Intelligent optimization and early warning method and system for computer room capacity based on machine learning

Abstract

The present invention belongs to the technical field of computer room capacity assessment and early warning. A method and system for intelligent optimization and early warning of computer room capacity based on machine learning is provided. The weighted sum of the computer room space utilization prediction results, the cooling efficiency prediction results, and the power utilization prediction results is used as the computer room capacity utilization, and the space, cooling, and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization; when any one of the computer room space utilization prediction results, the cooling efficiency prediction results, and the power utilization prediction results is less than a set threshold, an alarm message is generated; the present invention realizes accurate prediction and real-time monitoring of computer room capacity, as well as intelligent early warning when the capacity is close to the critical point, by establishing a three-dimensional data model of computer room space, cooling, and power distribution, thereby improving the operational efficiency and safety of the data center.

Description

Technical Field

[0001] The present invention relates to the technical field of computer room capacity assessment and early warning, and in particular to a computer room capacity intelligent optimization early warning method and system based on machine learning. Background Art

[0002] The statements in this section merely provide background art related to the present invention and do not necessarily constitute prior art.

[0003] Computer room capacity assessment can help enterprises fully understand the actual utilization of computer room resources, including the number of cabinets, equipment deployment density in the cabinets, power supply system capacity, etc. Through assessment, enterprises can identify areas where resources are under-utilized or over-utilized, and take corresponding optimization measures to improve resource utilization efficiency and reduce resource waste.

[0004] Accurate data center capacity assessment helps enterprises predict future capacity needs and develop corresponding expansion plans. This ensures that when business needs grow, the data center can provide sufficient resource support in a timely manner to avoid business interruptions or performance degradation due to insufficient resources. At the same time, capacity assessment can also help enterprises identify potential risk points and take corresponding preventive measures to ensure business continuity and stability.

[0005] The current capacity management of data center computer rooms lacks an effective real-time prediction and monitoring mechanism, and is unable to accurately assess the space, cooling and power distribution capabilities of the computer room, which in turn leads to unreasonable resource allocation, equipment overload and energy waste. Summary of the invention

[0006] In order to address the shortcomings of the prior art, the present invention provides a method and system for intelligent optimization and early warning of computer room capacity based on machine learning. By establishing a three-dimensional data model of computer room space, refrigeration and power distribution, accurate prediction and real-time monitoring of computer room capacity, as well as intelligent early warning when the capacity approaches the critical point, are achieved, thereby improving the operational efficiency and safety of the data center.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] In a first aspect, the present invention provides a method for intelligent optimization and early warning of computer room capacity based on machine learning.

[0009] A machine learning-based intelligent optimization and early warning method for computer room capacity includes the following processes:

[0010] According to the computer room space data and the pre-trained first machine learning model, the computer room space utilization prediction result is obtained;

[0011] According to the refrigeration data and the pre-trained second machine learning model, a refrigeration efficiency prediction result is obtained;

[0012] According to the power distribution data and the pre-trained third machine learning model, the power utilization rate prediction result is obtained;

[0013] The weighted sum of the computer room space utilization prediction results, the cooling efficiency prediction results and the power utilization prediction results is used as the computer room capacity utilization, and the space, cooling and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization.

[0014] When any one of the predicted results of the room space utilization, the cooling efficiency, and the power utilization is less than a set threshold, an alarm message is generated.

[0015] As a further limitation of the first aspect of the present invention, the first machine learning model and the third machine learning model both adopt long short-term memory neural network models, and the second machine learning model adopts a feedforward neural network model.

[0016] As a further limitation of the first aspect of the present invention, the computer room space data includes: the number of cabinets in the computer room, the cabinet size, the cabinet U position (cabinet U position is the unit used to represent the vertical space in the server cabinet, which is the abbreviation of Unit and represents a standard vertical height unit) and the computer room load-bearing capacity.

[0017] As a further limitation of the first aspect of the present invention, the refrigeration data includes: the number of air conditioners, the total refrigeration volume and the real-time refrigeration demand.

[0018] As a further limitation of the first aspect of the present invention, the power distribution data includes: active power, reactive power and apparent power of each power distribution cabinet, UPS (Uninterruptible Power Supply) load, and the bus and branch lines of the first cabinet.

[0019] As a further limitation of the first aspect of the present invention, the process of computer room control optimization includes: computer room space utilization constraint, cooling efficiency constraint and power utilization constraint.

[0020] In a second aspect, the present invention provides a computer room capacity intelligent optimization early warning system based on machine learning.

[0021] A machine learning-based intelligent optimization and early warning system for computer room capacity includes the following processes:

[0022] The computer room space utilization prediction unit is configured to: obtain a computer room space utilization prediction result according to the computer room space data and a pre-trained first machine learning model;

[0023] A refrigeration efficiency prediction unit is configured to: obtain a refrigeration efficiency prediction result according to the refrigeration data and a pre-trained second machine learning model;

[0024] The power utilization rate prediction unit is configured to: obtain a power utilization rate prediction result according to the power distribution data and the pre-trained third machine learning model;

[0025] The control optimization unit is configured to optimize the space, cooling and power control parameters of the computer room with the goal of maximizing the computer room capacity utilization rate according to the weighted sum of the computer room space utilization rate prediction result, the cooling efficiency prediction result and the power utilization rate prediction result as the computer room capacity utilization rate;

[0026] The early warning unit is configured to generate an alarm message when any one of the prediction results of the computer room space utilization, the prediction results of the cooling efficiency and the prediction results of the power utilization is less than a set threshold.

[0027] In a third aspect, the present invention provides a computer device, comprising: a processor and a computer-readable storage medium;

[0028] a processor adapted to execute a computer program;

[0029] A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the method for intelligent optimization and early warning of computer room capacity based on machine learning as described in the first aspect of the present invention is implemented.

[0030] In a fourth aspect, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program is suitable for being loaded by a processor and executing the machine learning-based intelligent optimization and early warning method for computer room capacity as described in the first aspect of the present invention.

[0031] In a fifth aspect, the present invention provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the machine learning-based intelligent optimization and early warning method for computer room capacity as described in the first aspect of the present invention.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. The present invention innovatively proposes an intelligent optimization and early warning method for computer room capacity based on machine learning. By establishing a three-dimensional data model of computer room space, refrigeration and power distribution, it realizes accurate prediction and real-time monitoring of computer room capacity, as well as intelligent early warning when the capacity approaches the critical point, thereby improving the operational efficiency and safety of the data center.

[0034] 2. The present invention can accurately evaluate the capacity of the computer room through data modeling and algorithm prediction, and can effectively guide the rational allocation of resources; the present invention realizes real-time monitoring of key indicators of the computer room and ensures the transparency of the operating status.

[0035] 3. The present invention can give a timely warning when the capacity is close to saturation, avoid equipment overload and energy waste, and ensure the safe operation of the computer room; the present invention optimizes the cooling strategy and power distribution, improves energy utilization efficiency, and meets the development requirements of green data centers.

[0036] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0038] Figure 1 A schematic diagram of the process of the machine room capacity intelligent optimization and early warning method based on machine learning provided in Example 1 of the present invention;

[0039] Figure 2 A schematic diagram of a machine room capacity intelligent optimization early warning system based on machine learning provided in Example 2 of the present invention;

[0040] Figure 3 A schematic diagram of a computer device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0041] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0042] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0043] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0044] Embodiment 1:

[0045] This implementation proposes a machine learning-based intelligent optimization and early warning method for computer room capacity. Figure 1 As shown, the following process is included:

[0046] S1: Obtain the prediction result of the computer room space utilization rate based on the computer room space data and the pre-trained first machine learning model;

[0047] S2: Obtain a cooling efficiency prediction result based on the cooling data and the pre-trained second machine learning model;

[0048] S3: Obtain power utilization prediction results based on power distribution data and a pre-trained third machine learning model;

[0049] S4: taking the weighted sum of the prediction results of the computer room space utilization rate, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate as the computer room capacity utilization rate, and optimizing the space, cooling, and power control parameters of the computer room with the goal of maximizing the computer room capacity utilization rate;

[0050] S5: When any one of the prediction results of the space utilization rate of the equipment room, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate is less than a set threshold, an alarm message is generated.

[0051] In this implementation, step S1 specifically includes:

[0052] Collect physical space data such as the load-bearing capacity of the computer room, the number of cabinets, and the U-position of the cabinets. At the same time, record in detail the dimensions of each cabinet and the dimensions of the equipment inside it, as well as the necessary safety distance between each device.

[0053] In this implementation, 3D modeling software is used to digitize the layout of the computer room, and detailed information such as the location, size, and load-bearing capacity of the cabinets are reflected in the model. The LSTM (Long Short Term Memory) model is used to predict the physical limitations of future equipment deployment. The model is trained with historical data to predict the trend of computer room space utilization. The LSTM model can process time series data and is suitable for analyzing the changing trend of equipment deployment over time. Specifically, the implementation process includes:

[0054] S1.1: Collect data related to the computer room space, including the number of cabinets, cabinet size, cabinet U position, and computer room load-bearing capacity;

[0055] S1.2: Build an LSTM model. The input data includes features related to the computer room space, including the number of cabinets, cabinet size, cabinet U position, and computer room load-bearing capacity.

[0056] S1.3: Use historical data related to the computer room space to train the LSTM model to predict the space constraints of future equipment deployment;

[0057] S1.5: Predict the changing trend of computer room space utilization based on the trained LSTM model;

[0058] S1.6: Output room space utilization S (i.e., the predicted result of computer room space utilization), that is, the degree of computer room space utilization, expressed as a percentage.

[0059] In this implementation, LSTM is a more advanced RNN (Recurrent Neural Network), which can effectively process sequence data. LSTM can selectively store information through gating devices to avoid gradient diffusion and gradient explosion problems. The state value of the node at a certain moment is not only related to the current input value, but also to Hidden layer node output at time Relevant, representing the impact of historical information on the present.

[0060] The forget gate, input gate, and output gate of the long short-term memory neural network work together to control and protect the state of the long short-term memory neural network cells. The forget gate is a gating device used to calculate the degree of information retention and control the historical state. Flow to How many entries are allowed after the time Status at all times ;

[0061] The formula of the forget gate is:

[0062] (1);

[0063] in, represents the forget gate gating function, is the current input vector, is the output vector of the previous layer, containing historical state information, Refers to the sigmoid activation function, is the weight matrix of the forget gate, is the bias term of the forget gate;

[0064] After being processed by the sigmoid activation function, it becomes a value between 0 and 1, where 0 means completely off and 1 means completely on. If the forget gate is completely off and takes a value of 0, the history has no effect on the current state, that is, it is all forgotten. If the forget gate is completely open and takes a value of 1, the historical information is transmitted intact to time t without any information loss, that is, it is all retained. The forget gate is more likely to take a value between 0 and 1, indicating a partial influx of historical information.

[0065] The input gate is a gating device used to calculate which information is saved in the state unit, controlling the input The above information includes two parts: one part can be regarded as how much information of the current input needs to be saved to the unit state, which can be understood as input gating, and its formula is:

[0066] (2);

[0067] The other part is used to add the new information generated by the current input to the cell state, which can be understood as the input value. The formula is:

[0068] (3);

[0069] The state update function formula is:

[0070] (4)

[0071] in, is the input gate gating function, is the current input value at time t, and the multiplication of the two represents How much current input information is allowed to enter at any time to determine In the current state, if the input gate is fully closed and takes the value 0, LSTM ignores the influence of the current input, which is equivalent to skipping the input gate directly without seeing this input; if the input gate is fully open and takes the value 1, the current input determines the current state to the maximum extent. , without any information loss; the input gate is more likely to take values ​​between 0 and 1, representing a partial inflow of input information. is the current input vector, is the output vector of the previous layer, containing historical state information, Refers to the sigmoid activation function, refers to the tanh activation function, and is the weight matrix of the forget gate, and is the bias term of the input gate, is the previous state, is the forget gate gating function.

[0072] The output gate is a gating device used to calculate the degree to which information is output at the current moment. Status value at the moment It is more or less visible to the outside world. The formula for the output gate is:

[0073] (5);

[0074] (6);

[0075] in, is the output gate gating function, is the output value of the hidden layer node, is the output vector of the previous layer, containing historical state information, refers to the tanh activation function, is the weight matrix of the output gate, is the bias term of the output gate, is the current status.

[0076] In order to measure the performance of the prediction model, the present invention uses the root mean square error (RMSE) as the performance evaluation index to evaluate the model, and the formula is:

[0077] (7);

[0078] in, N represents the number of samples, Represents the water demand in a certain period of time. The representative model predicts the drinking water demand in this time period. The RMSE value reflects the error between the predicted value and the actual value, which can measure the accuracy of the predicted value.

[0079] Optionally, in this implementation, the Keras deep learning network framework is used to build an LSTM model, adding 1 LSTM layer, 1 Leaky Relu layer, 1 Dropout layer, and 2 Dense layers. The LSTM input layer inputs data, the hidden layer sets 360 LSTM neurons, and the output layer performs single-value prediction; the Dropout layer discards 40% of the calculated data to prevent overfitting; the activation function of the Dense layer is set to ReLu and sigmoid. The configuration of the learning process is completed through the compile method, the loss function uses mean_squared_error, the optimizer uses Adam, and the training time is set to 100 cycles.

[0080] In this implementation, step S2 specifically includes:

[0081] In this implementation, the number of air conditioners, total cooling capacity, and real-time cooling demand are collected, where the total cooling capacity is composed of indoor equipment load and environmental heat load. In this implementation, the temperature difference and airflow distribution in different areas are considered, and a feedforward neural network is constructed to adjust the cooling parameters according to real-time monitoring data and optimize the efficiency of the cooling system. Through neural network model training, the cooling parameters can be adjusted more accurately according to the current environmental conditions.

[0082] The implementation steps include:

[0083] S2.1: Collect refrigeration-related data, including the number of air conditioners, total refrigeration capacity, and real-time refrigeration demand;

[0084] S2.2: Considering the temperature difference and airflow distribution in different areas of the room, a feedforward neural network is constructed. The input data includes the number of air conditioners, the total cooling volume, and the real-time cooling demand.

[0085] S2.3: Use historical data to train a feedforward neural network to optimize cooling parameters;

[0086] S2.4: Predict refrigeration system efficiency based on the trained model;

[0087] S2.5: Output cooling efficiency C (i.e. refrigeration efficiency prediction result), i.e. refrigeration system efficiency, expressed as a percentage.

[0088] The feedforward neural network used in this implementation is composed of a multi-layer neuron structure. Each layer of neurons has input (its input is the output of the previous layer of neurons) and output. The learning process consists of two processes: forward propagation of the signal and back propagation of the error. The input is from the input layer, processed by the hidden layer, and then transmitted to the output layer. If the actual output of the output layer does not match the expected output, it enters the back propagation stage of the error. The error back propagation is to propagate the output error back to the input layer through the hidden layer in some form, and distribute the error to all units in each layer, so as to obtain the error signal of the units in each layer. This error signal is used as the basis for correcting the weights of each unit until the output error meets certain conditions or the number of iterations reaches a certain number.

[0089] In this implementation, the feedforward neural network used uses weights , , and deviation , , , the artificial neural network propagates forward from the input data X to the output through the following three components of the logistic activation function σ = σ(·) with independent variable z .

[0090] (8);

[0091] The activation function is as follows:

[0092] (9);

[0093] The logistic function z is as follows:

[0094] (10);

[0095] The update of the n-dimensional vector x performed by the LM algorithm is Given by:

[0096] (11);

[0097] where e(x) is the error vector and J(x) is the Jacobian matrix consisting of the partial derivatives of the error function in each of the n dimensions. is the transpose of J(x), μ is the parameter that is iteratively modified during the algorithm, and I is the identity matrix.

[0098] In this implementation, step S3 specifically includes:

[0099] Monitor the active power, reactive power, and apparent power of each distribution cabinet, UPS load, bus and branch line, analyze the power flow and distribution, and use the LSTM model to predict future power demand based on historical power consumption data. During the model training process, historical power consumption is used as input and the predicted value is used as output. The training model learns the power consumption pattern. After the model training is completed, the current power parameters can be input and the prediction of future power load can be output. The implementation steps, specifically, include:

[0100] S3.1: Collect data related to power distribution, including active power, reactive power and apparent power of each distribution cabinet, UPS load, bus and branch line;

[0101] S3.2: Analyze the flow and distribution of electricity and build an LSTM model. The input data includes features related to power consumption, such as active power, reactive power, and apparent power.

[0102] S3.3: Train the LSTM model using historical electricity consumption data to predict future electricity demand;

[0103] S3.4: predicting power utilization based on the trained model;

[0104] S3.5: Output power utilization P (i.e., power utilization prediction result), that is, the effective use of power in the computer room, expressed as a percentage.

[0105] It can be understood that the LSTM model here has the same structure as the model in step S1, except that the input and output are changed accordingly for training, which will not be repeated here.

[0106] In this implementation, step S4 and step S5 specifically include:

[0107] Data fusion: Integrate the data of space, cooling and power distribution models, comprehensively consider space, cooling and power constraints, maximize the utilization of computer room capacity and optimize, while ensuring that each subsystem (space, cooling, power) is within the safe operating range. By setting thresholds, when the capacity of any subsystem approaches the warning value, the system automatically triggers a warning signal to remind management personnel to take measures.

[0108] In this implementation, the optimization goal can be defined as maximizing the utilization of the computer room capacity. 、 、 It is a weight coefficient, which is set manually to balance the importance of different goals. Specifically:

[0109] Computer room capacity utilization = ×S+ ×C+ ×P , according to the maximum utilization of the computer room capacity, the utilization of the computer room space is carried out S , Refrigeration system efficiency C and effective use P Optimization of the various input items during prediction to obtain the optimal space, cooling and power control parameters.

[0110] Constraints: The optimization process needs to meet the following constraints to ensure that each subsystem is within the safe operating range: (1) Computer room space utilization S ≤ the first maximum allowed value S max ; (2) Refrigeration efficiency C ≥Minimum allowed value C min ; (3) Electricity utilization rate P ≤ the second maximum allowable value P max ;

[0111] The first maximum allowable value refers to the maximum allowable value of the computer room space utilization, which defines the space usage limit that the computer room should not exceed under any circumstances to ensure that there is enough margin to cope with future expansion needs and to maintain sufficient maintenance and operation space; for example, if the maximum allowable computer room space utilization is 85%, then S max Set it to 0.85;

[0112] The minimum allowable value is the minimum allowable value of cooling efficiency, which defines the minimum efficiency level that the cooling system must achieve to ensure that the equipment does not malfunction or be damaged due to overheating; for example, if the minimum efficiency requirement of the cooling system is 60%, then C min Set it to 0.6;

[0113] The second maximum allowable value refers to the maximum allowable value of power utilization, which defines the usage limit that the power system should not exceed under any circumstances to ensure the stability and security of power supply; for example, if the maximum allowable utilization of the power system is 90%, then P max Set it to 0.9.

[0114] Embodiment 2:

[0115] like Figure 2 As shown, this implementation method provides a machine learning-based intelligent optimization early warning system for computer room capacity, including the following processes:

[0116] The computer room space utilization prediction unit is configured to: obtain a computer room space utilization prediction result according to the computer room space data and a pre-trained first machine learning model;

[0117] A refrigeration efficiency prediction unit is configured to: obtain a refrigeration efficiency prediction result according to the refrigeration data and a pre-trained second machine learning model;

[0118] The power utilization rate prediction unit is configured to: obtain a power utilization rate prediction result according to the power distribution data and the pre-trained third machine learning model;

[0119] The control optimization unit is configured to optimize the space, cooling and power control parameters of the computer room with the goal of maximizing the computer room capacity utilization rate according to the weighted sum of the computer room space utilization rate prediction result, the cooling efficiency prediction result and the power utilization rate prediction result as the computer room capacity utilization rate;

[0120] The early warning unit is configured to generate an alarm message when any one of the prediction results of the computer room space utilization, the prediction results of the cooling efficiency and the prediction results of the power utilization is less than a set threshold.

[0121] It is understandable that the above-mentioned units can be separately or completely combined into one or several other units to constitute, or one (some) of the units can be further divided into multiple smaller units in function to constitute, which can achieve the same operation without affecting the realization of the technical effects of the embodiments of the present application. The above-mentioned units are divided based on logical functions. In practical applications, the functions of one unit can also be implemented by multiple units, or the functions of multiple units can be implemented by one unit. In other embodiments of the present application, the system may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented by the collaboration of multiple units.

[0122] According to another embodiment of the present application, the system described in this embodiment can be constructed, and the method of Example 1 of the present application can be implemented by running a computer program (including program code) capable of executing the steps involved in the corresponding method described in Example 1 on a general-purpose computing device such as a computer, which includes processing elements and storage elements such as a central processing unit (CPU), random access memory (RAM), and read-only memory (ROM). The computer program can be recorded on, for example, a computer-readable recording medium, and loaded into the above-mentioned computing device through the computer-readable recording medium and run therein.

[0123] Embodiment 3:

[0124] like Figure 3 As shown, this implementation provides an electronic device, which includes a processor 1001, a communication interface 1002, and a computer-readable storage medium 1003. The processor 1001, the communication interface 1002, and the computer-readable storage medium 1003 may be connected via a bus or other means.

[0125] Among them, the communication interface 1002 is used to receive and send data, the computer-readable storage medium 1003 can be stored in the memory of the electronic device, the computer-readable storage medium 1003 is used to store a computer program, the computer program includes program instructions, and the processor 1001 is used to execute the program instructions stored in the computer-readable storage medium 1003.

[0126] The processor 1001 (or CPU (Central Processing Unit)) is the computing core and control core of the electronic device, which is suitable for implementing one or more instructions, and specifically suitable for loading and executing one or more instructions to implement corresponding method processes or corresponding functions.

[0127] The processor 1001 is configured to execute the following process:

[0128] According to the computer room space data and the pre-trained first machine learning model, the computer room space utilization prediction result is obtained;

[0129] According to the refrigeration data and the pre-trained second machine learning model, a refrigeration efficiency prediction result is obtained;

[0130] According to the power distribution data and the pre-trained third machine learning model, the power utilization rate prediction result is obtained;

[0131] The weighted sum of the prediction results of the computer room space utilization rate, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate is used as the computer room capacity utilization rate, and the space, cooling, and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization rate;

[0132] When any one of the predicted results of the room space utilization, the cooling efficiency, and the power utilization is less than a set threshold, an alarm message is generated.

[0133] The specific method is described in Example 1 and will not be repeated here.

[0134] Embodiment 4:

[0135] This implementation provides a computer-readable storage medium (Memory), which is a memory device in an electronic device for storing programs and data. It is understandable that the computer-readable storage medium here can include both built-in storage media in the electronic device and, of course, extended storage media supported by the electronic device. The computer-readable storage medium provides a storage space that stores the processing system of the electronic device.

[0136] In addition, the storage space also stores one or more instructions suitable for being loaded and executed by the processor, and these instructions may be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here may be a high-speed RAM memory, or a non-volatile memory, such as at least one disk memory; optionally, it may also be at least one computer-readable storage medium located away from the aforementioned processor.

[0137] In one embodiment, the computer-readable storage medium stores one or more instructions; the processor loads and executes the one or more instructions stored in the computer-readable storage medium to implement the following process:

[0138] According to the computer room space data and the pre-trained first machine learning model, the computer room space utilization prediction result is obtained;

[0139] According to the refrigeration data and the pre-trained second machine learning model, a refrigeration efficiency prediction result is obtained;

[0140] According to the power distribution data and the pre-trained third machine learning model, the power utilization rate prediction result is obtained;

[0141] The weighted sum of the prediction results of the computer room space utilization rate, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate is used as the computer room capacity utilization rate, and the space, cooling, and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization rate;

[0142] When any one of the predicted results of the room space utilization, the cooling efficiency, and the power utilization is less than a set threshold, an alarm message is generated.

[0143] The specific method is described in Example 1 and will not be repeated here.

[0144] Embodiment 5:

[0145] The present implementation provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the following process:

[0146] According to the computer room space data and the pre-trained first machine learning model, the computer room space utilization prediction result is obtained;

[0147] According to the refrigeration data and the pre-trained second machine learning model, a refrigeration efficiency prediction result is obtained;

[0148] According to the power distribution data and the pre-trained third machine learning model, the power utilization rate prediction result is obtained;

[0149] The weighted sum of the prediction results of the computer room space utilization rate, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate is used as the computer room capacity utilization rate, and the space, cooling, and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization rate;

[0150] When any one of the predicted results of the room space utilization, the cooling efficiency, and the power utilization is less than a set threshold, an alarm message is generated.

[0151] The specific method is described in Example 1 and will not be repeated here.

[0152] A person skilled in the art can appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0153] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from a website site, a computer, a server or a data center to another website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by the computer or a data processing device such as a server, a data center, etc. that includes one or more available media integrated. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

[0154] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A machine learning-based intelligent optimization and early warning method for computer room capacity, characterized in that: The process includes: According to the computer room space data and the pre-trained first machine learning model, the computer room space utilization prediction result is obtained; The computer room space data includes: the number of cabinets in the computer room, the cabinet size, the cabinet U position and the load-bearing capacity of the computer room; Obtaining a refrigeration efficiency prediction result according to refrigeration data and a pre-trained second machine learning model; the refrigeration data includes: the number of air conditioners, the total refrigeration volume, and the real-time refrigeration demand; The power utilization rate prediction result is obtained according to the power distribution data and the pre-trained third machine learning model; the power distribution data includes: the active power, reactive power and apparent power of each distribution cabinet, UPS load, bus of the first cabinet and branch line; The weighted sum of the prediction results of the computer room space utilization rate, the prediction results of the cooling efficiency, and the prediction results of the power utilization rate is used as the computer room capacity utilization rate, and the space, cooling, and power control parameters of the computer room are optimized with the goal of maximizing the computer room capacity utilization rate. The specific formula is: Computer room capacity utilization rate = ， The corresponding constraints are: computer room space utilization S ≤ the first maximum allowed value S max Refrigeration efficiency C ≥Minimum allowed value C min ; Power utilization P ≤ the second maximum allowable value P max ; The first maximum allowable value refers to the maximum allowable value of the computer room space utilization rate; the minimum allowable value refers to the minimum allowable value of the cooling efficiency; the second maximum allowable value refers to the maximum allowable value of the power utilization rate; When any one of the predicted results of the room space utilization, the cooling efficiency, and the power utilization is less than a set threshold, an alarm message is generated.

2. The machine learning-based intelligent optimization and early warning method for computer room capacity according to claim 1, characterized in that: The first machine learning model and the third machine learning model both use long short-term memory neural network models, and the second machine learning model uses a feedforward neural network model.

3. A machine learning-based intelligent optimization and early warning system for computer room capacity, characterized in that: The process includes: The computer room space utilization prediction unit is configured to: obtain a computer room space utilization prediction result according to the computer room space data and a pre-trained first machine learning model; The computer room space data includes: the number of cabinets in the computer room, the cabinet size, the cabinet U position and the load-bearing capacity of the computer room; A refrigeration efficiency prediction unit is configured to obtain a refrigeration efficiency prediction result according to refrigeration data and a pre-trained second machine learning model; the refrigeration data includes: the number of air conditioners, the total refrigeration volume, and the real-time refrigeration demand; The power utilization rate prediction unit is configured to obtain a power utilization rate prediction result according to power distribution data and a pre-trained third machine learning model; the power distribution data includes active power, reactive power and apparent power of each power distribution cabinet, UPS load, bus of the first cabinet and branch line; The control optimization unit is configured to optimize the space, cooling and power control parameters of the computer room with the goal of maximizing the computer room capacity utilization rate according to the weighted sum of the computer room space utilization rate prediction result, the cooling efficiency prediction result and the power utilization rate prediction result as the computer room capacity utilization rate, and the specific formula is: computer room capacity utilization rate = ， The corresponding constraints are: computer room space utilization S ≤ the first maximum allowed value S max Refrigeration efficiency C ≥Minimum allowed value C min ; Power utilization P ≤ the second maximum allowable value P max ; The first maximum allowable value refers to the maximum allowable value of the computer room space utilization rate; the minimum allowable value refers to the minimum allowable value of the cooling efficiency; the second maximum allowable value refers to the maximum allowable value of the power utilization rate; The early warning unit is configured to generate an alarm message when any one of the prediction results of the computer room space utilization, the prediction results of the cooling efficiency and the prediction results of the power utilization is less than a set threshold.

4. A computer device, characterized in that: include: A processor and a computer readable storage medium; a processor adapted to execute a computer program; A computer-readable storage medium, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, the method for intelligent optimization and early warning of computer room capacity based on machine learning as described in any one of claims 1 to 2 is implemented.

5. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is suitable for being loaded by a processor and executing the machine learning-based intelligent optimization and early warning method for computer room capacity as described in any one of claims 1 to 2.

6. A computer program product, characterized in that The computer program product includes a computer program, and when the computer program is executed by a processor, it implements the machine learning-based intelligent optimization and early warning method for computer room capacity as described in any one of claims 1 to 2.

Images