Methods, devices, equipment and storage media for adjusting operating parameters of computer room control equipment

Abstract

This disclosure relates to a method, apparatus, device, and storage medium for adjusting the operating parameters of a data center control device. The method includes: training a relationship model between the control variables and operating parameters of the data center control device using known control variables and operating parameters; calculating key operating parameters corresponding to each control variable based on the trained relationship model, wherein the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters; and adjusting the key operating parameters corresponding to the specified control variables of the data center control device when a request to change a specified control variable is received, so that the specified control variables of the data center control device change, and the ideal control variables can be obtained by adjusting the key operating parameters to ensure the optimal operating state of the data center control device.

Description

Technical Field

[0001] This disclosure relates to the field of computer room control technology, and in particular to a method, device, equipment and storage medium for adjusting the operating parameters of computer room control equipment. Background Technology

[0002] Currently, in computer rooms, equipment such as chillers, water pumps, and cooling towers, despite their highly coupled operating points, require decentralized control due to their individual requirements. This presents obstacles to the collaborative processing of equipment data. Numerous input parameters are generated during equipment operation, and even subtle changes in each parameter can affect the equipment's performance. Operating points consist of various operating parameters recorded during equipment operation, including both discrete and continuous data, such as the following equipment operating data:

[0003] Discrete data:

[0004] Switch status: including water pump switch status, fan switch status, etc., represented by numbers 0 and 1.

[0005] Continuous data:

[0006] Temperature: including chilled water inlet temperature, chilled water outlet temperature, cooling water inlet temperature, cooling water outlet temperature, etc.;

[0007] Flow rate: including chilled water flow rate, cooling water flow rate, etc.;

[0008] Rotation speed: including cooling water pump speed, chilled water pump speed, and cooling tower fan speed;

[0009] Environment: including indoor and outdoor ambient temperature, weather conditions, etc.;

[0010] Power consumption: including total system operating power and total system operating efficiency.

[0011] The above data together constitute multi-dimensional operating condition data at a specific moment. Under specific operating conditions, the current data center system maintains a certain operating mode and generates corresponding output. Due to the large number of operating parameters involved and the high dimensionality of the operating condition data, it is not possible to intuitively determine the impact of changes in specific parameters on the output of that state. Summary of the Invention

[0012] In order to solve the above-mentioned technical problems, or at least partially solve the above-mentioned technical problems, embodiments of this disclosure provide a method, apparatus, equipment and storage medium for adjusting the operating parameters of computer room control equipment.

[0013] In a first aspect, embodiments of this disclosure provide a method for adjusting the operating parameters of a computer room control device, the method comprising:

[0014] A model of the relationship between the control variables and operating parameters of the computer room control equipment is trained using known control variables and operating parameters of the computer room control equipment.

[0015] Based on the relationship model between the control variables and operating parameters of the pre-trained computer room control equipment, the key operating parameters corresponding to each control variable are calculated. Among them, the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters.

[0016] When a request to change the specified control variable of the computer room control equipment is received, the key operating parameters corresponding to the specified control variable of the computer room control equipment are adjusted so that the specified control variable of the computer room control equipment changes.

[0017] In one possible implementation, the relationship model is a neural network model, which is trained using known control variables and operating parameters of the computer room control equipment to obtain a relationship model between the control variables and operating parameters of the computer room control equipment, including:

[0018] The known control variables of the computer room control equipment are input into the pre-set calculation modules corresponding to each control variable. The output values ​​of the calculation modules are input into the original neural network model. The parameters of the original neural network model are trained by using the difference between the output value of the original neural network model and the operating parameters corresponding to the control variables. When the difference between the output value of the trained neural network model and the operating parameters corresponding to the control variables meets the first preset threshold, the trained neural network model is used as the relationship model between the control variables and the operating parameters of the computer room control equipment.

[0019] In one possible implementation, the calculation module includes a linear module, a differential module, and an integral module. Known control variables of the computer room control equipment are input into pre-set calculation modules corresponding to each control variable, including:

[0020] Determine whether the relationship between each control variable and operating parameter of the known computer room control equipment is linear, differential, or integral.

[0021] When the known correspondence between the control variables and operating parameters of the computer room control equipment is linear, the known control variables of the computer room control equipment are input into the linear module.

[0022] When the known relationship between the operating parameters and control variables of the computer room control equipment is an integral relationship, the known control variables of the computer room control equipment are input into the differential module;

[0023] When the known operating parameters of the computer room control equipment and the control variables have a differential relationship, the known control variables of the computer room control equipment are input into the integral module.

[0024] In one possible implementation, the key operating parameters corresponding to each control variable are calculated based on the relationship model between the control variables and operating parameters of the pre-trained computer room control equipment, including:

[0025] Select one control variable as the target control variable and set it to the first preset value, and set all other control variables other than the target control variable to 0;

[0026] The first preset value of the target control variable and the zero values ​​of other control variables are input into the pre-set calculation modules corresponding to each control variable. The output value of the calculation modules is input into the relationship model between the control variables and operating parameters of the trained computer room control equipment to obtain the first predicted value of each operating parameter.

[0027] While keeping all control variables other than the target control variable at 0, the target control variable is set to the second preset value to obtain the second predicted value of each operating parameter;

[0028] By comparing the difference between the first and second predicted values ​​of each operating parameter, the operating parameter corresponding to the largest difference is taken as the key operating parameter of the target control variable.

[0029] In one possible implementation, the relational model is a variational autoencoder model, which is trained using known control variables and operating parameters of the computer room control equipment to obtain a relationship model between the control variables and operating parameters, including:

[0030] The known operating parameters of the computer room control equipment are input into the encoder of the variational autoencoder model to obtain a low-dimensional representation of the operating parameters of the computer room control equipment. This low-dimensional representation of the operating parameters of the computer room control equipment is then used as the latent variable space.

[0031] The mean and variance of the operating parameters of the computer room control equipment are represented in a low-dimensional form, and the vector Z = {Z1, Z2, ..., Zn} is obtained by sampling from the latent variable space. n};

[0032] The sampled vector Z = {Z1, Z2, ..., Z} is obtained by... n The decoder of the input variational autoencoder model generates predictive control variables corresponding to the operating parameters input to the encoder.

[0033] Based on the difference between the predicted control variable and the actual control variable corresponding to the operating parameters input to the encoder, a variational autoencoder model is trained by adjusting the parameters of the encoder and decoder until the difference between the predicted control variable and the actual control variable corresponding to the operating parameters input to the encoder is less than a second preset threshold. The trained variational autoencoder model is then used as the relationship model between the control variables and operating parameters of the computer room control equipment.

[0034] In one possible implementation, the key operating parameters corresponding to each control variable are calculated based on the relationship model between the control variables and operating parameters of the pre-trained computer room control equipment, including:

[0035] Select one operating parameter of the computer room control equipment as the target operating parameter and set it as the third preset value; set all other operating parameters other than the target operating parameter to 0.

[0036] The third preset value of the target running parameter and the zero values ​​of other running parameters are input into the encoder of the trained variational autoencoder model to obtain the gradient response value of the target running parameter to the data of each channel in the latent space.

[0037] By comparing the gradient response values ​​of the target operating parameters to the data of each channel in the latent space with the preset third threshold, the channels in the latent space whose gradient response values ​​exceed the third preset threshold are obtained as the channels in the latent space corresponding to the target operating parameters, thereby obtaining the channels in the latent space corresponding to each operating parameter of the computer room control equipment.

[0038] Randomly sample the operating parameters of the computer room control equipment and input them into the trained variational autoencoder model to obtain the predictive control variables corresponding to the operating parameters of the computer room control equipment. Then, perform backpropagation to solve the gradient response values ​​of the predictive control variables to the data of each channel in the latent space.

[0039] By comparing the gradient response values ​​of the predictive control variables to the data of each channel in the latent space with the preset fourth threshold, the channels in the latent space whose gradient response values ​​exceed the fourth preset threshold are obtained as the channels in the latent space corresponding to the predictive control variables, thereby obtaining the channels in the latent space corresponding to each control variable of the computer room control equipment.

[0040] Each control variable is taken as the current control variable. In the implicit space channel corresponding to each operating parameter of the computer room control equipment, the same operating parameter as the channel of the implicit space corresponding to the current control variable is selected as the key operating parameter corresponding to the current control variable.

[0041] In one possible implementation, the gradient response value is obtained by differentiating the loss function of the variational autoencoder model with respect to the data of each channel in the latent space.

[0042] Secondly, embodiments of this disclosure provide a device for adjusting the operating parameters of a computer room control equipment, comprising:

[0043] The training module is used to train a model of the relationship between the control variables and operating parameters of the computer room control equipment using known control variables and operating parameters of the computer room control equipment.

[0044] The calculation module is used to calculate the key operating parameters corresponding to each control variable based on the relationship model between the control variables and operating parameters of the computer room control equipment, wherein the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters.

[0045] The adjustment module is used to adjust the key operating parameters corresponding to the specified control variables of the computer room control equipment when a request to change the specified control variables of the computer room control equipment is received, so that the specified control variables of the computer room control equipment change.

[0046] Thirdly, embodiments of this disclosure provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other via the communication bus;

[0047] Memory, used to store computer programs;

[0048] The processor, when executing the program stored in the memory, implements the above-mentioned method for adjusting the operating parameters of the computer room control equipment.

[0049] Fourthly, embodiments of this disclosure provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for adjusting the operating parameters of the computer room control equipment.

[0050] Compared with the prior art, the technical solutions provided in this disclosure have at least some or all of the following advantages:

[0051] The method for adjusting the operating parameters of the data center control equipment described in this embodiment utilizes known control variables and operating parameters of the data center control equipment to train a relationship model between the control variables and operating parameters. Based on the trained relationship model, the key operating parameters corresponding to each control variable are calculated, wherein the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters. When a request to change a specified control variable of the data center control equipment is received, the key operating parameters corresponding to the specified control variable are adjusted, so that the specified control variable of the data center control equipment changes. By adjusting the key operating parameters, the ideal control variable can be obtained to ensure the optimal operating state of the data center control equipment. Attached Figure Description

[0052] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0053] To more clearly illustrate the technical solutions in the embodiments of this disclosure or the prior art, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0054] Figure 1 The schematic diagram illustrates a flowchart of a method for adjusting the operating parameters of a computer room control device according to an embodiment of the present disclosure;

[0055] Figure 2 The illustration schematically shows a representation of the input and output values ​​of a neural network being trained in the case where the relational model is a neural network model, according to an embodiment of the present disclosure.

[0056] Figure 3 This schematically illustrates a structural block diagram of a computer room control equipment operating parameter adjustment device according to an embodiment of the present disclosure; and

[0057] Figure 4 A schematic block diagram of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0059] See Figure 1 The present disclosure provides a method for adjusting the operating parameters of a computer room control device, the method comprising:

[0060] S1. Using the known control variables and operating parameters of the computer room control equipment, a relationship model between the control variables and operating parameters of the computer room control equipment is trained.

[0061] S2, based on the relationship model between the control variables and operating parameters of the trained computer room control equipment, calculate the key operating parameters corresponding to each control variable, wherein the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters;

[0062] S3, when a request to change the specified control variable of the computer room control equipment is received, the key operating parameters corresponding to the specified control variable of the computer room control equipment are adjusted so that the specified control variable of the computer room control equipment changes.

[0063] In this embodiment, in step S1, the relationship model is a neural network model, which is trained using known control variables and operating parameters of the computer room control equipment to obtain a relationship model between the control variables and operating parameters of the computer room control equipment, including:

[0064] The known control variables of the computer room control equipment are input into pre-set calculation modules corresponding to each control variable. The output values ​​of the calculation modules are input into the original neural network model. The parameters of the original neural network model are trained using the difference between the output value of the original neural network model and the operating parameters corresponding to the control variables. This process continues until the difference between the output value of the trained neural network model and the operating parameters corresponding to the control variables meets a first preset threshold. The trained neural network model is then used as the relationship model between the control variables and the operating parameters of the computer room control equipment. The calculation modules include a linear module, a differential module, and an integral module. The process of inputting the known control variables of the computer room control equipment into the pre-set calculation modules corresponding to each control variable includes:

[0065] Determine whether the relationship between each control variable and operating parameter of the known computer room control equipment is linear, differential, or integral.

[0066] When the known correspondence between the control variables and operating parameters of the computer room control equipment is linear, the known control variables of the computer room control equipment are input into the linear module.

[0067] When the known relationship between the operating parameters and control variables of the computer room control equipment is an integral relationship, the known control variables of the computer room control equipment are input into the differential module;

[0068] When the known operating parameters of the computer room control equipment and the control variables have a differential relationship, the known control variables of the computer room control equipment are input into the integral module.

[0069] In this embodiment, in step S2, when the relationship model is a neural network model, the key operating parameters corresponding to each control variable are calculated based on the trained relationship model between the control variables and operating parameters of the computer room control equipment, including:

[0070] Select one control variable as the target control variable and set it to the first preset value, and set all other control variables other than the target control variable to 0;

[0071] The first preset value of the target control variable and the zero values ​​of other control variables are input into the pre-set calculation modules corresponding to each control variable. The output value of the calculation modules is input into the relationship model between the control variables and operating parameters of the trained computer room control equipment to obtain the first predicted value of each operating parameter.

[0072] While keeping all control variables other than the target control variable at 0, the target control variable is set to the second preset value to obtain the second predicted value of each operating parameter;

[0073] By comparing the difference between the first and second predicted values ​​of each operating parameter, the operating parameter corresponding to the largest difference is taken as the key operating parameter of the target control variable.

[0074] Taking the relational model as a neural network model as an example, the method for adjusting the operating parameters of the computer room control equipment in this embodiment is explained as follows:

[0075] The computer room control equipment system is a multi-input, multi-output model. The inputs are operating point data, i.e., parameters of the computer room control equipment, such as inlet water temperature and pump speed. The outputs are control variables within the computer room, such as power consumption and the controlled temperature. This system is reversible; therefore, a neural network can be constructed to fit and analyze the system outputs.

[0076] First, a preliminary analysis of the computer room control equipment system is conducted to identify the main influencing factors. Based on the number of main factors, a neural network is constructed by grouping them. (See also...) Figure 2 Based on the analysis, if a factor's influence on the system is linear, a linear module can be added before the neural network. For example, if the system output is the cooling water outlet temperature, this output is affected by the cooling water inlet temperature, and this influence can be fitted linearly. If a factor's influence on the system is integral, a differential module can be added before the neural network. For example, if the system output is the total flow rate per unit time, this output is affected by the pump speed, and this influence accumulates over time, it can be fitted integrally, thus a differential module can be added. Conversely, if a factor's influence on the system is differential, an integral module can be added to the inverse system. For example, if the system output is the instantaneous power consumption of a fan, this output is affected by the corresponding fan on / off state and speed, and this influence is differential, thus an integral module can be added. The corresponding calculation module is selected based on the different output types of the original system. When the output corresponds to multiple properties, the system output can be connected to multiple calculation modules. Adding a neural network after the calculation module constitutes the neural network inverse system. The input of this inverse system is the output of the original system, and the output is the input of the original system. Different types of factors are equivalent to different subsystems to achieve decoupling of the system's input and output. The constructed neural network includes an input layer, multiple hidden layers, and an output layer, with specific parameters set according to the training process. Multiple sets of input and output values ​​are collected based on the system's operational status. By training a specific module and the inverse system of the neural network, the inverse relationship between output and input values ​​is obtained, thus creating a network between each output and input value, achieving inverse fitting of the computer room system. Based on this inverse system, a specific output value can be given while other output values ​​are set to zero, allowing for inverse reasoning of the input value. For example, setting the original system output y1 to 1 unit and setting the remaining outputs to 0. While keeping all other outputs at 0, changing the value of output y1 and observing the change in the actual input of the original system (i.e., the output of the inverse system) reveals the input with the smallest change, which most directly affects the output value y1. This demonstrates the relationship between the actual input of the original system and each output value, thus achieving decoupling.

[0077] In this embodiment, in step S1, the relational model is a variational autoencoder model, which is trained using known control variables and operating parameters of the computer room control equipment to obtain a relational model between the control variables and operating parameters, including:

[0078] The known operating parameters of the computer room control equipment are input into the encoder of the variational autoencoder model to obtain a low-dimensional representation of the operating parameters of the computer room control equipment. This low-dimensional representation of the operating parameters of the computer room control equipment is then used as the latent variable space.

[0079] The mean and variance of the operating parameters of the computer room control equipment are represented in a low-dimensional form, and the vector Z = {Z1, Z2, ..., Zn} is obtained by sampling from the latent variable space. n};

[0080] The sampled vector Z = {Z1, Z2, ..., Z} is obtained by... n The decoder of the input variational autoencoder model generates predictive control variables corresponding to the operating parameters input to the encoder.

[0081] Based on the difference between the predicted control variable and the actual control variable corresponding to the operating parameters input to the encoder, a variational autoencoder model is trained by adjusting the parameters of the encoder and decoder until the difference between the predicted control variable and the actual control variable corresponding to the operating parameters input to the encoder is less than a second preset threshold. The trained variational autoencoder model is then used as the relationship model between the control variables and operating parameters of the computer room control equipment.

[0082] The disclosed method for adjusting the operating parameters of computer room control equipment uses deep learning methods such as neural network inverse system or decoupled representation learning to model the key factors affecting the control variables, analyze the impact of changes in a key factor on the control variables, analyze the meaning behind the characteristics of each operating parameter, further optimize the logic behind the decision, and thus better regulate the equipment operation in order to achieve the optimal operating state of the equipment.

[0083] In this embodiment, in step S2, when the relational model is a variational autoencoder model, the key operating parameters corresponding to each control variable are calculated based on the relationship model between the control variables and operating parameters of the trained computer room control equipment, including:

[0084] Select one operating parameter of the computer room control equipment as the target operating parameter and set it as the third preset value; set all other operating parameters other than the target operating parameter to 0.

[0085] The third preset value of the target running parameter and the zero values ​​of other running parameters are input into the encoder of the trained variational autoencoder model to obtain the gradient response value of the target running parameter to the data of each channel in the latent space. The gradient response value is the value obtained by taking the derivative of the loss function of the variational autoencoder model with respect to the data of each channel in the latent space.

[0086] By comparing the gradient response values ​​of the target operating parameters to the data of each channel in the latent space with the preset third threshold, the channels in the latent space whose gradient response values ​​exceed the third preset threshold are obtained as the channels in the latent space corresponding to the target operating parameters, thereby obtaining the channels in the latent space corresponding to each operating parameter of the computer room control equipment.

[0087] Randomly sample the operating parameters of the computer room control equipment and input them into the trained variational autoencoder model to obtain the predictive control variables corresponding to the operating parameters of the computer room control equipment. Then, perform backpropagation to solve the gradient response values ​​of the predictive control variables to the data of each channel in the latent space.

[0088] By comparing the gradient response values ​​of the predictive control variables to the data of each channel in the latent space with the preset fourth threshold, the channels in the latent space whose gradient response values ​​exceed the fourth preset threshold are obtained as the channels in the latent space corresponding to the predictive control variables, thereby obtaining the channels in the latent space corresponding to each control variable of the computer room control equipment.

[0089] Each control variable is taken as the current control variable. In the implicit space channel corresponding to each operating parameter of the computer room control equipment, the same operating parameter as the channel of the implicit space corresponding to the current control variable is selected as the key operating parameter corresponding to the current control variable.

[0090] Taking the relational model as a variational autoencoder model as an example, the method for adjusting the operating parameters of the computer room control equipment in this embodiment is further explained:

[0091] First, the encoder and decoder of the variational autoencoder model are trained: the system input X (operating point data) of the computer room control equipment is passed sequentially through the encoder and decoder. The encoder encodes the input data into layers of lower-dimensional representation. By calculating the low-dimensional mean μ and variance λ of each input data point, samples are taken from the latent variable space of the variational autoencoder model to obtain Z = {Z1, Z2, ..., Zn}. n The sampled data Z from the latent variable space follows the probability distribution of the original data X, thus the new data Y generated from the sampled data Z is also influenced by the probability distribution of the original data. The decoder generates new data Y = {Y1, Y2, ..., Y} by reconstructing the encoded representation. n The encoder and decoder are trained by continuously comparing the new data Y generated by the decoder with the actual system output Y0 of the computer room control equipment system.

[0092] Then, the operating parameters that have a significant impact on the control variables, i.e., the system inputs that have a significant impact on the system output, are calculated. Through training, the encoder encodes various attributes in the latent space, such as encoding factors affecting ambient temperature and storing them as probability distributions of the attributes. Z in the latent space is the most efficient representation of the input X of the computer room control equipment system. The data in each dimension of Z contains specific data features, i.e., an effective decoupled representation of the various interpretable generative factors within the data. Each component of the input X of the computer room control equipment system is input into the encoder to obtain the gradient response values ​​of each dimension (i.e., each channel) in the latent space. At the same time, a threshold is set, and the channels with larger response values ​​for that input are obtained based on the threshold, thus obtaining a certain channel range. Simultaneously, X is randomly sampled... ′The input system obtains the output Y′ and performs backpropagation on each channel of the latent space to obtain the gradient response value of the output to each channel. Based on a threshold, the channel with the larger response value for the output is selected. By comparing the two channels with larger response values, the input and output can be effectively adjusted for the channel with both large input and output response values, that is, there is a large correlation between the corresponding input and output, thereby achieving decoupling.

[0093] The method for adjusting the operating parameters of the computer room control equipment disclosed herein captures and represents each interpretable generative factor within the data using a subset of latent variables that are as independent as possible through a variational autoencoder model, and models the representation of real data from the perspective of maximum likelihood.

[0094] See Figure 3 The present disclosure provides an embodiment of a device for adjusting the operating parameters of a computer room control equipment, comprising:

[0095] Training module 11 is used to train a model of the relationship between the control variables and operating parameters of the computer room control equipment using known control variables and operating parameters of the computer room control equipment;

[0096] The calculation module 12 is used to calculate the key operating parameters corresponding to each control variable based on the relationship model between the control variables and operating parameters of the trained computer room control equipment. The key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters.

[0097] The adjustment module 13 is used to adjust the key operating parameters corresponding to the specified control variable of the computer room control device when a request to change the specified control variable of the computer room control device is received, so that the specified control variable of the computer room control device changes.

[0098] The specific implementation process of the functions and roles of each unit in the above device can be found in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0099] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can be referred to in the description of the method embodiments. The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the present invention according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0100] In the above embodiments, any plurality of the training module 11, calculation module 12, and adjustment module 13 can be combined into one module, or any one of the modules can be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules can be combined with at least part of the functionality of other modules and implemented in one module. At least one of the training module 11, calculation module 12, and adjustment module 13 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any appropriate combination of any of these three implementation methods. Alternatively, at least one of the training module 11, calculation module 12, and adjustment module 13 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.

[0101] Reference Figure 4 As shown, the electronic device provided in the embodiments of this disclosure includes a processor 1110, a communication interface 1120, a memory 1130, and a communication bus 1140, wherein the processor 1110, the communication interface 1120, and the memory 1130 communicate with each other through the communication bus 1140.

[0102] Memory 1130 is used to store computer programs;

[0103] When processor 1110 executes the program stored in memory 1130, it implements the following method for adjusting the operating parameters of the computer room control equipment:

[0104] A model of the relationship between the control variables and operating parameters of the computer room control equipment is trained using known control variables and operating parameters of the computer room control equipment.

[0105] Based on the relationship model between the control variables and operating parameters of the pre-trained computer room control equipment, the key operating parameters corresponding to each control variable are calculated. Among them, the key operating parameters corresponding to the control variables are the operating parameters that have the greatest impact on the control variables among all operating parameters.

[0106] When a request to change the specified control variable of the computer room control equipment is received, the key operating parameters corresponding to the specified control variable of the computer room control equipment are adjusted so that the specified control variable of the computer room control equipment changes.

[0107] The aforementioned communication bus 1140 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus 1140 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, it is represented by only one thick line in the figure, but this does not indicate that there is only one bus or one type of bus.

[0108] The communication interface 1120 is used for communication between the above-mentioned electronic device and other devices.

[0109] The memory 1130 may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory 1130 may also be at least one storage device located remotely from the aforementioned processor 1110.

[0110] The processor 1110 mentioned above can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0111] Embodiments of this disclosure also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for adjusting the operating parameters of the computer room control equipment as described above.

[0112] The computer-readable storage medium may be included in the device / apparatus described in the above embodiments; or it may exist independently and not assembled into the device / apparatus. The computer-readable storage medium carries one or more programs, which, when executed, implement the method for adjusting the operating parameters of the computer room control equipment according to the embodiments of this disclosure.

[0113] According to embodiments of this disclosure, the computer-readable storage medium can be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0114] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0115] The above description is merely a specific embodiment of this disclosure, enabling those skilled in the art to understand or implement it. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this disclosure. Therefore, this disclosure is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method of adjusting an operation parameter of a machine room control device, characterized by, The method comprises: training a relationship model between control variables and operating parameters of the machine room control equipment by using known control variables and operating parameters of the machine room control equipment; based on the trained relationship model between the control variables and the operating parameters of the machine room control equipment, calculating the key operating parameter corresponding to each control variable, wherein the key operating parameter corresponding to the control variable is the operating parameter that has the greatest impact on the control variable among all operating parameters; when a request to change a specified control variable of the machine room control equipment is received, adjusting the key operating parameter corresponding to the specified control variable of the machine room control equipment so that the specified control variable of the machine room control equipment changes; in response to the relationship model being a neural network model, the calculation of the key operating parameter corresponding to each control variable based on the trained relationship model between the control variables and the operating parameters of the machine room control equipment comprises: selecting a control variable as a target control variable and setting it to a first preset value, and setting other control variables except the target control variable to 0; inputting the first preset value of the target control variable and the 0 value of the other control variables into the pre-set calculation module corresponding to each control variable, and inputting the output value of the calculation module into the trained relationship model between the control variables and the operating parameters of the machine room control equipment to obtain the first predicted value of each operating parameter; setting the target control variable to a second preset value while keeping other control variables except the target control variable as 0 to obtain the second predicted value of each operating parameter; comparing the difference between the first predicted value and the second predicted value of each operating parameter, and taking the operating parameter corresponding to the maximum difference as the key operating parameter of the target control variable.

2. The method of claim 1, wherein, The relationship model is a neural network model, and the relationship model between the control variables and the operating parameters of the machine room control equipment is trained by using known control variables and operating parameters of the machine room control equipment, comprising: inputting the known control variables of the machine room control equipment into the pre-set calculation module corresponding to each control variable, inputting the output value of the calculation module into the original neural network model, training the parameters of the original neural network model by using the difference between the output value of the original neural network model and the operating parameter corresponding to the control variable, until the difference between the output value of the trained neural network model and the operating parameter corresponding to the control variable satisfies a first preset threshold, and taking the trained neural network model as the relationship model between the control variables and the operating parameters of the machine room control equipment.

3. The method of claim 2, wherein, The calculation module comprises a linear module, a differential module and an integral module, and the known control variables of the machine room control equipment are input into the pre-set calculation module corresponding to each control variable, comprising: determining whether the corresponding relationship between each control variable and the operating parameter of the known machine room control equipment is a linear relationship, a differential relationship or an integral relationship: when the corresponding relationship between the control variable and the operating parameter of the known machine room control equipment is a linear relationship, inputting the control variable of the known machine room control equipment into the linear module; when the correspondence between the operating parameter of the known machine room control device and the control variable is an integral relationship, inputting the control variable of the known machine room control device into the differential module; when the correspondence between the operating parameter of the known machine room control device and the control variable is a differential relationship, inputting the control variable of the known machine room control device into the integral module.

4. The method of claim 1, wherein, The relationship model is a variational auto-encoding model, and the relationship model between the control variable and the operating parameter is trained by using the control variable and the operating parameter of the known machine room control device, including: inputting the operating parameter of the known machine room control device into the encoder of the variational auto-encoding model to obtain a low-dimensional representation of the operating parameter of the machine room control device, and taking the low-dimensional representation of the operating parameter of the machine room control device as a latent variable space; The computer room control device calculates the mean and variance of the low-dimensional representation of the operating parameters and samples from the latent variable space according to the mean and variance to obtain a vector Z = {Z1, Z2,..., ZN}. n} The sampled vector Z = {Z1, Z2,..., Z n} is input into the decoder of the variational self-encoding model to generate a predicted control variable corresponding to the operating parameters input into the encoder. training the variational auto-encoding model by adjusting the parameters of the encoder and the decoder according to the difference between the predicted control variable and the actual control variable corresponding to the operating parameter input into the encoder, until the difference between the predicted control variable and the actual control variable corresponding to the operating parameter input into the encoder is less than a second preset threshold, and taking the trained variational auto-encoding model as the relationship model between the control variable and the operating parameter of the machine room control device.

5. The method of claim 4, wherein, The trained relationship model between the control variable and the operating parameter of the machine room control device is used to calculate the key operating parameter corresponding to each control variable, including: selecting one operating parameter of the machine room control device as a target operating parameter and setting it to a third preset value, and setting other operating parameters to 0; inputting the third preset value of the target operating parameter and the 0 value of the other operating parameters into the encoder of the trained variational auto-encoding model to obtain the gradient response value of the target operating parameter to the data of each channel of the latent space; comparing the gradient response value of the target operating parameter to the data of each channel of the latent space with a third preset threshold to obtain the channel of the latent space whose gradient response value exceeds the third preset threshold as the channel of the latent space corresponding to the target operating parameter, thereby obtaining the channel of the latent space corresponding to each operating parameter of the machine room control device; randomly sampling the operating parameter of the machine room control device and inputting it into the trained variational auto-encoding model to obtain the predicted control variable corresponding to the operating parameter of the machine room control device, and performing back propagation to solve the gradient response value of the predicted control variable to the data of each channel of the latent space; comparing the gradient response value of the predicted control variable to the data of each channel of the latent space with a fourth preset threshold to obtain the channel of the latent space whose gradient response value exceeds the fourth preset threshold as the channel of the latent space corresponding to the predicted control variable, thereby obtaining the channel of the latent space corresponding to each control variable of the machine room control device; taking each control variable as a current control variable, selecting the operating parameter corresponding to the channel of the latent space corresponding to the current control variable as the key operating parameter corresponding to the current control variable in the channel of the latent space corresponding to each operating parameter of the machine room control device.

6. The method of claim 5, wherein, The gradient response value is a value obtained by deriving the loss function of the variational auto-encoding model with respect to the data of each channel of the latent space.

7. A device for adjusting operating parameters of computer room control equipment, characterized in that, The method comprises the following steps: a training module is used to train a relationship model between control variables and operating parameters of the machine room control device by using known control variables and operating parameters of the machine room control device; a calculation module is used to calculate a key operating parameter corresponding to each control variable based on the trained relationship model between the control variables and the operating parameters of the machine room control device, wherein the key operating parameter corresponding to the control variable is an operating parameter that has the greatest impact on the control variable among all operating parameters; an adjustment module is used to adjust the key operating parameter corresponding to a specified control variable of the machine room control device when a request to change the specified control variable of the machine room control device is received, so that the specified control variable of the machine room control device is changed; in response to the relationship model being a neural network model, the calculation of the key operating parameter corresponding to each control variable based on the trained relationship model between the control variables and the operating parameters of the machine room control device comprises the following steps: a control variable is selected as a target control variable and is set to a first preset value, and other control variables except the target control variable are set to 0; the first preset value of the target control variable and the 0 values of the other control variables are input into a calculation module corresponding to each control variable, and an output value of the calculation module is input into the trained relationship model between the control variables and the operating parameters of the machine room control device to obtain first predicted values of the operating parameters; the target control variable is set to a second preset value while keeping the other control variables except the target control variable as 0 to obtain second predicted values of the operating parameters; the first predicted values and the second predicted values of the operating parameters are compared, and a difference value between the first predicted values and the second predicted values is obtained, and a corresponding operating parameter when the difference value is the largest is taken as the key operating parameter of the target control variable.

8. An electronic device, comprising: The device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete communication with each other through the communication bus; the memory is used to store a computer program; the processor is used to execute the program stored in the memory to implement the machine room control device operating parameter adjustment method in any one of claims 1-6.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the machine room control device operating parameter adjustment method in any one of claims 1-6.

Images