in the temperature data sequence and the fluctuating data to be analyzed in the data sequence to be analyzed; According to the order of each data to be analyzed in the data sequence to be analyzed and the fluctuating data to be analyzed in the data sequence to be analyzed, obtain the coefficient adjustment factor of the fluctuating data to be analyzed; Obtain the first coefficient of each data to be analyzed in the data sequence to be analyzed, and according to the first coefficient of each data to be analyzed and the coefficient adjustment factor of the fluctuating data to be analyzed, obtain the second coefficient of each data to be analyzed in the data sequence to be analyzed; according to the second coefficient of the data to be analyzed and the prediction model, obtain the predicted coolant temperature data of the water-cooling equipment at the future monitoring moment; Adjust the working power of the water-cooling equipment in the water-cooling system according to the predicted coolant temperature data.

2. The intelligent operation and maintenance management system of the water-cooled system based on the cloud platform according to claim 1, characterized in that, The method for obtaining the suspected fluctuating temperature data in the temperature data sequence includes: Perform STL time series decomposition on the temperature data sequence to obtain the residuals of each temperature data in the temperature data sequence; Construct a first mapping coordinate system, where the abscissa of the first mapping coordinate system represents time and the ordinate represents the residual of the temperature data; map the acquisition time of each temperature data in the temperature data sequence and the residual of the temperature data to the first mapping coordinate system to obtain the residual points corresponding to each temperature data in the temperature data sequence; in the first mapping coordinate system, obtain the straight line with the ordinate value of 0 and denote it as the first straight line; According to the residual points corresponding to the temperature data and the first straight line, obtain the initial fluctuation index value of each temperature data in the temperature data sequence; For any temperature data a in any temperature data sequence A, if the initial fluctuation characterization value of the temperature data a is not less than the preset first fluctuation threshold, then record the temperature data a as the suspected fluctuating temperature data.

3. The intelligent operation and maintenance management system for the water-cooled system based on the cloud platform according to claim 2, characterized in that, The method for obtaining the initial fluctuation index value of the temperature data includes: For the temperature data a in the temperature data sequence A: Denote the set composed of the residual points corresponding to all temperature data in the temperature data sequence A except the temperature data a as the first set. In the first set, denote the residual point closest to the residual point corresponding to the temperature data a as the nearest neighbor point corresponding to the temperature data a, and denote the Euclidean distance between the residual point corresponding to the temperature data a and the nearest neighbor point corresponding to the temperature data a as the first distance; Denote the distance between the residual point corresponding to the temperature data a and the first straight line as the second distance; Denote the normalized value of the result obtained by adding the first distance and the second distance as the initial fluctuation index value of the temperature data a.

4. The intelligent operation and maintenance management system of the water-cooling system based on the cloud platform according to claim 1, characterized in that, The method for obtaining the target fluctuation characterization value of the suspected fluctuating temperature data includes: For any suspected fluctuating temperature data b in any temperature data sequence B: In the temperature data sequence B, denote the sequence composed of a continuous preset number of temperature data located on the left side of the suspected fluctuating temperature data b as the left adjacent sequence of the suspected fluctuating temperature data b, and denote the sequence composed of a continuous preset number of temperature data located on the right side of the suspected fluctuating temperature data b as the right adjacent sequence of the suspected fluctuating temperature data b; Denote the variance of the difference sequence of the left adjacent sequence as the first variance, and denote the variance of the difference sequence of the right adjacent sequence as the second variance; Construct a second mapping coordinate system, where the abscissa of the second mapping coordinate system represents time and the ordinate represents temperature data; Map each temperature data in the temperature data sequence B and the acquisition time of each temperature data into the second mapping coordinate system to obtain the temperature data points corresponding to each temperature data in the temperature data sequence B; According to the temperature data points corresponding to each temperature data in the temperature data sequence B, obtain the near-neighbor suspected data points and the feature set of the suspected fluctuating temperature data b; According to the first variance, the second variance, the near-neighbor suspected data points and the feature set, obtain the target fluctuation characterization value of the suspected fluctuating temperature data b.

5. The intelligent operation and maintenance management system of the water-cooling system based on the cloud platform according to claim 4, characterized in that The method for obtaining the near-neighbor suspected data points and the feature set of the suspected fluctuating temperature data b includes: In the temperature data sequence B, denote all temperature data points corresponding to the suspected fluctuating temperature data except the suspected fluctuating temperature data b as suspected data points; Denote the temperature data point corresponding to the suspected fluctuating temperature data b as the first data point; In the second mapping coordinate system, denote the suspected data point located on the left side of the first data point and closest to the first data point as the left near-neighbor suspected data point of the suspected fluctuating temperature data b, and denote the suspected data point located on the right side of the first data point and closest to the first data point as the right near-neighbor suspected data point of the suspected fluctuating temperature data b. The near-neighbor suspected data points include the left near-neighbor suspected data point and the right near-neighbor suspected data point; Denote the interval formed by the abscissa value of the left neighboring suspected data point and the abscissa value of the first data point as the left interval, and denote the interval formed by the abscissa value of the right neighboring suspected data point and the abscissa value of the first data point as the right interval; in the second mapping coordinate system, denote the set formed by all temperature data points whose abscissa values belong to the left interval as the first feature set, and denote the set formed by all temperature data points whose abscissa values belong to the right interval as the second feature set, and the feature set includes the first feature set and the second feature set.

6. The intelligent operation and maintenance management system of the water-cooling system based on the cloud platform according to claim 4, characterized in that, The method for obtaining the target fluctuation characterization value of the suspected fluctuating temperature data b includes: Perform a negative correlation mapping on the result of adding the first variance and the second variance, and denote the mapping result as the first index value; Denote the Euclidean distance between the left neighboring suspected data point and the first data point as the left neighboring distance, and denote the Euclidean distance between the right neighboring suspected data point and the first data point as the right neighboring distance; denote the negative correlation mapping value of the result obtained by adding the left neighboring distance and the right neighboring distance as the second index value; Denote the absolute value of the difference between the mean value of the ordinate values of all data points in the first feature set and the ordinate value of the first data point as the first difference value; denote the absolute value of the difference between the mean value of the ordinate values of all data points in the second feature set and the ordinate value of the first data point as the second difference value; denote the negative correlation mapping value of the result obtained by adding the first difference value and the second difference value as the third index value; Denote the normalized value of the result obtained by adding the first index value, the second index value, and the third index value as the target fluctuation characterization value of the suspected fluctuating temperature data b.

7. The intelligent operation and maintenance management system for the water-cooling system based on the cloud platform according to claim 1, wherein, The method for obtaining the data sequence to be analyzed of the temperature data sequence and the fluctuating data to be analyzed in the data sequence to be analyzed includes: For any suspected fluctuating temperature data, if the target fluctuation characterization value of the suspected fluctuating temperature data is not less than the preset second fluctuation threshold, then determine that the suspected fluctuating temperature data is fluctuating temperature data, otherwise, determine that the suspected fluctuating temperature data is noise temperature data; Obtain the replacement value of the noise temperature data, and the replacement value of the noise temperature data is the mean value of the adjacent temperature data pairs of the noise temperature data, and the adjacent temperature data pairs of the noise temperature data are composed of the temperature data adjacent to and on the left of the noise temperature data and the temperature data adjacent to and on the right of the noise temperature data; Replace each noise temperature data in the temperature data sequence with the replacement value of the corresponding noise temperature data, and denote the new sequence obtained after the replacement as the data sequence to be analyzed of the temperature data sequence; If the f-th temperature data in the temperature data sequence is fluctuating temperature data, then denote the f-th data to be analyzed in the data sequence to be analyzed of the temperature data sequence as the fluctuating data to be analyzed.

8. The intelligent operation and maintenance management system of the water-cooling system based on the cloud platform according to claim 1, characterized in that, The method for obtaining the coefficient adjustment factor of the fluctuating data to be analyzed includes: According to the order of each data to be analyzed in the data sequence to be analyzed, obtain the characteristic marker values of each data to be analyzed in the data sequence to be analyzed, and the characteristic marker value of the g-th data to be analyzed in the data sequence to be analyzed is g; In all the data sequences to be analyzed, divide all the data to be analyzed with the same characteristic marker value into the same set, and all are denoted as characteristic subsets. Take the characteristic marker value of the data to be analyzed in the characteristic subset as the identification value of the corresponding characteristic subset; Use the clustering algorithm to cluster the data to be analyzed in each characteristic subset respectively to obtain the corresponding clustering clusters of each characteristic subset; For any fluctuating data to be analyzed h in any data sequence to be analyzed: Denote the characteristic subset with the same identification value as the characteristic marker value of the fluctuating data to be analyzed h as the target subset corresponding to the fluctuating data to be analyzed h; Among the clustering clusters corresponding to the target subset, denote the clustering cluster containing the fluctuating data to be analyzed h as the clustering cluster to be analyzed; Denote the ratio of the total number of data in the clustering cluster to be analyzed to the total number of data in the target subset as the first ratio; Denote the reciprocal of the number of clustering clusters corresponding to the target subset as the second ratio; In the clustering cluster to be analyzed, obtain the nearest neighbor data to be analyzed of each data to be analyzed in the clustering cluster to be analyzed, and denote it as the nearest neighbor data within the cluster corresponding to the corresponding data to be analyzed; Denote the absolute value of the difference between each data to be analyzed in the clustering cluster to be analyzed and its corresponding nearest neighbor data within the cluster as the characteristic difference of the corresponding data to be analyzed; Denote the mean value of the characteristic differences of all the data to be analyzed in the clustering cluster to be analyzed as the first mean value, and denote the negative correlation mapping value of the first mean value as the first mapping value; Denote the normalized value of the result obtained by adding the first ratio, the second ratio, and the first mapping value as the coefficient adjustment factor of the fluctuating data to be analyzed h.

9. The intelligent operation and maintenance management system of the water-cooled system based on the cloud platform according to claim 1, characterized in that The method for obtaining the second coefficient of each data to be analyzed in the data sequence to be analyzed includes: For any data to be analyzed in any data sequence to be analyzed, if the data to be analyzed does not belong to the fluctuating data to be analyzed, take the first coefficient of the data to be analyzed as the second coefficient of the data to be analyzed. If the data to be analyzed belongs to the fluctuating data to be analyzed, take the product of the first coefficient of the data to be analyzed and the coefficient adjustment factor of the data to be analyzed as the second coefficient of the data to be analyzed; The first coefficient of the data to be analyzed is the coefficient calculated when the data sequence to be analyzed is imported into the ARIMA prediction model.

10. The intelligent operation and maintenance management system of the water-cooling system based on the cloud platform according to claim 1, characterized in that, The method for obtaining the predicted coolant temperature data of the water cooling device at the future monitoring moment includes: For any water cooling device, according to the data sequence to be analyzed of the temperature data sequence corresponding to the water cooling device, the second coefficients of each data to be analyzed in the data sequence to be analyzed corresponding to the temperature data sequence of the water cooling device, and the ARIMA prediction model, obtain the predicted coolant temperature data of the water cooling device at the future monitoring moment.

Images