An energy-saving control method, system, device and medium for computer room air conditioners

Abstract

The present invention provides an energy-saving control method, system, device and medium for a computer room. The method includes: obtaining the operation data of the computer room equipment, where the computer room equipment includes a cabinet equipment group and a refrigeration equipment group; performing standardization processing on the operation data to obtain standardized data; predicting the temperature of the computer room in the next state based on a preset first prediction subunit for the standardized data, and predicting the refrigeration energy consumption data in the next state based on a preset second prediction subunit for the standardized data; determining an optimal temperature control strategy according to the computer room temperature data and the refrigeration energy consumption data, and generating a corresponding control instruction according to the optimal temperature control strategy. The control instruction is used to control the refrigeration equipment group to operate under optimal operating parameters so that the cabinet environment temperature does not exceed a preset warning temperature and the total refrigeration power consumption is the lowest. The present invention can alleviate the problem of excessive refrigeration in the existing computer room, reduce power loss, and improve the energy-saving level of the computer room.

Description

Technical Field

[0001] The present invention relates to the field of energy conservation in big data and AI computer rooms, and particularly to an energy-saving control method, system, device and medium for computer room air conditioners. Background Technique

[0002] With the booming development of data centers and the promotion of green, low-carbon, energy-saving and emission-reduction policies, the problem of computer room energy consumption has become the biggest challenge. To avoid server failures in the computer room due to high temperatures, currently, the temperature of the cabinets is mainly adjusted by means of air supply and refrigeration of air conditioners. Due to the different loads of servers in each unit and the different distances between the servers and the refrigeration equipment, the temperatures at different positions of the cabinets are different, resulting in failures of some servers.

[0003] At present, the control and refrigeration output of computer room refrigeration equipment adopt traditional thermodynamic control models or fixed strategies, while the operating states of each cabinet in the computer room are diverse, with large fluctuations and great differences. Traditional refrigeration control strategies are mostly fixed modes, such as cooling by means of all-weather relatively excessive cooling, and cannot perform optimal refrigeration output in real time according to the real-time heat generation dynamics of the cabinets, resulting in insufficient refrigeration capacity when the servers are under high load and high heat, and redundant refrigeration output when the servers are under low load and low heat, resulting in a large amount of ineffective power consumption.

[0004] Therefore, researching the operating state data of the computer room, studying the operating modes of computer room equipment and refrigeration systems, how to accurately control each refrigeration equipment, and how to effectively utilize electric energy have great social value and significance for implementing the concept of green and low carbon. Summary of the Invention

[0005] Embodiments of the present invention provide an energy-saving control method, system, device and medium for computer room air conditioners to solve the problems existing in the related technologies. The technical solutions are as follows:

[0006] In the first aspect, embodiments of the present invention provide an energy-saving control method for a computer room air conditioner, including:

[0007] Obtain the operating data of computer room equipment, where the computer room equipment includes a cabinet equipment group and a refrigeration equipment group; perform standardization processing on the operating data to obtain standardized data;

[0008] Based on a preset first prediction subunit, perform temperature prediction on the standardized data to obtain the computer room temperature data in the next state, and based on a preset second prediction subunit, perform energy consumption prediction on the standardized data to obtain the refrigeration energy consumption data in the next state;

[0009] Determine the optimal temperature control strategy based on the computer room temperature data and the refrigeration energy consumption data, and generate corresponding control instructions according to the optimal temperature control strategy. The control instructions are used to control the refrigeration equipment group to operate under the optimal operating parameters so that the cabinet environment temperature does not exceed the preset warning temperature and the total refrigeration power consumption is the lowest.

[0010] In one embodiment, the operating data includes the cabinet equipment temperature, the air outlet temperature of the refrigeration equipment, the indoor temperature, the cabinet equipment power, and the refrigeration equipment power.

[0011] In one embodiment, the method of standardization processing includes:

[0012] Perform data standardization processing on the mean value of each original data and the standard deviation of each original data in the operating data by using Z-score standardization; or,

[0013] Perform non-linear scaling on the values of each original data in the operating data.

[0014] In one embodiment, the temperature prediction method is:

[0015] Take the air conditioner power at the current time, the air outlet wind speed of the air conditioner at the current time, and the cabinet temperature at the current time as independent variables, take the computer room temperature at the next moment as the dependent variable, and input the independent variables and the dependent variable into the trained first prediction subunit to obtain the computer room temperature data in the next state.

[0016] In one embodiment, the energy consumption prediction method is:

[0017] Take the air conditioner power at the current time, the air outlet wind speed of the air conditioner at the current time, and the cabinet load at the current time as independent variables, take the air conditioner energy consumption at the next moment as the dependent variable, and input the independent variables and the dependent variable into the trained second prediction subunit to obtain the refrigeration energy consumption data in the next state.

[0018] In one embodiment, the method for determining the optimal temperature control strategy is:

[0019] Search all possible values of the air conditioner operating parameters according to the genetic algorithm to obtain the optimal operating parameters;

[0020] Generate the optimal temperature control strategy according to the optimal operating parameters.

[0021] In one embodiment, the method for determining the optimal temperature control strategy is:

[0022] Traverse each possible value of the air conditioner operating parameters, input any possible value, the equipment load, and the meteorological conditions into the pre-constructed first model, and predict the computer room temperature under the corresponding conditions;

[0023] Filter out the possible values where the computer room temperature is lower than the set threshold to obtain the target operating parameters;

[0024] Input all the target operating parameters into the pre-constructed second model to predict the corresponding air-conditioning energy consumption;

[0025] Filter out the target operating parameters corresponding to the minimum air-conditioning energy consumption to obtain the optimal operating parameters.

[0026] In a second aspect, an embodiment of the present invention provides a computer room air-conditioning energy-saving control system that executes the computer room air-conditioning energy-saving control method as described above.

[0027] In a third aspect, an embodiment of the present invention provides an electronic device, which includes: a memory and a processor. Among them, the memory and the processor communicate with each other through an internal connection path. The memory is used to store instructions, and the processor is used to execute the instructions stored in the memory. And when the processor executes the instructions stored in the memory, the processor executes the methods in any one of the above aspects.

[0028] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program runs on a computer, the methods in any one of the above aspects are executed.

[0029] The advantages or beneficial effects in the above technical solutions at least include:

[0030] The purpose of the present invention is to utilize computer room data information, based on deep learning technology and computational intelligence methods, mainly including collecting status data such as the indoor temperature of the computer room, the running time of the air conditioner, and the cabinet power, respectively predicting the computer room temperature and the air-conditioning energy consumption, determining the control quantity constraints and state quantity constraints of the air-conditioning system, and calculating the optimal variables in the energy consumption model through computational intelligence methods within a reasonable range, so as to obtain the air-conditioning system control optimization strategy, generate the optimal refrigeration control logic, and dynamically adjust the cooling parameters of the refrigeration equipment to meet the computer room refrigeration demand with the minimum refrigeration power output. The present invention can alleviate the problem of excessive refrigeration in existing computer rooms, reduce power loss, and thus improve the energy-saving level of computer rooms.

[0031] The above summary is only for the purpose of the specification and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of the present invention will be readily apparent by reference to the drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the several views denote the same or similar components or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments disclosed in accordance with the present invention and should not be considered as limiting the scope of the present invention.

[0033] Figure 1 It is a schematic flow chart of the energy-saving control method for the computer room air conditioner of the present invention;

[0034] Figure 2 It is a schematic diagram of the multi-variable time series model of the present invention;

[0035] Figure 3 It is a schematic diagram of the modules of the energy-saving control system for the computer room of the present invention;

[0036] Figure 4 It is a schematic diagram of the algorithm prediction module of the present invention;

[0037] Figure 5 It is a structural block diagram of an electronic device according to an embodiment of the present invention. Detailed implementation manners

[0038] In the following, only some exemplary embodiments are briefly described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the spirit or scope of the present invention. Therefore, the drawings and the description are considered to be exemplary in nature rather than restrictive.

[0039] Embodiment 1

[0040] This embodiment provides an energy-saving control method for a computer room air conditioner, as Figure 1 shown, including:

[0041] Step S1: Obtain the operation data of the computer room equipment, where the computer room equipment includes a cabinet equipment group and a refrigeration equipment group; perform standardization processing on the operation data to obtain standardized data;

[0042] Step S2: Based on a preset first prediction subunit, predict the temperature of the standardized data to obtain the computer room temperature data of the next state, and based on a preset second prediction subunit, predict the energy consumption of the standardized data to obtain the refrigeration energy consumption data of the next state;

[0043] Step S3: Determine the optimal temperature control strategy according to the computer room temperature data and the refrigeration energy consumption data, and generate a corresponding control instruction according to the optimal temperature control strategy. The control instruction is used to control the refrigeration equipment group to operate under the optimal operation parameters so that the cabinet environment temperature does not exceed the preset warning temperature and the total refrigeration power consumption is the lowest.

[0044] In this embodiment, the data acquisition device is used to collect the operation data of air conditioners and cabinet equipment in the computer room. The operation data includes, but is not limited to, the operation data of cabinet equipment groups and refrigeration equipment groups, such as the temperature at the air outlet of the terminal air conditioner, the indoor temperature, the cabinet power, the power of the terminal air conditioner, the operation duration, and the energy consumption meter parameters of each cabinet and each intelligent equipment group. The data acquisition frequency can be per second, per minute, per hour, per day, or other custom time.

[0045] In order to make the values of each data index obtained by collection be at the same order of magnitude after standardization, so that the features in different dimensions are comparable in terms of numerical values, it is necessary to perform standardization processing on the collected historical operation data to improve the accuracy of subsequent deep learning prediction.

[0046] The standardization processing method adopted in this embodiment includes:

[0047] Z-score standardization is adopted, and the data is standardized through the mean of each original data and the standard deviation of each original data in the historical operation data. The processed data conforms to the standard normal distribution, that is, the mean is 0 and the standard deviation is 1. The conversion function is:

[0048]

[0049] where μ is the sample mean and σ is the sample standard deviation; this standardization method is also applicable to the case of outlier data with values outside the range.

[0050] In one implementation manner, a non-linear standardization method is also provided. This method performs non-linear standardization processing on the data, non-linearly scales the values of each data point without changing the data sequence and the overall distribution, expands the data-dense interval while compressing the data-sparse interval. The conversion function is:

[0051]

[0052] The data is mapped to [0, 1] through the Logistic function, where α and β are undetermined coefficients determined by the actual data in the sample. The determination method is as follows:

[0053] ① β = media(x), that is, β is the median of the sample;

[0054] ②

[0055] where t = min(|x 90％ - α|, |x 10％ - β|), that is, the minimum value of the distance from the 90% quantile to the median and the distance from the 10% quantile to the median.

[0056] After standardization processing, standardized data is obtained, and temperature prediction and energy consumption prediction are performed on the standardized data.

[0057] In this embodiment, the first prediction subunit and the second prediction subunit are used to predict the temperature at the next moment and the energy consumption at the next moment.

[0058] The prediction algorithm principles of the first prediction subunit and the second prediction subunit are as follows:

[0059] For a given set of multivariate time series samples: (x1, x2,..., x L ), the lookback window length is L, where each x t is an M-dimensional vector corresponding to time step t, and the next T values are predicted (x L + 1, x L+2 ,..., x L+T ). It is mainly divided into the following parts:

[0060] S1: Forward process

[0061] The i-th sequence in the multivariate time series is represented as:

[0062]

[0063] Therefore, the input (x1, x2,..., x L ) is divided into M single-variable sequences x (i) ∈ R 1*L . According to the Channel-independence setting, each sequence independently enters the Transformer backbone network. The specific structure of the multivariate time series segmentation model is as Figure 2 shown.

[0064] After that, the Transformer will provide the prediction results:

[0065]

[0066] S2: Patching process

[0067] Each input single-variable time series x (i) is first divided into N patches, which can be overlapping or non-overlapping.

[0068] Let the length of each patch be P, and the non-overlapping area between two consecutive patches be denoted as S. Then the generated patch sequence is:

[0069]

[0070] Through the patch operation, the number of tokens in the input model is reduced. Therefore, the patch design allows the model to see longer historical sequences, thus significantly improving the prediction performance.

[0071] S3: Transformer Encoding

[0072] On top of the trainable linear mapping relationship W p = R D*P the patch in step S2 is mapped into the D-dimensional Transformer space, and a learnable positional encoding is used to mark the order of each patch. The input representation finally fed into the Transformer encoder is where

[0073] Multi-head attention output Calculation formula:

[0074]

[0075] Finally, the prediction result is obtained using a flattening layer with a linear head as:

[0076]

[0077] S4: Instance Normalization

[0078] Before the patch, each x (i) is normalized, and then the mean and deviation are added back to the output prediction.

[0079] S5: Loss Function

[0080] The MSE loss is used to measure the difference between the prediction result and the true value. The loss of each channel is calculated and averaged over M time series to obtain the overall objective loss as:

[0081]

[0082] In particular, the present invention also provides a feedback correction mechanism. Because there are many unknown factors, such as model mismatch and environmental interference, which may cause the predicted value to deviate from the actual value. If the real-time information is not used for feedback correction in time, further optimization will be based on falsehood. The correction method adopted in this article is:

[0083]

[0084]

[0085] where h is the compensation coefficient, which can be adjusted according to the actual effect; e(k) is the error between the actual output y(k) of the system at time k and the model predicted output between them.

[0086] In this embodiment, the first prediction subunit predicts the temperature of the next state of the computer room according to the experimental acquisition data set. The main factors affecting the temperature of the computer room are the power of the air conditioner at the current time, the air outlet speed of the air conditioner, and the cabinet temperature. The power of the air conditioner at the current time, the air outlet speed of the air conditioner, and the cabinet temperature are selected as independent variables, the temperature of the computer room at the next moment is used as the dependent variable, and this data is used as the input of the algorithm prediction unit to train the computer room temperature prediction model, so as to predict the temperature of the computer room in the next state.

[0087] The second prediction subunit predicts the energy consumption of the next state of the terminal air conditioner in the computer room. The power of the air conditioner at the current time, the air outlet speed of the air conditioner, and the cabinet load are selected as independent variables, and the energy consumption of the air conditioner at the next moment is used as the dependent variable. This data is used as the input of the algorithm prediction unit to train the air conditioner energy consumption prediction model, so as to predict the energy consumption of the air conditioner in the next state.

[0088] Specifically, in order to generate the prediction result, the data set after standardization processing is received, and the data set is divided into a validation set and a test set by using the cross-validation verification method. In this embodiment, 7600 time steps are selected as the validation set, and 1500 time steps are selected as the test set. During the process of building the prediction model, 96 time steps are used by default.

[0089] Secondly, in order to objectively evaluate each model, the input size is set to twice the time span, that is, 192 time steps, and the maximum number of iterations is set to 50, and other hyperparameters remain the default values.

[0090] Next, by specifying the Transformer model to be used and the prediction frequency, the prediction frequency in this embodiment is to predict once per hour, and the result value predicted by the previous model is updated. Then the cross-validation method is used to utilize the test set and the validation set to return the predicted values and true values of all models.

[0091] Finally, the predicted values and the true values are spliced onto the same time dimension, and the mean squared error (MSE) and the mean absolute error (MAE) are used to evaluate the model performance metrics until the model performance metrics are within the ideal range. In this embodiment, MAE ≤ 0.4 and MSE ≤ 0.2. The MSE and MAE are:

[0092]

[0093]

[0094] where n is the number of samples, y i is the true value, is the predicted value. In this embodiment, the MSE and MAE metrics are used to evaluate in order to eliminate the factors insensitive to outliers and better reflect the distribution of prediction errors.

[0095] The method for determining the optimal temperature control strategy in this embodiment may be:

[0096] Search for all possible values of the air conditioner operation parameters according to the genetic algorithm to obtain the optimal operation parameters;

[0097] Generate the optimal temperature control strategy according to the optimal operation parameters.

[0098] Among them, a global optimization genetic algorithm is used to calculate the optimal temperature control strategy because of its characteristics such as simple algorithm, high generality, strong robustness, and parallel processing ability. Search for the optimal operation parameters among all possible values of the air conditioner operation parameters according to the next-state computer room temperature and air conditioner energy consumption.

[0099] The genetic algorithm is mainly divided into the selection, crossover, and mutation of operators. Among them, the selection operator selects different pairs of individuals in the population according to a certain rule, and the probability of different individuals being selected is proportional to their fitness. The crossover operator, during the crossover process, the gene chains of the two selected individuals are crossed with a certain probability, and then new individuals are generated, and their crossover positions are randomly generated. The mutation operator mutates the gene chains of the new individuals with a certain probability.

[0100] The steps of the genetic algorithm are as follows:

[0101] S1: Initialize the population setting, default k = 0, x(0) = (x1(0), x2(0),..., x n (0)) ∈ S n ;

[0102] S2: Independently select n pairs of parent individuals according to a certain rule in the current population;

[0103] S3: Perform crossover on the n parent individuals to obtain n intermediate individuals, where the crossover processes do not interfere with each other;

[0104] S4: Mutate the n crossed individuals to obtain the next-generation population. The mutation process is the same as the crossover process and is carried out independently, that is:

[0105]

[0106] S5: If the stop criterion is met, stop the iteration; otherwise, k = k + 1, and return to S2 in the genetic algorithm.

[0107] The genetic algorithm of the control strategy module obtains the optimal operating parameters by searching all possible values of the air conditioner operating parameters.

[0108] In one implementation, the method for determining the optimal temperature control strategy can also be:

[0109] Traverse each possible value of the air conditioner operating parameters, input any possible value, the device load, and the meteorological conditions into a pre-constructed first model, and predict the machine room temperature under the corresponding conditions;

[0110] Screen out the possible values for which the machine room temperature is lower than the set threshold to obtain the target operating parameters;

[0111] Input all the target operating parameters into a pre-constructed second model to predict the corresponding air conditioner energy consumption;

[0112] Screen out the target operating parameters corresponding to the minimum air conditioner energy consumption to obtain the optimal operating parameters.

[0113] Specifically, according to the layout, location, and equipment parameters of the data center in practice, select the air outlet temperature, wind speed, and the temperature and humidity at the rack of each air conditioner as independent variables, and take the temperature at the rack at the next moment as the dependent variable. To solve the optimal operating parameters of the air conditioner, the steps for designing the optimal solution are as follows: First, list all possible values of the air conditioner operating parameters, input each value together with the device load and meteorological conditions into the model, and predict the machine room temperature under the corresponding conditions respectively; Second, screen out the possible values that can make the machine room temperature less than a specific threshold; Then, input the selected multiple groups of air conditioner operating parameters into the model respectively to predict the corresponding air conditioner energy consumption; Finally, according to the air conditioner energy consumption value, find the air conditioner operating parameters corresponding to the minimum value as the optimal solution.

[0114] After obtaining the optimal solution, generate the corresponding control instruction, change the operating state of the refrigeration equipment according to the received control instruction, and adjust the power of the refrigeration equipment after parsing the instruction. The operations that the control device can perform include turning off a certain refrigeration equipment; turning on a certain refrigeration equipment; adjusting the operating power of a certain refrigeration equipment to a certain power value.

[0115] The effects that can be obtained by the embodiments of the present invention are as follows:

[0116] 1. The provided energy-saving method based on deep learning prediction not only solves the problem of poor model accuracy caused by the design methods of traditional non-linear modeling of approximate linear or similar linear systems, but also overcomes the defects of local convergence and overfitting brought by other traditional machine learning or deep learning methods.

[0117] 2. The provided Transformer model, thanks to its attention mechanism, can automatically learn the associations between elements in a sequence and has remarkable effects in long-term time series prediction.

[0118] 3. To avoid model mismatch and environmental interference, and special situations that cause model mismatch and environmental interference, the embodiments of the present invention provide a method for feedback correction of real-time information to prevent the control strategy from making wrong control strategies based on wrong data.

[0119] 4. The provided genetic algorithm model is a global optimization algorithm with characteristics such as simplicity, high generality, and strong robustness, and has the ability of parallel processing.

[0120] Embodiment 2

[0121] This embodiment provides an energy-saving control system for computer room air conditioners, and the system executes the energy-saving control method for computer room air conditioners as in Embodiment 1. As Figure 3 shown, the system includes a data acquisition module, an algorithm prediction module, a control strategy module, and a control logic module.

[0122] The data acquisition module is a set of data acquisition devices for collecting the operation data of air conditioners and IT equipment in the computer room. The operation data includes, but is not limited to, the operation data of cabinet equipment groups and refrigeration equipment groups, such as the temperature at the outlet of the terminal air conditioner, the indoor temperature, the cabinet power, the terminal air conditioner power, the operation duration, the energy consumption meters of each cabinet and intelligent equipment group, etc. The acquisition frequency can be per second, per minute, per hour, per day, or other custom time.

[0123] The algorithm prediction module includes a data feature engineering unit and an algorithm prediction unit.

[0124] The data feature engineering unit standardizes the data index values of the data acquisition module so that they are all at the same quantity level, enabling the feature values between different dimensions of the algorithm prediction unit to be comparable in terms of numerical values, which can improve the prediction accuracy of the algorithm prediction unit. Additionally, standardization can also enhance the convergence speed of the algorithm prediction unit model and prevent gradient explosion. Meanwhile, in order to make the variable values of non-poset relations non-poset and equidistant from the origin point, one-hot encoding is used to extend the value range of discrete features to the Euclidean space, and a certain value of a discrete feature corresponds to a certain point in the Euclidean space. Using one-hot encoding for discrete features makes the distance calculation between features more reasonable, such as variables like working days, holidays, and wind directions. One-hot encoding is the representation of binary vectors, using an N-bit status register to encode N states, with each state having its independent register bit and only one bit being valid at any time. Finally, the sample data is divided into three parts: a training sample set, a validation sample set, and a test sample set. Among them, the role of the validation sample set is to evaluate the trained network. To avoid the influence of too small a validation sample set or contingency on parameter adjustment and reliable evaluation of the model, the K-fold cross-validation method is adopted for validation. The specific approach is as follows: divide the data into K partitions; then, instantiate K identical models. For each model, use K - 1 partitions as its training set and the remaining 1 partition as the validation set; take the mean of the K results as the validation result of the algorithm error. In an embodiment of the present invention, K = 4.

[0125] The algorithm prediction unit executes a deep learning method based on the attention mechanism. The algorithm prediction unit includes a first prediction sub-unit and a second prediction sub-unit. As Figure 4 shown, the first prediction sub-unit is a computer room temperature prediction model, and the second prediction sub-unit is an air-conditioning energy consumption prediction model. The prediction models are all based on the cabinet status data and cabinet operation data in the computer room. Among them, the first prediction sub-unit combines the experimentally collected data set to predict the temperature of the next state of the computer room, and the second prediction sub-unit is used to predict the energy consumption of the next state of the end air-conditioning in the computer room.

[0126] This module can continuously learn the data relationship between elements in the cabinet operation status data sequence. Meanwhile, a deep learning method based on the attention mechanism provided by the present invention uses self-supervised representation learning to capture the abstract representation of data, which can further improve the prediction performance of the model.

[0127] The control strategy module executes an optimization method based on computational intelligence. Based on the computer room temperature and air conditioner energy consumption output by the algorithm prediction module, this module is used to calculate the optimal solution of the parameters when the air conditioner operates in the next state. The optimal solution should satisfy that the computer room temperature is lower than the given threshold and the energy consumption of the air conditioner is the lowest. In this control strategy module, its inputs are the cabinet operation status, the output of the algorithm prediction module, and the operation power of the terminal air conditioner. Through parameter traversal and combination, business rule guarantee, the optimal temperature control strategy is found among the optimal strategies, and the output is a control instruction. The optimal control strategy is the best strategy among all control strategies. In specific cabinet states, the operation instructions generated by different control strategies are different; the operation instructions generated by the optimal control strategy are optimal instructions in any state. The optimal instruction means that after the control device executes this instruction in a specific state, the computer room environment temperature is not higher than the warning temperature, and the total cooling power consumption is the lowest.

[0128] The control logic module changes the operation state of the refrigeration equipment according to the received control instruction, and adjusts the power of the refrigeration equipment after parsing the instruction. The operations corresponding to the control instruction include turning off a certain refrigeration equipment, turning on a certain refrigeration equipment, and adjusting the operation power of a certain refrigeration equipment to a certain power value, etc.

[0129] For the functions of the various modules in the system of the embodiments of the present invention, reference can be made to the corresponding descriptions in the above methods, which will not be elaborated here.

[0130] Embodiment III

[0131] Figure 5 Show a structural block diagram of an electronic device according to an embodiment of the present invention. As Figure 5 shown, the electronic device includes: a memory 100 and a processor 200. The memory 100 stores a computer program that can run on the processor 200. When the processor 200 executes the computer program, it implements the computer room air conditioner energy saving control method in the above embodiments. The number of the memory 100 and the processor 200 can be one or more.

[0132] The electronic device further includes:

[0133] A communication interface 300, used for communicating with external devices and performing data interaction and transmission.

[0134] If the memory 100, the processor 200, and the communication interface 300 are implemented independently, the memory 100, the processor 200, and the communication interface 300 can be interconnected through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 only a thick line is used to represent it in Figure 5 , but it does not mean that there is only one bus or one type of bus.

[0135] Optionally, in a specific implementation, if the memory 100, the processor 200, and the communication interface 300 are integrated on a single chip, the memory 100, the processor 200, and the communication interface 300 can communicate with each other through an internal interface.

[0136] An embodiment of the present invention provides a computer-readable storage medium that stores a computer program, and when the program is executed by a processor, the method provided in the embodiment of the present invention is implemented.

[0137] An embodiment of the present invention further provides a chip, which includes a processor for calling and running instructions stored in a memory, so that a communication device installed with the chip executes the method provided in the embodiment of the present invention.

[0138] An embodiment of the present invention further provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, the output interface, the processor, and the memory are connected through an internal connection path. The processor is used to execute code in the memory, and when the code is executed, the processor is used to execute the method provided in the embodiment of the invention.

[0139] It should be understood that the above-mentioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. It is worth noting that the processor can be a processor that supports the advanced RISC machines (ARM) architecture.

[0140] Further, optionally, the above-mentioned memory can include a read-only memory and a random access memory, and can also include a non-volatile random access memory. The memory can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can include a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can include a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available. For example, static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM).

[0141] 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 a computer, the processes or functions according to the present invention are 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 from one computer-readable storage medium to another computer-readable storage medium.

[0142] In the description of this specification, the descriptions with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0143] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" can explicitly or implicitly include at least one of these features. In the description of the present invention, the meaning of "a plurality" is two or more, unless otherwise specifically defined.

[0144] Any process or method description shown in the flowchart or described in other ways herein can be understood to represent a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process. And the scope of the preferred embodiments of the present invention includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the involved functions, rather than in the order shown or discussed.

[0145] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as an ordered list of executable instructions for implementing a logical function and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatus, or devices.

[0146] It should be understood that each part of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the method in the above embodiments can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0147] In addition, each functional unit in various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. If the above integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. The storage medium can be a read-only memory, a disk, an optical disc, or the like.

[0148] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various changes or substitutions, and these should all be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An energy-saving control method for a computer room air conditioner, characterized in that Including: Obtain the operation data of the computer room equipment, where the computer room equipment includes a cabinet equipment group and a refrigeration equipment group; Perform standardization processing on the operation data to obtain standardized data; among them, the method of the standardization processing includes: performing data standardization processing on the mean value of each original data and the standard deviation of each original data in the operation data by using Z-score standardization; or, performing non-linear scaling on the value of each original data in the operation data; Use the air conditioner power at the current time, the air outlet wind speed of the air conditioner at the current time, and the cabinet temperature at the current time as independent variables, use the computer room temperature at the next moment as the dependent variable, and input the independent variables and the dependent variable into the trained first prediction subunit; based on the first prediction subunit, perform temperature prediction on the standardized data to obtain the computer room temperature data in the next state; Use the air conditioner power at the current time, the air outlet wind speed of the air conditioner at the current time, and the cabinet load at the current time as independent variables, use the refrigeration energy consumption at the next moment as the dependent variable, and input the independent variables and the dependent variable into the trained second prediction subunit; based on the second prediction subunit, perform energy consumption prediction on the standardized data to obtain the refrigeration energy consumption data in the next state; Determine the optimal temperature control strategy according to the computer room temperature data and the refrigeration energy consumption data. The method for determining the optimal temperature control strategy is: traverse each possible value of the refrigeration equipment operation parameters, where the refrigeration equipment operation parameters include air conditioner power and air outlet wind speed of the air conditioner; input any one of the possible values and the cabinet temperature into the pre-constructed first prediction subunit to predict the computer room temperature under the corresponding conditions; screen out the possible values for which the computer room temperature is lower than the set threshold to obtain the target operation parameters; input all the target operation parameters into the pre-constructed second prediction subunit to predict the corresponding refrigeration energy consumption; screen out the target operation parameters corresponding to the minimum refrigeration energy consumption to obtain the optimal operation parameters; Generate a corresponding control instruction according to the optimal temperature control strategy, and the control instruction is used to control the refrigeration equipment group to operate under the optimal operation parameters.

2. The energy-saving control method for computer room air conditioners according to claim 1, characterized in that, The operation data includes the cabinet equipment temperature, the air outlet temperature of the refrigeration equipment, the indoor temperature, the cabinet equipment power, and the refrigeration equipment power.

3. The energy-saving control method for computer room air conditioners according to claim 1, characterized in that, The method for determining the optimal temperature control strategy is: Search for all possible values of the refrigeration equipment operation parameters according to the genetic algorithm to obtain the optimal operation parameters; Generate the optimal temperature control strategy according to the optimal operation parameters.

4. A computer room air conditioning energy saving control system, characterized in that, The system includes computer room equipment, and the computer room equipment includes a cabinet equipment group and a refrigeration equipment group; the computer room equipment executes the computer room air-conditioning energy-saving control method as described in any one of claims 1 to 3.

5. An electronic device, characterized in that, Including: A processor and a memory, where instructions are stored in the memory, and the instructions are loaded and executed by the processor to implement the computer room air-conditioning energy-saving control method as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, A computer program is stored in the computer-readable storage medium, and when the computer program is executed by the processor, it implements the computer room air-conditioning energy-saving control method as described in any one of claims 1 to 3.

Images