CN118265275A - Air cooling line control cooling optimization control method - Google Patents
Air cooling line control cooling optimization control method Download PDFInfo
- Publication number
- CN118265275A CN118265275A CN202410690375.9A CN202410690375A CN118265275A CN 118265275 A CN118265275 A CN 118265275A CN 202410690375 A CN202410690375 A CN 202410690375A CN 118265275 A CN118265275 A CN 118265275A
- Authority
- CN
- China
- Prior art keywords
- time
- time sequence
- real
- vector
- feature vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001816 cooling Methods 0.000 title claims abstract description 352
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000005457 optimization Methods 0.000 title claims abstract description 33
- 230000004044 response Effects 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 9
- 239000013598 vector Substances 0.000 claims description 291
- 230000004927 fusion Effects 0.000 claims description 13
- 238000000605 extraction Methods 0.000 claims description 12
- 239000011159 matrix material Substances 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 10
- 238000007499 fusion processing Methods 0.000 claims description 9
- 238000003062 neural network model Methods 0.000 claims description 7
- 230000001537 neural effect Effects 0.000 claims description 6
- 238000010606 normalization Methods 0.000 claims description 3
- 230000017525 heat dissipation Effects 0.000 abstract description 11
- 230000008569 process Effects 0.000 abstract description 7
- 238000004134 energy conservation Methods 0.000 abstract description 6
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 238000013473 artificial intelligence Methods 0.000 abstract description 4
- 238000013135 deep learning Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000010219 correlation analysis Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000005265 energy consumption Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000002035 prolonged effect Effects 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000007792 addition Methods 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 238000005215 recombination Methods 0.000 description 1
- 230000006798 recombination Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000003252 repetitive effect Effects 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- H05K7/00—Constructional details common to different types of electric apparatus
- H05K7/20—Modifications to facilitate cooling, ventilating, or heating
- H05K7/20709—Modifications to facilitate cooling, ventilating, or heating for server racks or cabinets; for data centers, e.g. 19-inch computer racks
- H05K7/20836—Thermal management, e.g. server temperature control
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- H05K7/00—Constructional details common to different types of electric apparatus
- H05K7/20—Modifications to facilitate cooling, ventilating, or heating
- H05K7/20009—Modifications to facilitate cooling, ventilating, or heating using a gaseous coolant in electronic enclosures
- H05K7/20209—Thermal management, e.g. fan control
-
- H—ELECTRICITY
- H05—ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
- H05K—PRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
- H05K7/00—Constructional details common to different types of electric apparatus
- H05K7/20—Modifications to facilitate cooling, ventilating, or heating
- H05K7/20709—Modifications to facilitate cooling, ventilating, or heating for server racks or cabinets; for data centers, e.g. 19-inch computer racks
- H05K7/20718—Forced ventilation of a gaseous coolant
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2123/00—Data types
- G06F2123/02—Data types in the time domain, e.g. time-series data
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Microelectronics & Electronic Packaging (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Thermal Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Hardware Design (AREA)
- Air Conditioning Control Device (AREA)
Abstract
The invention provides an air cooling line control cooling optimal control method, and relates to the technical field of intelligent control. Comprising the following steps: acquiring a time sequence of state parameters of the air-cooled line through a sensor network; and adjusting the fan rotating speed of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to enable the actual cooling capacity of the air cooling line to be close to the theoretical cooling capacity. The state parameters of the air-cooled lines are monitored and collected in real time through a sensor network, a data processing and analyzing algorithm based on artificial intelligence and deep learning technology is introduced into the rear end to perform time sequence collaborative correlation analysis on the state parameters of the air-cooled lines, and nonlinear response compensation is performed on extracted time sequence characteristics of the parameters in the process, so that the actual cold quantity is estimated and controlled more accurately. Therefore, the real-time monitoring and intelligent optimization control of the running state of the air cooling line can be realized, so that the heat dissipation efficiency of the air cooling line is improved, and the stability and energy conservation of the air cooling line are ensured.
Description
Technical Field
The invention relates to the technical field of intelligent control, in particular to an air cooling wire control cooling optimal control method.
Background
The air cooling line is a common device in a data center heat dissipation system, and air is sent into cooling equipment through a fan, and cold air is blown through a server rack to take away heat generated by a server, so that the temperature of the server is reduced. Conventional air cooling line control methods are typically based on simple heuristic rules or empirical settings, meaning that these methods rely on manually defined rules or empirical-based thresholds to control the air cooling line fan speed. Specifically, heuristic rules refer to rules formulated empirically or intuitively, which are typically based on simple observations or assumptions about system behavior. For example, one common heuristic is to set the air cooler fan speed to be proportional to the server rack temperature. When the server rack temperature increases, the fan speed also increases to remove more heat. Whereas empirical settings refer to specific parameter values obtained based on experience or trial and error. For example, one common empirical setting is to set the air cooled line fan speed to a fixed value that is believed to provide adequate heat dissipation.
However, these conventional methods have a disadvantage in that they cannot utilize timing information and association relationships between various state parameter data of the device. In addition, as the sensor has nonlinear response when collecting data, the control precision of the rotating speed of the fan can be influenced, and the control precision of the rotating speed of the fan can not reach the preset requirement, so that the cooling effect is reduced. In addition, the traditional air cooling line control cooling method lacks an intelligent control mode, and cannot be adjusted dynamically in response according to the actual running condition of the air cooling line, so that the difference between the actual cooling capacity of the air cooling line and the theoretical cooling capacity is large, and the air cooling line control cooling effect is affected. These drawbacks can result in the air cooling lines not responding effectively to changes in data center conditions, resulting in reduced cooling effectiveness. In addition, overcooling or undercooling can lead to energy waste. Insufficient cooling can also cause the servers to overheat, affecting performance.
Accordingly, an air-cooled, wire-controlled, optimal control scheme is desired.
Disclosure of Invention
In order to solve the technical problems, the invention provides an air cooling line control cooling optimal control method.
In a first aspect, an air-cooling drive-by-wire cooling optimization control method is provided, which includes:
acquiring a time sequence of state parameters of the air-cooled line through a sensor network;
And adjusting the rotating speed of a fan of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to enable the actual cooling capacity of the air cooling line to be close to the theoretical cooling capacity.
Optionally, adjusting the fan speed of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to make the actual cooling capacity of the air cooling line approach to the theoretical cooling capacity, including:
Determining actual cooling capacity based on a time sequence of state parameters of the air cooling line, wherein the state parameters of the air cooling line comprise a real-time temperature value, a real-time humidity value and a real-time wind speed value;
determining the actual cooling efficiency of the air cooling line based on a comparison between the actual cooling capacity and the theoretical cooling capacity;
and adjusting the rotating speed of a fan of the air cooling line based on the actual cooling efficiency of the air cooling line and the preset target cooling efficiency.
Optionally, determining the actual cooling capacity based on the time sequence of the state parameters of the air-cooled line includes:
data normalization is carried out on the time sequence of the state parameters of the air cooling line according to the dimension of the state parameter sample, so that a real-time temperature time sequence input vector, a real-time humidity time sequence input vector and a real-time wind speed time sequence input vector are obtained;
The time sequence mode feature extractor based on the deep neural network model is used for extracting features of the real-time temperature time sequence input vector, the real-time humidity time sequence input vector and the real-time wind speed time sequence input vector respectively so as to obtain a real-time temperature time sequence associated feature vector, a real-time humidity time sequence associated feature vector and a real-time wind speed time sequence associated feature vector;
Nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector so as to obtain a compensated real-time temperature time sequence related characteristic vector, a compensated real-time humidity time sequence related characteristic vector and a compensated real-time wind speed time sequence related characteristic vector;
Performing autocorrelation attention fusion processing on the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector to obtain an air cooling line global state time sequence related characteristic;
And determining the estimated value of the actual cooling capacity based on the time sequence related characteristics of the air cooling line global state.
Optionally, the feature extraction is performed on the real-time temperature time sequence input vectors through a time sequence pattern feature extractor based on a deep neural network model to obtain real-time temperature time sequence associated feature vectors, which comprises the following steps:
Processing the real-time temperature time sequence input vector by a time sequence mode feature extractor based on a one-dimensional extended convolutional neural model through a feature extraction formula to obtain the real-time temperature time sequence associated feature vector;
The feature extraction formula is as follows:
Wherein X 1,X2,...,Xn represents each real-time temperature local time sequence input vector in the real-time temperature time sequence input vectors, X 1∶n represents a cascade vector of each real-time temperature local time sequence input vector, X i∶i+j-l represents a cascade vector of X i,Xi+1,...,Xi+j-1 in each real-time temperature local time sequence input vector, ω and b represent a weight matrix and an offset vector, f (·) represents a convolution operation, C i represents each real-time temperature local time sequence correlation feature vector in the real-time temperature time sequence correlation feature vectors, and C is the real-time temperature time sequence correlation feature vector.
Optionally, performing nonlinear response compensation on the real-time temperature time sequence related feature vector, the real-time humidity time sequence related feature vector and the real-time wind speed time sequence related feature vector to obtain a compensated real-time temperature time sequence related feature vector, a compensated real-time humidity time sequence related feature vector and a compensated real-time wind speed time sequence related feature vector, including:
Nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector by using a correction formula so as to obtain the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector;
the correction formula is:
Wherein v si is the feature value of the i position of the s-th time sequence correlation feature vector in the real-time temperature time sequence correlation feature vector, the real-time humidity time sequence correlation feature vector and the real-time wind speed time sequence correlation feature vector, A, B, C and D are the adjustment super parameters, and v i is the feature value of the i position of the s-th corrected time sequence correlation feature vector in the compensated real-time temperature time sequence correlation feature vector, the compensated real-time humidity time sequence correlation feature vector and the compensated real-time wind speed time sequence correlation feature vector.
Optionally, performing autocorrelation attention fusion processing on the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector to obtain an air cooling line global state time sequence associated feature, including:
Inputting the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector into an autocorrelation attention fusion network to perform autocorrelation attention fusion processing according to a fusion formula so as to obtain an air cooling line global state time sequence related characteristic vector as the air cooling line global state time sequence related characteristic;
the fusion formula is as follows:
wherein h i is the ith time sequence associated feature vector in the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, And W i represents a weight coefficient vector and a weight coefficient matrix respectively, B i is an offset vector, tanh (·) represents a tanh function, e i is an attention score value of an ith time sequence associated feature vector, exp (·) represents an exponential operation of the vector, e k is an attention score value of a kth time sequence associated feature vector, t is the vector numbers of the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, a i is a weight value of the ith time sequence associated feature vector, and v 1 is the air cooling line global state time sequence associated feature vector.
Optionally, determining the estimated value of the actual cooling capacity based on the air-cooled line global state time sequence association feature includes:
and the air cooling line global state time sequence related characteristic vector passes through an actual cooling capacity estimator based on a decoder to obtain an estimated value of the actual cooling capacity.
Optionally, passing the air cooling line global state time sequence associated feature vector through an actual cooling capacity estimator based on a decoder to obtain an estimated value of the actual cooling capacity, including:
Using an actual cold quantity estimator based on a decoder to carry out decoding regression on the air cooling line global state time sequence associated feature vector by a decoding formula so as to obtain an estimated value of the actual cold quantity;
the decoding formula is: y= Σ (W V d + B), wherein V d represents the global state timing related feature vector of the air cooling line, Y represents the estimated value of the actual cooling capacity, W represents a weight matrix, B represents a bias vector,Representing a matrix multiplication.
In a second aspect, an air-cooled drive-by-wire cooling optimization control system is provided, which includes:
the state parameter acquisition module is used for acquiring a time sequence of state parameters of the air-cooled line through the sensor network;
and the fan rotating speed adjusting module is used for adjusting the fan rotating speed of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to enable the actual cooling capacity of the air cooling line to be close to the theoretical cooling capacity.
Compared with the prior art, the air cooling line control cooling optimizing control method provided by the invention has the advantages that the time sequence of the state parameters of the air cooling line is obtained through the sensor network, the fan rotating speed of the air cooling line is regulated based on the time sequence of the state parameters of the air cooling line, so that the actual cooling capacity of the air cooling line is close to the theoretical cooling capacity, the fan rotating speed of the air cooling line is intelligently controlled, the fan rotating speed is optimized through accurately predicting the actual cooling capacity of the air cooling line, the heat dissipation efficiency of the air cooling line can be improved, the stable operation of a server or other equipment can be ensured, the operation state of the air cooling line is monitored in real time, the abnormal condition is timely found, the stability of the air cooling line can be enhanced, and the equipment fault caused by poor heat dissipation is prevented. Unnecessary energy consumption can be reduced by optimizing the rotation speed of the fan, so that the energy conservation of the air cooling line is improved. Through real-time monitoring and intelligent control, the service life of the air cooling line can be prolonged, and the maintenance cost is reduced.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of an air-cooling drive-by-wire cooling optimization control method provided by an embodiment of the invention.
Fig. 2 is a flowchart of step 120 in the air-cooling drive-by-wire cooling optimization control method according to an embodiment of the present invention.
Fig. 3 is a block diagram of an air-cooled drive-by-wire cooling optimization control system provided by an embodiment of the invention.
Fig. 4 is a schematic diagram of a scenario of an air-cooled drive-by-wire cooling optimization control method according to an embodiment of the present invention.
Detailed Description
The following description of the technical solutions according to the embodiments of the present invention will be given with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Fig. 1 is a flowchart of an air-cooling drive-by-wire cooling optimization control method provided by an embodiment of the invention. As shown in fig. 1, the air-cooling drive-by-wire cooling optimization control method includes: 110, acquiring a time sequence of state parameters of the air-cooled line through a sensor network; 120, adjusting the fan rotating speed of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to enable the actual cooling capacity of the air cooling line to be close to the theoretical cooling capacity.
In the step 110, to ensure that the sensor network can reliably and accurately collect the state parameters of the air-cooled line, a state parameter capable of capturing time series data of key aspects of the operation condition of the air-cooled line is selected. In this way, a comprehensive view of the air cooling line operating conditions may be provided in order to detect and analyze patterns and trends in the air cooling line behavior.
In the step 120, a control algorithm is used, which can predict the cooling demand of the air-cooled line by using the time series data, and the nonlinear response of the sensor in the air-cooled line is considered, so as to optimize the control algorithm to minimize the difference between the actual cooling capacity and the theoretical cooling capacity of the air-cooled line. Therefore, the cooling effect of the air cooling line can be improved, the energy waste is reduced, and the performance of the server is improved.
Accordingly, in the process of performing adaptive adjustment of the fan speed of the air-cooled line, it is important to determine the actual cooling capacity based on the time sequence of the state parameters of the air-cooled line, because the actual cooling capacity is a direct representation of the actual heat dissipation capacity of the air-cooled line. The current heat dissipation performance of the air cooling line can be accurately known by determining the actual cooling capacity, so that the rotating speed of the fan is dynamically adjusted. Specifically, if the actual cooling capacity is lower than the theoretical cooling capacity, the cooling efficiency of the air cooling line is insufficient, and the rotating speed of the fan needs to be increased.
Fig. 2 is a flowchart of step 120 in the air-cooling drive-by-wire cooling optimization control method according to an embodiment of the present invention. As shown in fig. 2, step 120, adjusting a fan speed of the air cooling line based on a time sequence of state parameters of the air cooling line to make an actual cooling capacity of the air cooling line approach to a theoretical cooling capacity, includes: 121, determining an actual cooling capacity based on a time sequence of state parameters of the air cooling line, wherein the state parameters of the air cooling line comprise a real-time temperature value, a real-time humidity value and a real-time wind speed value; 122, determining the actual cooling efficiency of the air cooling line based on the comparison between the actual cooling capacity and the theoretical cooling capacity; 123, adjusting the fan rotating speed of the air cooling line based on the actual cooling efficiency of the air cooling line and the preset target cooling efficiency.
The step 121 uses an accurate and reliable formula or model to calculate the actual cooling capacity, considers the nonlinear response of the sensor in the air cooling line, and determines the actual cooling capacity based on the time sequence of the state parameters of the air cooling line. In this way, an accurate estimate of the actual cooling performance of the air-cooled line may be provided, allowing for detection and analysis of changes in the cooling capacity of the air-cooled line. The step 122 uses the theoretical cooling capacity of the air cooling line as a reference for comparison. The actual cooling efficiency of the air-cooled line is determined based on a comparison between the actual cooling capacity and a theoretical cooling capacity, taking into account the operating conditions of the air-cooled line, such as ambient temperature and humidity. In this way, a quantitative measure of the cooling efficiency of the air-cooled line may be provided, allowing the identification of the cause of the reduced or degraded performance of the air-cooled line. Step 123 adjusts the fan speed of the air cooling line based on the actual cooling efficiency of the air cooling line and a preset target cooling efficiency. Therefore, the cooling effect of the air cooling line can be improved, the energy waste is reduced, and the performance of the server is improved.
In the process of determining the actual cooling capacity based on the time sequence of the state parameters of the air cooling line, the state parameters of the air cooling line comprise a real-time temperature value, a real-time humidity value and a real-time wind speed value, wherein the state parameters have a time sequence dynamic change rule in a time dimension, that is, the time sequence data of the real-time temperature value, the real-time humidity value and the real-time wind speed value have time sequence association relations at different time points respectively.
In a specific embodiment of the present invention, determining the actual cooling capacity based on the time sequence of the state parameters of the air-cooled line includes: data normalization is carried out on the time sequence of the state parameters of the air cooling line according to the dimension of the state parameter sample, so that a real-time temperature time sequence input vector, a real-time humidity time sequence input vector and a real-time wind speed time sequence input vector are obtained; the time sequence mode feature extractor based on the deep neural network model is used for extracting features of the real-time temperature time sequence input vector, the real-time humidity time sequence input vector and the real-time wind speed time sequence input vector respectively so as to obtain a real-time temperature time sequence associated feature vector, a real-time humidity time sequence associated feature vector and a real-time wind speed time sequence associated feature vector; nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector so as to obtain a compensated real-time temperature time sequence related characteristic vector, a compensated real-time humidity time sequence related characteristic vector and a compensated real-time wind speed time sequence related characteristic vector; performing autocorrelation attention fusion processing on the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector to obtain an air cooling line global state time sequence related characteristic; and determining the estimated value of the actual cooling capacity based on the time sequence related characteristics of the air cooling line global state.
In the technical scheme of the invention, the time sequence of the state parameters (such as temperature, humidity and wind speed) of the air cooling line is regulated according to the dimension of the state parameter sample, so that the original parameter time sequence data can be converted into a format suitable for deep learning model processing, the structure and the characteristics of the data can be better utilized, and the model can be helped to better capture the association and the change of the time sequence data of the state parameters at different time points.
Further, the one-dimensional convolution layer is used for respectively carrying out feature analysis and extraction on the real-time temperature time sequence input vector, the real-time humidity time sequence input vector and the real-time wind speed time sequence input vector so as to respectively capture local time sequence features of each state parameter in a local time period, including local time sequence modes and features of temperature, humidity and wind speed, and the system can be helped to better understand the state time sequence change rule of the air cooling line. In particular, in order to extract higher-level, more abstract and more comprehensive features from time series data of these state parameters so as to better understand state time sequence changes of an air cooling line, in the technical scheme of the invention, the real-time temperature time sequence input vector, the real-time humidity time sequence input vector and the real-time wind speed time sequence input vector are processed by a time sequence pattern feature extractor based on a one-dimensional extended convolutional neural model so as to obtain a real-time temperature time sequence related feature vector, a real-time humidity time sequence related feature vector and a real-time wind speed time sequence related feature vector. It should be appreciated that the one-dimensional extended convolutional neural model introduces a super-parameter named expansion rate, as compared to the one-dimensional convolutional neural model, to control the number of values 0 in the convolutional kernel. The expansion rate defines the intervals among the state parameters of each time point when the convolution kernel processes the time sequence data, and after the convolution operation is expanded, the time sequence correlation characteristics of the state parameters far away can be captured, the generalization capability of the model is improved, and the feature dimension reduction is facilitated.
In a specific embodiment of the present invention, the feature extraction of the real-time temperature time sequence input vector by the time sequence pattern feature extractor based on the deep neural network model to obtain the real-time temperature time sequence associated feature vector includes: processing the real-time temperature time sequence input vector by a time sequence mode feature extractor based on a one-dimensional extended convolutional neural model through a feature extraction formula to obtain the real-time temperature time sequence associated feature vector; the feature extraction formula is as follows:
wherein X 1,X2,...,Xn represents each real-time temperature local time sequence input vector in the real-time temperature time sequence input vectors, X 1∶n represents a cascade vector of each real-time temperature local time sequence input vector, X i∶i+j-l represents a cascade vector of X i,Xi+1,...,Xi+j-1 in each real-time temperature local time sequence input vector, ω and b represent a weight matrix and an offset vector, f (·) represents a convolution operation, C i represents each real-time temperature local time sequence correlation feature vector in the real-time temperature time sequence correlation feature vectors, and C is the real-time temperature time sequence correlation feature vector. Here, it should be understood that the above-described feature extraction formula may also be referred to by a process of performing feature extraction on the real-time humidity time series input vector and the real-time wind speed time series input vector by a time series pattern feature extractor based on a deep neural network model, respectively.
Then, it is considered that different state parameters may have different data distribution ranges, and at the same time, different sensors may have differences in sensitivity, response speed, and the like, and also have nonlinear responses, so that acquired data have deviation. Therefore, in order to eliminate or reduce measurement errors and nonlinear response errors between different sensors in a data processing process, so as to ensure the accuracy and reliability of data, and standardize the data to a uniform range at the same time, avoid the influence of numerical value differences between different parameters on model training. It will be appreciated that by gamma correction, the differences between these sensors can be eliminated, making the data more consistent and comparable. In addition, because in actual data, state parameter data may have deflection or nonlinear relation, nonlinear influence and error in the data can be eliminated through gamma correction, so that the data is more in accordance with the assumption of a linear model, the fitting effect of the model is improved, and the data is more accurate and reliable.
In a specific embodiment of the present invention, performing nonlinear response compensation on the real-time temperature time-series related feature vector, the real-time humidity time-series related feature vector, and the real-time wind speed time-series related feature vector to obtain a compensated real-time temperature time-series related feature vector, a compensated real-time humidity time-series related feature vector, and a compensated real-time wind speed time-series related feature vector, including: nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector by using a correction formula so as to obtain the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector; the correction formula is:
Wherein v si is the feature value of the i position of the s-th time sequence correlation feature vector in the real-time temperature time sequence correlation feature vector, the real-time humidity time sequence correlation feature vector and the real-time wind speed time sequence correlation feature vector, A, B, C and D are the adjustment super parameters, and v i is the feature value of the i position of the s-th corrected time sequence correlation feature vector in the compensated real-time temperature time sequence correlation feature vector, the compensated real-time humidity time sequence correlation feature vector and the compensated real-time wind speed time sequence correlation feature vector.
It should be appreciated that the global state of the air cooling line may be affected by a number of factors, including temperature, humidity, wind speed, and the like. The compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector respectively comprise time sequence characteristic information about different state parameters after nonlinear compensation. Therefore, in order to take the relevance among the time sequence characteristics of different state parameters into consideration, so as to better model the global state of the air cooling line and capture the time sequence relevance characteristics among the global states, in the technical scheme of the invention, the compensated real-time temperature time sequence relevance characteristic vector, the compensated real-time humidity time sequence relevance characteristic vector and the compensated real-time wind speed time sequence relevance characteristic vector are input into an autocorrelation attention fusion network so as to obtain the air cooling line global state time sequence relevance characteristic vector. It should be appreciated that the autocorrelation attentiveness mechanism may help the model automatically learn importance weights between time series features of different state parameters, enabling the model to focus more on features that are more critical to the current task, which helps to improve the characterization ability and prediction accuracy of the model. That is, the self-correlation attention fusion network is used for processing, so that the correlation information between time sequence correlation feature vectors of different state parameters can be effectively fused, the model is helped to better understand the correlation between different features, the comprehensive and comprehensive global state feature representation of the air cooling line is obtained, and more powerful support is provided for subsequent state prediction and analysis.
In a specific embodiment of the present invention, performing an autocorrelation attention fusion process on the compensated real-time temperature time sequence related feature vector, the compensated real-time humidity time sequence related feature vector, and the compensated real-time wind speed time sequence related feature vector to obtain an air cooling line global state time sequence related feature, including: inputting the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector into an autocorrelation attention fusion network to perform autocorrelation attention fusion processing according to a fusion formula so as to obtain an air cooling line global state time sequence related characteristic vector as the air cooling line global state time sequence related characteristic; the fusion formula is as follows:
wherein h i is the ith time sequence associated feature vector in the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, And W i represents a weight coefficient vector and a weight coefficient matrix respectively, B i is an offset vector, tanh (·) represents a tanh function, e i is an attention score value of an ith time sequence associated feature vector, exp (·) represents an exponential operation of the vector, e k is an attention score value of a kth time sequence associated feature vector, t is the vector numbers of the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, a i is a weight value of the ith time sequence associated feature vector, and v 1 is the air cooling line global state time sequence associated feature vector.
In one embodiment of the present invention, determining the estimated value of the actual cooling capacity based on the air-cooled line global state timing correlation feature includes: and the air cooling line global state time sequence related characteristic vector passes through an actual cooling capacity estimator based on a decoder to obtain an estimated value of the actual cooling capacity.
Specifically, the air cooling line global state time sequence associated feature vector passes through an actual cooling capacity estimator based on a decoder to obtain an estimated value of the actual cooling capacity. That is, the decoding regression is performed through the global state time sequence correlation characteristic of the air cooling line, so that the actual cooling capacity is predicted more accurately. And further, determining the actual cooling efficiency of the air cooling line based on the comparison between the actual cooling capacity and the theoretical cooling capacity, and adjusting the fan rotating speed of the air cooling line based on the actual cooling efficiency of the air cooling line and the preset target cooling efficiency. Therefore, the real-time monitoring and intelligent optimization control of the running state of the air cooling line can be realized, so that the heat dissipation efficiency of the air cooling line is improved, and the stability and energy conservation of the air cooling line are ensured.
In a specific embodiment of the present invention, the method for obtaining the estimated value of the actual cooling capacity by passing the air cooling line global state time sequence associated feature vector through an actual cooling capacity estimator based on a decoder includes: performing decoding regression on the air cooling line global state time sequence association characteristic vector by using the actual cooling capacity estimator based on the decoder according to a decoding formula so as to obtain an estimated value of the actual cooling capacity;
the decoding formula is: y= Σ (W V d + B), wherein V d represents the global state timing related feature vector of the air cooling line, Y represents the estimated value of the actual cooling capacity, W represents a weight matrix, B represents a bias vector,Representing a matrix multiplication.
In the above technical solution, the compensated real-time temperature time sequence related feature vector, the compensated real-time humidity time sequence related feature vector and the compensated real-time wind speed time sequence related feature vector respectively express the time sequence related features of the real-time temperature value, the real-time humidity value and the real-time wind speed value of the air cooling line after nonlinear feature response correction, so that after the compensated real-time temperature time sequence related feature vector, the compensated real-time humidity time sequence related feature vector and the compensated real-time wind speed time sequence related feature vector pass through an autocorrelation attention fusion network, autocorrelation attention strengthening can be further performed based on the time sequence distribution features of the real-time temperature value, the real-time humidity value and the real-time wind speed value of the air cooling line, so as to emphasize the significance and the importance of the time sequence feature representation under the local sample space domain.
In this way, after the real-time temperature value, the real-time humidity value and the real-time wind speed value of the air cooling line are considered, after time sequence correlation characteristic extraction and nonlinear characteristic response correction, larger characteristic distribution inconsistency is introduced into time sequence correlation characteristic representation, so that after the compensated real-time temperature time sequence correlation characteristic vector, the compensated real-time humidity time sequence correlation characteristic vector and the compensated real-time wind speed time sequence correlation characteristic vector are input into an autocorrelation attention fusion network, the characteristic representation significance of time sequence characteristics of certain state parameters in the air cooling line global state time sequence correlation characteristic vector is improved, and meanwhile, the air cooling line global state time sequence correlation characteristic vector deviates from the original characteristic source domain distribution of time sequence characteristics of state parameters, so that the overall distribution regression constraint of the air cooling line global state time sequence correlation characteristic vector relative to a decoding regression domain is insufficient.
Preferably, in an exemplary embodiment, the method for obtaining the estimated value of the actual cooling capacity by passing the air cooling line global state time sequence associated feature vector through an actual cooling capacity estimator based on a decoder includes the following steps: performing point adding operation on the square root of the length of the air cooling line global state time sequence related characteristic vector and the reciprocal of the square root of the two norms of the air cooling line global state time sequence related characteristic vector to obtain an air cooling line global state time sequence related offset characteristic vector; calculating an exponential function based on a natural constant of the air cooling line global state time sequence associated offset characteristic vector to obtain an air cooling line global state time sequence associated class offset characteristic vector; performing point multiplication operation on a norm and a weight super parameter of the air cooling line global state time sequence related characteristic vector and the air cooling line global state time sequence related characteristic vector to obtain an air cooling line global state time sequence related boundary characteristic vector; performing point adding operation on the air cooling line global state time sequence related class offset feature vector and the air cooling line global state time sequence related boundary feature vector to obtain an optimized air cooling line global state time sequence related feature vector; and the optimized global state time sequence associated feature vector of the air cooling line passes through an actual cooling capacity estimator based on a decoder so as to obtain an estimated value of the actual cooling capacity.
In the above-described exemplary embodiment, the local canonical coordinates of each feature value of the air-cooling line global state time series related feature vector are represented by the structured norm of the air-cooling line global state time series related feature vector, the vector overall distribution representation of the air-cooling line global state time series related feature vector is determined based on the scale model and the rotational offset of the feature distribution model relative to the feature values, so as to set the offset prediction direction for each feature value of the air-cooling line global state time series related feature vector serving as the center, and the feature value constraint is performed for the overall distribution boundary box representation by the vector feature value of the air-cooling line global state time series related feature vector, so that the constraint of the air-cooling line global state time series related feature vector under the overall decoding regression distribution is improved, and the training speed of the model and the accuracy of the actual cold quantity estimated value of the air-cooling line global state time series related feature vector obtained by the actual cold quantity estimator based on the decoder are improved.
In summary, the air-cooling line control cooling optimization control method provided by the embodiment of the invention is clarified, wherein the state parameters of the air-cooling line are monitored and collected in real time through a sensor network, the state parameters comprise a real-time temperature value, a real-time humidity value and a real-time wind speed value, a data processing and analyzing algorithm based on artificial intelligence and deep learning technology is introduced into the rear end to perform time sequence collaborative association analysis on the state parameters of the air-cooling line, and in the process, nonlinear response compensation is performed on extracted parameter time sequence characteristics, so that the actual cooling capacity is estimated and controlled more accurately. Therefore, the real-time monitoring and intelligent optimization control of the running state of the air cooling line can be realized, so that the heat dissipation efficiency of the air cooling line is improved, and the stability and energy conservation of the air cooling line are ensured.
According to the invention, the state parameters of the air cooling line are monitored in real time through the sensor network, and the time sequence collaborative correlation analysis is carried out by combining the artificial intelligence and the deep learning technology, so that the actual cooling capacity of the air cooling line can be accurately predicted, the rotating speed of the fan can be optimized, the heat dissipation efficiency of the air cooling line can be improved, and the stable operation of a server or other equipment is ensured. The running state of the air cooling line is monitored in real time, and abnormal conditions are found in time, so that the stability of the air cooling line can be enhanced, and equipment faults caused by poor heat dissipation are prevented. Unnecessary energy consumption can be reduced by optimizing the rotation speed of the fan, so that the energy conservation of the air cooling line is improved. Through real-time monitoring and intelligent control, the service life of the air cooling line can be prolonged, and the maintenance cost is reduced. The analysis algorithm based on the artificial intelligence and the deep learning technology can automatically detect and diagnose the problem of the air cooling line, improve the operation and maintenance efficiency and reduce the manual intervention.
In one embodiment of the present invention, fig. 3 is a block diagram of an air-cooled wire-controlled cooling optimization control system provided in an embodiment of the present invention. As shown in fig. 3, an air-cooling drive-by-wire cooling optimization control system 200 provided in an embodiment of the present invention includes: a state parameter obtaining module 210, configured to obtain a time sequence of state parameters of the air-cooled line through the sensor network; the fan speed adjustment module 220 of the air cooling line is configured to adjust a fan speed of the air cooling line based on a time sequence of state parameters of the air cooling line, so that an actual cooling capacity of the air cooling line is close to a theoretical cooling capacity.
Here, it will be understood by those skilled in the art that the specific functions and operations of the respective modules in the above-described air-cooling-by-wire cooling-optimization control system have been described in detail in the above description of the air-cooling-by-wire cooling-optimization control method with reference to fig. 1 to 2, and thus, repetitive descriptions thereof will be omitted.
As described above, the air-cooling drive-by-wire cooling optimization control system 200 provided by the embodiment of the invention can be implemented in various terminal devices, for example, a server for air-cooling drive-by-wire cooling optimization control, and the like. In one example, the air-cooled wire-controlled cooling optimization control system 200 provided by the embodiment of the present invention may be integrated into a terminal device as a software module and/or a hardware module. For example, the air-cooled, wire-cooled, optimal control system 200 may be a software module in the operating system of the terminal device, or may be an application developed for the terminal device; of course, the air-cooled wire-controlled cooling optimization control system 200 can also be one of a plurality of hardware modules of the terminal device.
Alternatively, in another example, the air-cooled, wire-controlled, optimal control system 200 and the terminal device may be separate devices, and the air-cooled, wire-controlled, optimal control system 200 may be connected to the terminal device via a wired and/or wireless network and transmit the interaction information in a agreed data format.
In one embodiment of the present invention, there is provided an air-cooling drive-by-wire cooling optimization control method including the steps of:
(1) According to parameters such as real-time temperature, humidity and wind speed of the air cooling line, calculating the actual cooling capacity and theoretical cooling capacity of the air cooling line;
(2) Comparing the actual cooling capacity with the theoretical cooling capacity to obtain the cooling efficiency of the air-out cooling line;
(3) According to the cooling efficiency and the preset target cooling efficiency, control parameters such as the fan rotating speed, the water pump flow, the valve opening and the like of the air cooling line are adjusted, so that the actual cooling capacity of the air cooling line is close to the theoretical cooling capacity, and the aim of optimizing control is fulfilled;
(4) And repeating the steps, and monitoring and adjusting the running state of the air cooling line in real time, so as to ensure the stability and energy conservation of the air cooling line.
The system comprises a central controller and a plurality of air cooling units, wherein each air cooling unit comprises a fan, a water pump, a valve and a sensor, the sensor is used for collecting parameters such as temperature, humidity and wind speed of the air cooling unit, the central controller is used for receiving signals of the sensor, calculating the actual cooling capacity and the theoretical cooling capacity of an air cooling line, comparing the difference of the actual cooling capacity and the theoretical cooling capacity, outputting control signals according to preset target cooling efficiency, adjusting control parameters such as fan rotating speed, water pump flow and valve opening, and realizing optimal control of the air cooling line.
Fig. 4 is a schematic diagram of a scenario of an air-cooled drive-by-wire cooling optimization control method according to an embodiment of the present invention. As shown in fig. 4, in this application scenario, first, a time sequence of state parameters of an air-cooled line is acquired through a sensor network (e.g., C as illustrated in fig. 4); the time series of acquired state parameters of the air-cooled line is then input into a server (e.g., S as illustrated in fig. 4) deployed with an air-cooled line-controlled cooling optimization control algorithm, wherein the server is capable of processing the time series of state parameters of the air-cooled line based on the air-cooled line-controlled cooling optimization control algorithm to adjust a fan speed of the air-cooled line to bring an actual cooling capacity of the air-cooled line close to a theoretical cooling capacity.
It is also noted that in the apparatus, devices and methods of the present invention, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent aspects of the present invention.
The previous description is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the invention. Thus, the present invention is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit embodiments of the invention to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.
Claims (6)
1. An air-cooling line-control cooling optimization control method is characterized by comprising the following steps:
acquiring a time sequence of state parameters of the air-cooled line through a sensor network;
adjusting the fan rotating speed of the air cooling line based on the time sequence of the state parameters of the air cooling line so as to enable the actual cooling capacity of the air cooling line to be close to the theoretical cooling capacity;
Wherein, based on the time sequence of the state parameter of the forced air cooling line, adjust the fan rotational speed of forced air cooling line to make the actual cold volume of forced air cooling line be close to theoretical cold volume, include:
Determining actual cooling capacity based on a time sequence of state parameters of the air cooling line, wherein the state parameters of the air cooling line comprise a real-time temperature value, a real-time humidity value and a real-time wind speed value;
determining the actual cooling efficiency of the air cooling line based on a comparison between the actual cooling capacity and the theoretical cooling capacity;
Based on the actual cooling efficiency of the air cooling line and a preset target cooling efficiency, adjusting the rotating speed of a fan of the air cooling line;
wherein determining the actual cooling capacity based on the time sequence of the state parameters of the air cooling line comprises:
data normalization is carried out on the time sequence of the state parameters of the air cooling line according to the dimension of the state parameter sample, so that a real-time temperature time sequence input vector, a real-time humidity time sequence input vector and a real-time wind speed time sequence input vector are obtained;
The time sequence mode feature extractor based on the deep neural network model is used for extracting features of the real-time temperature time sequence input vector, the real-time humidity time sequence input vector and the real-time wind speed time sequence input vector respectively so as to obtain a real-time temperature time sequence associated feature vector, a real-time humidity time sequence associated feature vector and a real-time wind speed time sequence associated feature vector;
Nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector so as to obtain a compensated real-time temperature time sequence related characteristic vector, a compensated real-time humidity time sequence related characteristic vector and a compensated real-time wind speed time sequence related characteristic vector;
Performing autocorrelation attention fusion processing on the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector to obtain an air cooling line global state time sequence related characteristic;
And determining the estimated value of the actual cooling capacity based on the time sequence related characteristics of the air cooling line global state.
2. The air-cooling drive-by-wire cooling optimization control method according to claim 1, wherein the step of extracting features of the real-time temperature time sequence input vectors by a time sequence pattern feature extractor based on a deep neural network model to obtain real-time temperature time sequence associated feature vectors comprises the steps of:
Processing the real-time temperature time sequence input vector by a time sequence mode feature extractor based on a one-dimensional extended convolutional neural model through a feature extraction formula to obtain the real-time temperature time sequence associated feature vector;
The feature extraction formula is as follows:
Wherein X 1,X2,...,Xn represents each real-time temperature local time sequence input vector in the real-time temperature time sequence input vectors, X 1∶n represents a cascade vector of each real-time temperature local time sequence input vector, X i∶i+j-l represents a cascade vector of X i,Xi+1,...,Xi+j-1 in each real-time temperature local time sequence input vector, ω and b represent a weight matrix and an offset vector, f (·) represents a convolution operation, C i represents each real-time temperature local time sequence correlation feature vector in the real-time temperature time sequence correlation feature vectors, and C is the real-time temperature time sequence correlation feature vector.
3. The air-cooling drive-by-wire cooling optimization control method according to claim 2, wherein performing nonlinear response compensation on the real-time temperature-time-sequence-related feature vector, the real-time humidity-time-sequence-related feature vector, and the real-time wind speed-time-sequence-related feature vector to obtain a compensated real-time temperature-time-sequence-related feature vector, a compensated real-time humidity-time-sequence-related feature vector, and a compensated real-time wind speed-time-sequence-related feature vector, comprises:
Nonlinear response compensation is carried out on the real-time temperature time sequence related characteristic vector, the real-time humidity time sequence related characteristic vector and the real-time wind speed time sequence related characteristic vector by using a correction formula so as to obtain the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector;
the correction formula is:
Wherein v si is the feature value of the i position of the s-th time sequence correlation feature vector in the real-time temperature time sequence correlation feature vector, the real-time humidity time sequence correlation feature vector and the real-time wind speed time sequence correlation feature vector, A, B, C and D are the adjustment super parameters, and v i is the feature value of the i position of the s-th corrected time sequence correlation feature vector in the compensated real-time temperature time sequence correlation feature vector, the compensated real-time humidity time sequence correlation feature vector and the compensated real-time wind speed time sequence correlation feature vector.
4. The air-cooled drive-by-wire cooling optimization control method of claim 3, wherein performing autocorrelation attention fusion processing on the compensated real-time temperature time sequence-related feature vector, the compensated real-time humidity time sequence-related feature vector, and the compensated real-time wind speed time sequence-related feature vector to obtain an air-cooled drive-by-wire global state time sequence-related feature, comprises:
Inputting the compensated real-time temperature time sequence related characteristic vector, the compensated real-time humidity time sequence related characteristic vector and the compensated real-time wind speed time sequence related characteristic vector into an autocorrelation attention fusion network to perform autocorrelation attention fusion processing according to a fusion formula so as to obtain an air cooling line global state time sequence related characteristic vector as the air cooling line global state time sequence related characteristic;
the fusion formula is as follows:
wherein h i is the ith time sequence associated feature vector in the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, And W i represents a weight coefficient vector and a weight coefficient matrix respectively, B i is an offset vector, tanh (·) represents a tanh function, e i is an attention score value of an ith time sequence associated feature vector, exp (·) represents an exponential operation of the vector, e k is an attention score value of a kth time sequence associated feature vector, t is the vector numbers of the compensated real-time temperature time sequence associated feature vector, the compensated real-time humidity time sequence associated feature vector and the compensated real-time wind speed time sequence associated feature vector, a i is a weight value of the ith time sequence associated feature vector, and v 1 is the air cooling line global state time sequence associated feature vector.
5. The air-cooled drive-by-wire cooling optimization control method of claim 4, wherein determining the estimated value of the actual cooling capacity based on the air-cooled drive-by-wire global state timing correlation characteristics comprises:
and the air cooling line global state time sequence related characteristic vector passes through an actual cooling capacity estimator based on a decoder to obtain an estimated value of the actual cooling capacity.
6. The air-cooled drive-by-wire cooling optimization control method of claim 5, wherein passing the air-cooled drive-by-wire global state timing related feature vector through an actual cooling capacity estimator based on a decoder to obtain the estimated value of the actual cooling capacity, comprises:
Using an actual cold quantity estimator based on a decoder to carry out decoding regression on the air cooling line global state time sequence associated feature vector by a decoding formula so as to obtain an estimated value of the actual cold quantity;
the decoding formula is: y= Σ (W V d + B), wherein V d represents the global state timing related feature vector of the air cooling line, Y represents the estimated value of the actual cooling capacity, W represents a weight matrix, B represents a bias vector,Representing a matrix multiplication.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410690375.9A CN118265275B (en) | 2024-05-30 | 2024-05-30 | Air cooling line control cooling optimization control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410690375.9A CN118265275B (en) | 2024-05-30 | 2024-05-30 | Air cooling line control cooling optimization control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118265275A true CN118265275A (en) | 2024-06-28 |
| CN118265275B CN118265275B (en) | 2024-08-27 |
Family
ID=91603617
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410690375.9A Active CN118265275B (en) | 2024-05-30 | 2024-05-30 | Air cooling line control cooling optimization control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118265275B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118497447A (en) * | 2024-07-19 | 2024-08-16 | 山西新泰钢铁有限公司 | High-efficiency dust removing method for converter of steel mill |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140079533A1 (en) * | 2012-09-18 | 2014-03-20 | Nec Computertechno, Ltd. | Electronic device and method for controlling temperature of electronic device |
| US20200301487A1 (en) * | 2017-09-25 | 2020-09-24 | Eizo Corporation | Ambient temperature estimating device, ambient temperature estimating method, program and system |
| EP3850926A1 (en) * | 2018-09-13 | 2021-07-21 | Bripco (UK) Limited | Data centre |
| CN114818515A (en) * | 2022-06-24 | 2022-07-29 | 中国海洋大学 | Multidimensional time sequence prediction method based on self-attention mechanism and graph convolution network |
| CN116538127A (en) * | 2023-06-16 | 2023-08-04 | 湖州越球电机有限公司 | Axial flow fan and control system thereof |
| CN117170473A (en) * | 2023-09-21 | 2023-12-05 | 深圳汉光电子技术有限公司 | High heat dispersion's respiratory server |
| CN117404853A (en) * | 2023-12-14 | 2024-01-16 | 山西省水利建筑工程局集团有限公司 | External circulating water cooling system and method for tunnel boring machine |
| CN117610884A (en) * | 2023-12-22 | 2024-02-27 | 国网河南省电力公司信息通信分公司 | Power dispatching system and method based on data management platform |
| CN117743772A (en) * | 2023-12-29 | 2024-03-22 | 维达纸业(浙江)有限公司 | Toilet paper drying parameter optimization method and system based on artificial intelligent model |
-
2024
- 2024-05-30 CN CN202410690375.9A patent/CN118265275B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140079533A1 (en) * | 2012-09-18 | 2014-03-20 | Nec Computertechno, Ltd. | Electronic device and method for controlling temperature of electronic device |
| US20200301487A1 (en) * | 2017-09-25 | 2020-09-24 | Eizo Corporation | Ambient temperature estimating device, ambient temperature estimating method, program and system |
| EP3850926A1 (en) * | 2018-09-13 | 2021-07-21 | Bripco (UK) Limited | Data centre |
| CN114818515A (en) * | 2022-06-24 | 2022-07-29 | 中国海洋大学 | Multidimensional time sequence prediction method based on self-attention mechanism and graph convolution network |
| CN116538127A (en) * | 2023-06-16 | 2023-08-04 | 湖州越球电机有限公司 | Axial flow fan and control system thereof |
| CN117170473A (en) * | 2023-09-21 | 2023-12-05 | 深圳汉光电子技术有限公司 | High heat dispersion's respiratory server |
| CN117404853A (en) * | 2023-12-14 | 2024-01-16 | 山西省水利建筑工程局集团有限公司 | External circulating water cooling system and method for tunnel boring machine |
| CN117610884A (en) * | 2023-12-22 | 2024-02-27 | 国网河南省电力公司信息通信分公司 | Power dispatching system and method based on data management platform |
| CN117743772A (en) * | 2023-12-29 | 2024-03-22 | 维达纸业(浙江)有限公司 | Toilet paper drying parameter optimization method and system based on artificial intelligent model |
Non-Patent Citations (2)
| Title |
|---|
| 宋奎勇: "面向试验数据的多源信息融合方法研究", 工程科技Ⅱ辑, 15 May 2024 (2024-05-15) * |
| 柳亦兵: "风速时间序列的非线性特性分析", 华北电力大学学报(自然科学版), 30 November 2008 (2008-11-30) * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118497447A (en) * | 2024-07-19 | 2024-08-16 | 山西新泰钢铁有限公司 | High-efficiency dust removing method for converter of steel mill |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118265275B (en) | 2024-08-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN118265275B (en) | Air cooling line control cooling optimization control method | |
| CN111352408B (en) | Multi-working-condition process industrial process fault detection method based on evidence K nearest neighbor | |
| KR101941854B1 (en) | System and method of estimating load with null data correction | |
| CN117387172B (en) | Terminal air conditioner energy saving method and system based on accurate recommended equipment control parameters | |
| CN118312746B (en) | Equipment state evaluation method and system based on digital twin | |
| CN115358281B (en) | Machine learning-based cold and hot all-in-one machine monitoring control method and system | |
| CN112070353A (en) | Method and system for accurately detecting energy efficiency of data center | |
| CN116882079A (en) | Water pump characteristic curve self-adaptive calibration and prediction method | |
| MirhoseiniNejad et al. | ALTM: Adaptive learning-based thermal model for temperature predictions in data centers | |
| CN118171873A (en) | Machine room energy-saving operation and maintenance management method and system | |
| CN112560339B (en) | Method for predicting guide bearing bush temperature of hydroelectric generating set by utilizing machine learning | |
| CN113390641A (en) | Intelligent early warning and online diagnosis method and system for equipment faults of wind and smoke system | |
| CN111623905B (en) | Wind turbine generator bearing temperature early warning method and device | |
| CN116213095B (en) | Intelligent clean coal product ash content adjusting method and system based on dense medium separation | |
| CN115907138B (en) | Method, system and medium for predicting PUE value of data center | |
| CN116085290A (en) | Sliding window thermal imaging-based fan thermal fault detection method and system | |
| CN116432524A (en) | Transformer oil temperature prediction method, device, equipment and storage medium | |
| CN113486953B (en) | Method and device for predicting replacement time of filter screen of frequency converter and computer readable medium | |
| CN118134044B (en) | Intelligent biological laboratory instrument sharing management method and system | |
| CN112255141A (en) | Thermal imaging gas monitoring system | |
| CN118132987B (en) | Heat energy meter heat data acquisition method and system | |
| CN117111661B (en) | Centralized control system and method for production workshops | |
| CN117338600B (en) | Drug stir-frying control system and control method | |
| CN118426318B (en) | Internet of things service self-adaptive system based on process mining | |
| CN116150666B (en) | Energy storage system fault detection method and device and intelligent terminal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |