Efficient refrigerating machine room air conditioner energy-saving intelligent control system and method

Abstract

The invention discloses an efficient refrigerating machine room air conditioner energy-saving intelligent control system and method, and relates to the field of air conditioner intelligent control. The method comprises the steps that air conditioner operation parameters, machine room equipment operation data and machine room environment data are collected and preprocessed; building a deep neural network by using a deep Q-learning reinforcement learning algorithm to obtain a thermal load demand of the machine room; introducing a nonlinear dynamic compensation function to calculate an air-conditioner dynamic energy efficiency coefficient and air-conditioner refrigeration power, and calculating the rotating speed of an air-conditioner compressor by referring to air-conditioner operation parameters; a PID (Proportion Integration Differentiation) controller is used for carrying out primary frequency conversion on a machine room air conditioner, and a multi-stage adjusting mechanism is used for correcting a primary frequency conversion result in real time according to machine room equipment operation data, machine room environment data change and frequency conversion conditions. According to the air conditioner energy-saving control method, efficient air conditioner energy-saving control is achieved by combining the deep Q-learning reinforcement learning algorithm, the nonlinear dynamic compensation function and the PID controller.

Description

Technical Field

[0001] The present invention relates to the field of intelligent air conditioner control, and in particular to an energy-saving intelligent control system and method for an air conditioner in a high-efficiency refrigeration room. Background Art

[0002] With the rapid development of information technology, especially in computer room environments such as data centers and communication base stations, the energy consumption of air conditioning systems has become an important factor restricting computer room management costs and environmental sustainability. Traditional computer room air conditioning systems rely on preset temperature control strategies and fixed cooling capacity, usually maintaining indoor temperature through constant cooling power of air conditioning equipment. However, this method has a large energy efficiency loss when dealing with changes in computer room loads, environmental fluctuations, and dynamic equipment requirements. In recent years, with the development of the Internet of Things, artificial intelligence, and control technology, more and more research and practice have begun to focus on how to improve the operating efficiency of computer room air conditioning systems and reduce energy consumption through intelligent and dynamic adjustment methods. At present, air conditioning energy efficiency optimization technologies based on methods such as model predictive control (MPC), fuzzy control, and classical PID control have gradually been applied. However, most of these methods rely on static control rules or assumptions about system models, and it is difficult to adapt to changes in computer room heat load demand and fluctuations in air conditioning system status in real time in complex actual environments.

[0003] However, existing air conditioning control technologies still have many shortcomings in terms of dynamic energy efficiency optimization. Although the traditional PID control method performs well in some application scenarios, its inherent limitation is that it lacks sufficient adaptive capabilities and cannot make accurate real-time adjustments based on changes in environmental data and dynamic changes in the heat load demand of the computer room. Model-based control methods, such as MPC, can provide more accurate predictions and adjustments, but they rely on accurate system models and have high computational complexity, making it difficult to achieve rapid response in practical applications. Therefore, how to accurately adjust the air conditioning refrigeration power and compressor speed through real-time monitoring data analysis and heat load prediction to achieve intelligent energy-saving optimization of the computer room air conditioning system is still an urgent problem to be solved. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a high-efficiency energy-saving intelligent control method for refrigeration room air conditioners to solve the problem of being unable to make accurate real-time adjustments according to changes in environmental data and dynamic changes in the heat load demand of the room.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, the present invention provides an energy-saving intelligent control method for a high-efficiency refrigeration room air conditioner, which includes collecting air conditioning operating parameters, room equipment operating data and room environment data, and preprocessing the room equipment operating data and room environment data; based on the preprocessed room equipment operating data and room environment data, using a deep Q-learning reinforcement learning algorithm to build a deep neural network, constructing a heat load assessment model by fitting a Q value function, and obtaining the room heat load demand; according to the air conditioning operating parameters, room equipment operating data and room environment data, introducing a nonlinear dynamic compensation function to calculate the air conditioning dynamic energy efficiency coefficient, and then calculating the air conditioning refrigeration power through the air conditioning dynamic energy efficiency coefficient and the room heat load demand, and finally calculating the air conditioning compressor speed with reference to the air conditioning operating parameters; according to the air conditioning compressor speed, using a PID controller to perform preliminary frequency conversion on the room air conditioner, while monitoring the changes in the room equipment operating data and room environment data, using a multi-stage adjustment mechanism, and performing real-time correction on the preliminary frequency conversion results according to the changes in the room equipment operating data and room environment data and the frequency conversion conditions.

[0008] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: the air conditioning operation parameters, the room equipment operation data and the room environment data are collected, and the room equipment operation data and the room environment data are pre-processed, and the specific steps are:

[0009] The air conditioner operating parameters include the maximum energy efficiency coefficient of the air conditioner, the current speed of the air conditioner compressor and the maximum speed of the air conditioner compressor, which are collected through the sensor interface of the air conditioner;

[0010] The equipment room equipment operation data includes equipment power consumption, equipment load rate, equipment recommended operating temperature, equipment recommended operating humidity and equipment maximum load, which are collected through smart meters, power consumption metering equipment and equipment documents;

[0011] The computer room environment data includes the internal temperature of the computer room, the external temperature of the computer room, the internal humidity of the computer room, and the external humidity of the computer room, which are collected through temperature and humidity sensors;

[0012] The collected computer room equipment operation data and computer room environment data are cleaned, normalized and standardized, and timestamp calibrated, and transmitted to the data processing unit and stored in the database.

[0013] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: based on the pre-processed room equipment operation data and room environment data, a deep neural network is built using a deep Q-learning reinforcement learning algorithm, and a heat load evaluation model is constructed by fitting the Q value function to obtain the room heat load demand. The specific steps are:

[0014] Extract the pre-processed equipment operation data and environment data of the computer room from the database, and set the state space and action space;

[0015] According to the state space and action space, the Q value function is set. The expression of the Q value function is:

[0016]

[0017] Among them, Q(s,a) is the expected return of taking action a in state s, R(s,a) is the immediate reward after taking action a in state s, γ is the discount factor, is the maximum value after taking all possible actions a' in the next state s', where a is the action taken in the action space and s is the state in the state space. is the weighted average symbol;

[0018] A multi-layer perceptron structure is used to construct a feedforward neural network, the input features of the input layer, the number of layers and neurons of the hidden layer and the output layer are set, and the ReLU nonlinear activation function is used to improve the fitting ability of the feedforward neural network to build a heat load assessment model;

[0019] The heat load assessment model is trained using the Q-value function. The Q-value is updated through the error between the actual observed instant reward and the predicted value to output the heat load demand of the computer room.

[0020] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: the nonlinear dynamic compensation function is introduced to calculate the dynamic energy efficiency coefficient of the air conditioner according to the air conditioner operation parameters, the room equipment operation data and the room environment data, and the specific steps are:

[0021] Extract the internal temperature and humidity of the computer room from the computer room environment data, extract the recommended operating temperature and humidity of the equipment from the equipment operation data of the computer room, and extract the maximum energy efficiency coefficient of the air conditioner from the air conditioner operation parameters;

[0022] According to the above extracted parameters, the dynamic energy efficiency coefficient of the air conditioner is calculated, and the nonlinear dynamic compensation function expression is:

[0023]

[0024] Among them, η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time, η max is the maximum energy efficiency coefficient of the air conditioner, T p (t) is the internal temperature of the computer room at the current time, H p (t) is the humidity inside the computer room at the current time, T o Recommended operating temperature for the equipment, H o Recommended operating humidity for the device, α 1is the dynamic energy efficiency temperature adjustment coefficient, α 2 is the dynamic energy efficiency humidity adjustment coefficient, and t is the current time.

[0025] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: the air conditioning refrigeration power is calculated by the air conditioning dynamic energy efficiency coefficient and the heat load demand of the room, and finally the air conditioning compressor speed is calculated with reference to the air conditioning operating parameters. The specific steps are:

[0026] The air conditioning cooling power is calculated based on the dynamic energy efficiency coefficient of the air conditioner and the heat load demand of the computer room. The expression is:

[0027] P = E·η(t);

[0028] Among them, P is the cooling power of the air conditioner, E is the heat load demand of the computer room, and η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time;

[0029] The maximum speed of the air conditioner compressor is extracted from the air conditioner operating parameters. Combined with the heat load demand of the computer room, the cooling power of the air conditioner and the environmental data of the computer room, the speed of the air conditioner compressor is calculated through the speed adjustment function. The expression is:

[0030]

[0031] Where N is the air conditioning compressor speed, that is, the target speed, N max is the maximum speed of the air conditioning compressor, β 1 is the influence factor of temperature change on cooling power, β 2 is the nonlinear adjustment coefficient caused by temperature fluctuation, β 3 is the influence factor of humidity on cooling power.

[0032] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: according to the speed of the air conditioner compressor, the PID controller is used to perform preliminary frequency conversion on the room air conditioner, and the specific steps are:

[0033] The current speed of the air-conditioning compressor is extracted from the air-conditioning operating parameters, and the control signal is calculated using the PID controller to adjust the current speed of the air-conditioning compressor. The expression is:

[0034]

[0035] e(t)=NN now (t);

[0036] Among them, u(t) is the control signal, K 1 is the proportional coefficient, e(t) is the speed error at the current time t, N now (t) is the current speed of the air-conditioning compressor, K 2is the integration coefficient, is the error accumulation, τ is the time integral variable, K 3 is the differential coefficient, is the rate of change of error;

[0037] Adjust and optimize PID parameters through manual adjustment method;

[0038] The control signal is input to the inverter, and the inverter adjusts the speed of the compressor by changing the frequency of the current input to the compressor.

[0039] As a preferred solution of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner of the present invention, wherein: the multi-stage adjustment mechanism is used, and the specific steps are:

[0040] The operating status of the computer room is divided into the initial adjustment stage, the stable adjustment stage and the fine adjustment stage;

[0041] In the initial adjustment stage, the speed of the air conditioning compressor is adjusted to a large extent through the PID controller;

[0042] Reduce the adjustment range of the PID controller during the stable regulation stage;

[0043] In the fine adjustment stage, the speed of the air conditioning compressor is fine-tuned through the PID controller.

[0044] In a second aspect, the present invention provides an efficient energy-saving intelligent control system for refrigeration room air conditioners, including a data acquisition module, a heat load evaluation module, a compressor speed calculation module, and a PID multi-stage control module;

[0045] The data acquisition module is used to collect air-conditioning operating parameters, computer room equipment operating data and computer room environment data, and pre-process the computer room equipment operating data and computer room environment data;

[0046] The heat load assessment module is used to build a deep neural network based on the pre-processed equipment operation data and environment data of the computer room using a deep Q-learning reinforcement learning algorithm, and to construct a heat load assessment model by fitting a Q value function to obtain the heat load demand of the computer room;

[0047] The compressor speed calculation module is used to calculate the dynamic energy efficiency coefficient of the air conditioner by introducing a nonlinear dynamic compensation function according to the air conditioner operating parameters, the equipment operating data of the computer room and the computer room environment data, and then calculate the air conditioner refrigeration power by the air conditioner dynamic energy efficiency coefficient and the heat load demand of the computer room, and finally calculate the air conditioner compressor speed with reference to the air conditioner operating parameters;

[0048] The PID multi-stage control module is used to perform preliminary frequency conversion on the computer room air conditioner using a PID controller based on the speed of the air conditioner compressor, while monitoring changes in the computer room equipment operating data and computer room environmental data, and using a multi-stage adjustment mechanism to perform real-time corrections on the preliminary frequency conversion results based on changes in the computer room equipment operating data and computer room environmental data and frequency conversion conditions.

[0049] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the high-efficiency refrigeration room air conditioner energy-saving intelligent control method as described in the first aspect of the present invention is implemented.

[0050] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the high-efficiency refrigeration room air conditioner energy-saving intelligent control method as described in the first aspect of the present invention is implemented.

[0051] The beneficial effects of the present invention are as follows: by collecting and preprocessing air-conditioning operating parameters, computer room equipment data and environmental data, combined with deep Q-learning reinforcement learning algorithm, nonlinear dynamic compensation function and PID controller, an efficient computer room air-conditioning energy-saving intelligent control method is constructed. This method can dynamically optimize the operation of the air-conditioning system according to real-time environment and load changes, and accurately calculate the air-conditioning refrigeration power and compressor speed. By optimizing the heat load assessment through deep learning, combined with the dynamic energy efficiency coefficient and multi-stage adjustment mechanism, the problems of low energy efficiency and energy waste of traditional air-conditioning systems are avoided, and more efficient energy-saving control is achieved. It not only improves the adaptability and energy efficiency of the air conditioner, extends the life of the equipment, but also reduces the energy consumption of the computer room. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0053] Figure 1 This is a flow chart of the energy-saving intelligent control method for the high-efficiency refrigeration room air conditioner in Example 1.

[0054] Figure 2 This is a module diagram of the energy-saving intelligent control system for the high-efficiency refrigeration room air conditioner in Example 1. DETAILED DESCRIPTION

[0055] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0056] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0057] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0058] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, provides a high-efficiency refrigeration room air conditioner energy-saving intelligent control method, comprising the following steps:

[0059] S1: Collect air conditioning operation parameters, computer room equipment operation data and computer room environment data, and pre-process the computer room equipment operation data and computer room environment data.

[0060] Specifically, the following steps are included:

[0061] S1.1: The air conditioner operating parameters include the maximum energy efficiency coefficient of the air conditioner, the current speed of the air conditioner compressor and the maximum speed of the air conditioner compressor, which are collected through the sensor interface of the air conditioner.

[0062] Specifically, the maximum energy efficiency coefficient is obtained through the air conditioner operation manual or device firmware parameters as the theoretical maximum energy efficiency standard of the air conditioner. The compressor speed (current speed and maximum speed) is directly collected through the built-in sensor of the air conditioner and transmitted to the data processing unit through the data transmission interface.

[0063] S1.2: The equipment operation data of the computer room includes equipment power consumption, equipment load rate, equipment recommended operating temperature, equipment recommended operating humidity and equipment maximum load, which are collected through smart meters, power consumption metering equipment and equipment documents.

[0064] Specifically, by using smart meters or power consumption metering devices to collect the current, voltage and power data of the device in real time, accurate device power consumption is provided. The load rate of the device is calculated through the built-in sensor of the device or the ratio of power consumption to the maximum rated power. The load rate reflects the current working intensity of the device and affects the energy efficiency and operating status of the device. The recommended operating temperature and humidity of the device are extracted from the device's operating manual or the standard values ​​provided by the manufacturer, or collected through the built-in temperature and humidity sensor of the device. The recommended operating temperature and humidity are the best balance point for device performance, ensuring that the device is in an efficient operating state.

[0065] S1.3: The computer room environment data includes the temperature inside the computer room, the temperature outside the computer room, the humidity inside the computer room, and the humidity outside the computer room, which are collected through temperature and humidity sensors.

[0066] Specifically, temperature and humidity sensors are installed at different locations inside and outside the computer room to collect data. The temperature and humidity inside the computer room are used to adjust the state of the air conditioner, and the temperature and humidity outside the computer room are used to input the heat load assessment model to assess the impact of the external environment on the operation of the air conditioner in the computer room and assist in energy optimization.

[0067] S1.4: Perform data cleaning, data normalization and standardization, and time stamp calibration on the collected computer room equipment operation data and computer room environment data, and transmit them to the data processing unit and store them in the database.

[0068] It should be understood that data cleaning refers to removing outliers, missing values ​​and noise data to ensure the quality and effectiveness of data collection. Interpolation methods are used to fill missing data and remove unreasonable outliers.

[0069] Data normalization refers to scaling the collected equipment operation data and environment data of the computer room to a uniform range (such as [0,1]) to eliminate the impact between data of different dimensions.

[0070] Data standardization refers to converting the collected equipment room equipment operation data and room environment data into a standard normal distribution, that is, the mean is 0 and the standard deviation is 1. This is to prevent the training effect from being poor due to the large difference in data dimensions when constructing the heat load assessment model in the future, which will affect the generalization ability of the heat load assessment model.

[0071] Timestamp calibration refers to the time synchronization of different data sources to ensure data consistency due to the possible delay in data collection sources. Time synchronization technologies such as the Network Time Protocol (NTP) can help adjust the timestamps of each device so that data can be integrated according to the same timeline.

[0072] The collected data are transmitted to the data processing unit (central server or cloud platform) through the LAN connection and stored in the MySQL database.

[0073] Preferably, by collecting and preprocessing the air conditioner operating parameters, equipment operating data in the computer room, and environmental data in the computer room, it is possible to monitor in real time and accurately understand the operating status, load conditions, and environmental changes of the air conditioner and equipment. The maximum energy efficiency coefficient and compressor speed data of the air conditioner provide the basis for energy efficiency evaluation and optimization. The power consumption, load rate, and recommended operating temperature and humidity data of the equipment help improve the equipment operating efficiency and extend the equipment life, while the temperature and humidity data inside and outside the computer room optimize the operating status of the air conditioner. Data preprocessing ensures the quality and consistency of the collection, and provides high-quality data support for subsequent heat load evaluation and air conditioner energy efficiency optimization, ultimately achieving the effects of energy saving, improving energy efficiency, and enhancing stability.

[0074] S2: Based on the pre-processed equipment operation data and environment data of the computer room, a deep neural network is built using the deep Q-learning reinforcement learning algorithm. A heat load assessment model is constructed by fitting the Q value function to obtain the heat load demand of the computer room.

[0075] Specifically, the following steps are included:

[0076] S2.1: Extract the pre-processed computer room equipment operation data and computer room environment data from the database, and set the state space and action space.

[0077] Specifically, SQL query statements are used to extract pre-processed computer room equipment operation data and computer room environment data from the MySQL database.

[0078] The state space includes all relevant information about the current state of the computer room air conditioner. Each state is a description of the computer room environment and equipment operation status, including: temperature and humidity inside the computer room, temperature and humidity outside the computer room, equipment load rate, and equipment power consumption.

[0079] The action space includes all possible operations or adjustment measures that can be selected, including: air conditioning compressor speed adjustment, air conditioning temperature setting and equipment operating status adjustment.

[0080] S2.2: According to the state space and action space, set the Q value function. The expression of the Q value function is:

[0081]

[0082] Among them, Q(s,a) is the expected return of taking action a in state s, R(s,a) is the immediate reward after taking action a in state s, γ is the discount factor, is the maximum value after taking all possible actions a' in the next state s', where a is the action taken in the action space and s is the state in the state space. is the weighted average symbol.

[0083] S2.3: Use a multi-layer perceptron structure to build a feedforward neural network, set the input features of the input layer, the number of layers and neurons in the hidden layer and output layer, use the ReLU nonlinear activation function to improve the fitting ability of the feedforward neural network, and build a heat load assessment model.

[0084] Specifically, the feedforward neural network structure includes:

[0085] Input layer: The input layer is used to receive the equipment operation data and the environment data of the computer room, and pass it to the next layer of the network. The input layer receives the temperature and humidity inside the computer room, the temperature and humidity outside the computer room, the equipment load rate, the equipment power consumption, the current speed of the air conditioner compressor, and the maximum energy efficiency coefficient of the air conditioner; these input features will be used as the input vector of the network, and the number of neurons in the input layer is equal to the number of features.

[0086] Hidden layer: The hidden layer is the core of the multilayer perceptron, which contains multiple neurons and is responsible for extracting higher-level features from the input data. Each hidden layer neuron is connected to all neurons in the previous layer, and after weighted summation, a nonlinear transformation is performed using the ReLU activation function. The number of hidden layers is 2, and the number of neurons in each layer is determined based on experimental tuning during the training process.

[0087] Output layer: The number of neurons in the output layer is equal to the number of actions in the action space, and the output layer outputs the heat load demand of the computer room.

[0088] Furthermore, the ReLU function can effectively avoid the gradient vanishing problem and has strong nonlinear fitting capabilities, making it suitable for processing complex environments and device data relationships.

[0089] S2.4: Use the Q-value function to train the heat load assessment model, update the Q-value through the error between the actual observed instant reward and the predicted value, and output the heat load demand of the computer room.

[0090] Preferably, the Q value function is used to train the heat load assessment model, and the Q value is updated by the error between the actual observed instant reward and the predicted value. The expression is:

[0091]

[0092] Among them, Q(s t ,a t ) is the Q value at the current time, θ is the learning rate, which controls the step size of each update, R(s t ,a t ) is the instantaneous reward at the current time, i.e., the actual heat load demand, is the optimal Q value at the next time;

[0093] The weights and biases of the feedforward neural network are randomly initialized, and the weights of the feedforward neural network are adjusted through the back-propagation algorithm. Through multiple iterations, the Q value of the feedforward neural network becomes more and more accurate, and finally the heat load demand of the computer room air conditioner can be effectively evaluated.

[0094] Finally, by using the trained heat load assessment model and inputting new computer room environment data, the network can output a heat load prediction value. This prediction value is the heat load demand required by the current computer room, which is used for the adjustment and optimization of air conditioning equipment.

[0095] Preferably, the heat load demand of the computer room air conditioner is accurately evaluated and optimized by combining a multi-layer perceptron neural network with a deep Q-learning reinforcement learning algorithm. By extracting pre-processed computer room environmental data from the database and setting the state space and action space, it is possible to intelligently learn and adjust the air conditioner operating parameters, such as compressor speed and temperature settings, to improve energy efficiency and stability. Reinforcement learning trains the model through the Q-value function to achieve automated decision-making, continuously optimize heat load assessment, and dynamically adjust the control strategy based on real-time data. This method not only improves the operating efficiency, energy saving effect and equipment life of the air conditioner, but also can flexibly adapt to the changing environment and equipment status, reducing the operating cost of the computer room and improving environmental stability.

[0096] S3: According to the air conditioner operating parameters, the equipment operating data of the computer room and the computer room environment data, a nonlinear dynamic compensation function is introduced to calculate the dynamic energy efficiency coefficient of the air conditioner. Then, the air conditioner cooling power is calculated by the dynamic energy efficiency coefficient of the air conditioner and the heat load demand of the computer room. Finally, the air conditioner compressor speed is calculated by referring to the air conditioner operating parameters.

[0097] Specifically, the following steps are included:

[0098] S3.1: Extract the internal temperature and humidity of the computer room from the computer room environment data, extract the recommended operating temperature and humidity of the equipment from the equipment operation data of the computer room, and extract the maximum energy efficiency coefficient of the air conditioner from the air conditioner operation parameters.

[0099] S3.2: Based on the above extracted parameters, the dynamic energy efficiency coefficient of the air conditioner is calculated. The expression of the nonlinear dynamic compensation function is:

[0100]

[0101] Among them, η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time, η max is the maximum energy efficiency coefficient of the air conditioner, T p (t) is the internal temperature of the computer room at the current time, H p (t) is the humidity inside the computer room at the current time, T o Recommended operating temperature for the equipment, H oRecommended operating humidity for the device, α 1 is the dynamic energy efficiency temperature adjustment coefficient, α 2 is the dynamic energy efficiency humidity adjustment coefficient, and t is the current time.

[0102] Preferably, the dynamic energy efficiency coefficient of the air conditioner reflects the working efficiency of the air conditioner under specific conditions. Ideally, the air conditioner should adjust its energy efficiency according to the temperature and humidity of the computer room and the recommended working conditions of the equipment. If the temperature and humidity of the computer room deviate from the recommended values, it will affect the working efficiency of the air conditioner. Therefore, it is necessary to adjust the energy efficiency coefficient of the air conditioner through a nonlinear dynamic compensation function. This function will dynamically adjust the energy efficiency coefficient of the air conditioner based on the current temperature and humidity deviation (i.e., the difference between the current temperature and the recommended temperature, and the difference between the humidity and the recommended humidity). When the temperature and humidity deviation of the computer room is too large, the energy efficiency coefficient of the air conditioner will decrease, which means that the air conditioner needs to consume more energy to maintain the same cooling effect, that is, the energy efficiency of the air conditioner deteriorates and the cooling effect decreases. Therefore, the air conditioner will optimize its energy efficiency according to real-time environmental changes to adapt to the heat load requirements of the computer room.

[0103] S3.3: Calculate the air conditioning cooling power based on the dynamic energy efficiency coefficient of the air conditioner and the heat load demand of the computer room. The expression is:

[0104] P = E·η(t);

[0105] Among them, P is the cooling power of the air conditioner, E is the heat load demand of the computer room, and η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time.

[0106] It should be understood that the heat load demand should be combined with the air conditioning energy efficiency to ensure that the air conditioner provides sufficient cooling power at the most appropriate energy efficiency state to meet the actual needs of the computer room.

[0107] S3.4: Extract the maximum speed of the air conditioner compressor from the air conditioner operating parameters, combine the heat load demand of the computer room, the cooling power of the air conditioner and the computer room environment data, and calculate the speed of the air conditioner compressor through the speed adjustment function. The expression is:

[0108]

[0109] Where N is the air conditioning compressor speed, that is, the target speed, N max is the maximum speed of the air conditioning compressor, β 1 is the influence factor of temperature change on cooling power, β 2 is the nonlinear adjustment coefficient caused by temperature fluctuation, β 3 is the influence factor of humidity on cooling power.

[0110] Preferably, by combining the effects of temperature and humidity, the air conditioner can achieve more precise dynamic regulation, ensuring that the optimal air conditioner compressor speed is calculated under different environmental conditions.

[0111] Preferably, the energy efficiency coefficient of the air conditioner is adjusted by introducing a nonlinear dynamic compensation function, and the cooling power and compressor speed of the air conditioner are optimized by combining the heat load demand of the computer room, the dynamic energy efficiency coefficient of the air conditioner and the computer room environment data. The dynamic energy efficiency coefficient is adjusted in real time according to the temperature and humidity deviation of the computer room to ensure that the air conditioner maintains efficient operation under different environmental conditions. By incorporating the influence of temperature and humidity into the speed adjustment function, the speed of the air conditioner compressor is accurately calculated to achieve the best cooling effect and energy efficiency. This optimization mechanism improves the adaptive ability of the air conditioner, reduces energy waste, ensures the stability of the computer room environment, and improves the overall operating efficiency of the air conditioner.

[0112] S4: Based on the air-conditioning compressor speed, use the PID controller to perform preliminary frequency conversion on the computer room air conditioner. At the same time, monitor the changes in the computer room equipment operation data and computer room environment data. Use a multi-stage adjustment mechanism to make real-time corrections to the preliminary frequency conversion results based on the changes and frequency conversion conditions of the computer room equipment operation data and computer room environment data.

[0113] Specifically, the following steps are included:

[0114] S4.1: Extract the current speed of the air-conditioning compressor from the air-conditioning operating parameters, use the PID controller to calculate the control signal, and adjust the current speed of the air-conditioning compressor. The expression is:

[0115]

[0116] e(t)=NN now (t);

[0117] Among them, u(t) is the control signal, K 1 is the proportional coefficient, e(t) is the speed error at the current time t, N now (t) is the current speed of the air-conditioning compressor, K 2 is the integration coefficient, is the error accumulation, τ is the time integral variable, K 3 is the differential coefficient, is the rate of change of error.

[0118] Specifically, K 1 Used to control the proportional effect of error on output. This coefficient adjusts the response speed of the air conditioner speed. K 2 Used to control the cumulative impact of historical errors on output. By integrating and accumulating errors, long-term deviations can be reduced. 3 Used to control the impact of the error change rate. This coefficient helps predict the error trend so that adjustments can be made in advance.

[0119] S4.2: Adjust and optimize PID parameters through manual adjustment method.

[0120] Specifically, adjust the PID parameters step by step based on actual operation data and response. First, adjust the proportional coefficient K 1 , then adjust the integral coefficient K 2 To eliminate the steady-state error, finally adjust the differential coefficient K 3 To reduce overshoot and improve response speed. Considering that the room environment may have large temperature fluctuations, the differential term K of the PID controller 3 Relatively small values ​​may be needed to avoid overscaling.

[0121] S4.3: The control signal is input to the inverter, and the inverter adjusts the speed of the compressor by changing the frequency of the current input to the compressor.

[0122] Specifically, the inverter adjusts the speed of the compressor by changing the frequency of the current input to the compressor, thereby adjusting the cooling capacity of the air conditioner. According to the control signal output by the PID, the inverter converts the frequency adjustment into a voltage signal and outputs it to the air conditioner compressor, thereby affecting its speed.

[0123] S4.4: The operating status of the computer room is divided into the preliminary adjustment stage, the stable adjustment stage and the fine adjustment stage.

[0124] Preferably, by dividing the air-conditioning operation into different adjustment stages and dynamically adjusting the parameters of the PID controller according to the changes in the computer room environment data, more flexible and accurate air-conditioning control can be achieved. The multi-stage adjustment mechanism can ensure that the air-conditioning can operate efficiently under different working conditions, thereby reducing energy waste, improving the stability of the computer room environment, and further improving the overall energy efficiency and stability of the air-conditioning operation.

[0125] S4.4.1: During the initial adjustment phase, the speed of the air conditioning compressor is adjusted to a large extent through the PID controller.

[0126] Specifically, when the heat load demand E of the computer room is greater than 80% of the maximum load of the equipment, the temperature fluctuation exceeds 2°C, and the humidity fluctuation exceeds 10%, the preliminary adjustment stage is entered. The goal is to quickly respond to the fluctuation of temperature and humidity in the computer room and prevent the temperature and humidity from exceeding the set thresholds to protect the equipment in the computer room.

[0127] Furthermore, increasing K 1 , so that the PID controller can respond quickly to fluctuations in temperature or humidity. This can help the air conditioner make larger speed adjustments when the load changes. 2 , reduce long-term errors and ensure that temperature and humidity can quickly return to a stable range under large fluctuations. 3 , in order to improve the response to the rate of change of temperature and humidity, avoid overreaction of the control signal and prevent oscillation.

[0128] Increasing the amplitude of the control signal allows the air conditioner compressor speed to be adjusted to the required level more quickly to cope with changes in the room's heat load. By increasing the adjustment amplitude, the air conditioner compressor speed is controlled to ensure that the air conditioner maintains appropriate cooling capacity when the load changes greatly.

[0129] S4.4.2: Reduce the adjustment range of the PID controller during the stable regulation phase.

[0130] Specifically, when the heat load demand of the computer room is 50% ≤ E ≤ 80% of the maximum load of the equipment, the temperature change rate is less than 0.1°C per minute, and the humidity change rate is less than 0.5% per minute, it enters the stable adjustment stage. The goal is to maintain stable temperature and humidity in the computer room, reduce the energy consumption of air conditioning, and maintain a balance between cooling capacity and heat load demand.

[0131] Furthermore, reducing K 1 , which slows down the response speed of the control signal and avoids over-regulation. 2 , avoid over-reliance on historical errors and prevent the controller from entering an "over-compensation" state. 3 , reduce excessive response to the rate of change of temperature and humidity, and avoid frequent control signal fluctuations.

[0132] The change range of the control signal becomes smaller, and the speed of the compressor is gradually adjusted to a more stable state. By reducing the speed adjustment range, the air conditioner's overreaction to environmental changes is reduced, energy consumption is saved, and the room environment is stable.

[0133] S4.4.3: During the fine adjustment stage, the speed of the air conditioning compressor is fine-tuned through the PID controller.

[0134] Specifically, when the heat load demand E of the computer room is less than 50% of the maximum load of the equipment, the temperature fluctuation is less than 0.5°C, and the humidity fluctuation is less than 2%, the fine adjustment stage is entered. The goal is to make detailed adjustments to the air conditioner so that it can maintain the stability of the temperature and humidity of the computer room with minimum energy consumption under low load conditions.

[0135] Furthermore, we can further reduce K 1 , ensuring fine adjustment to small changes in temperature and humidity. Keeping the K low 2 , to avoid over-response to the accumulation of long-term errors. Increase K 3 , quickly respond to subtle changes in temperature and humidity, ensuring that temperature and humidity stability can be accurately maintained under light load.

[0136] Minimize the change of control signal, and adjust the speed of air conditioner compressor only within a very small range to accurately match the actual heat load demand of the computer room. At this stage, slight speed adjustment is mainly used to ensure that the air conditioner can maintain the best energy efficiency under the condition that the computer room environment reaches a stable state.

[0137] Preferably, by adopting PID controller and multi-stage adjustment mechanism, the air conditioner can respond to the changes in the room environment in real time, realize the precise control of temperature and humidity, optimize the energy efficiency of the air conditioner, reduce energy consumption and reduce operating costs. Dynamically adjust the air conditioner compressor speed according to the room environment data to ensure stable operation under different load conditions, reduce equipment loss and extend service life. In addition, the adjustment of PID parameters improves the response speed and adaptive ability of the air conditioner, effectively reduces the impact of environmental fluctuations on the equipment, and ensures that the air conditioner always operates in the best working state through fine adjustment, which improves the reliability, maintainability and overall performance of the air conditioner.

[0138] The present embodiment also provides an efficient energy-saving intelligent control system for refrigerating computer room air conditioners, including: a data acquisition module, which is used to collect air conditioner operating parameters, computer room equipment operating data and computer room environmental data, and pre-process the computer room equipment operating data and computer room environmental data; a heat load assessment module, which is used to build a deep neural network based on the pre-processed computer room equipment operating data and computer room environmental data using a deep Q-learning reinforcement learning algorithm, and to construct a heat load assessment model by fitting a Q value function to obtain the computer room heat load demand; a compressor speed calculation module, which is used to calculate the dynamic energy efficiency coefficient of the air conditioner according to the air conditioner operating parameters, computer room equipment operating data and computer room environmental data, and then calculate the air conditioner refrigeration power according to the dynamic energy efficiency coefficient of the air conditioner and the computer room heat load demand, and finally calculate the air conditioner compressor speed with reference to the air conditioner operating parameters; a PID multi-stage control module, which is used to perform preliminary frequency conversion on the computer room air conditioner using a PID controller according to the air conditioner compressor speed, and at the same time monitor the changes in the computer room equipment operating data and computer room environmental data, and use a multi-stage adjustment mechanism to make real-time corrections to the preliminary frequency conversion results according to the changes in the computer room equipment operating data and computer room environmental data and the frequency conversion situation.

[0139] This embodiment also provides a computer device, which is suitable for the case of a high-efficiency refrigeration room air conditioner energy-saving intelligent control method, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the high-efficiency refrigeration room air conditioner energy-saving intelligent control method proposed in the above embodiment.

[0140] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0141] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for realizing energy-saving intelligent control of a high-efficiency refrigeration room air conditioner as proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0142] In summary, the present invention designs an efficient intelligent control method for energy saving of computer room air conditioners by collecting and preprocessing air conditioner operating parameters, computer room equipment data and environmental data, combining deep Q-learning reinforcement learning algorithm, nonlinear dynamic compensation function and PID controller. This method can dynamically optimize the operation of the air conditioning system according to real-time environment and load changes, and accurately calculate the air conditioning refrigeration power and compressor speed. Through deep learning to optimize the thermal load assessment, combined with the dynamic energy efficiency coefficient and multi-stage adjustment mechanism, the problems of low energy efficiency and energy waste of traditional air conditioning systems are avoided, and more efficient energy-saving control is achieved. It not only improves the adaptability and energy efficiency of the air conditioner, extends the equipment life, but also reduces the energy consumption of the computer room.

[0143] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A high-efficiency energy-saving intelligent control method for air conditioners in refrigeration rooms, characterized in that: include, Collect air-conditioning operation parameters, computer room equipment operation data and computer room environment data, and pre-process the computer room equipment operation data and computer room environment data; Based on the pre-processed equipment operation data and environment data of the computer room, a deep neural network is built using the deep Q-learning reinforcement learning algorithm. A heat load assessment model is constructed by fitting the Q value function to obtain the heat load demand of the computer room. According to the air conditioner operating parameters, equipment operating data of the computer room and the computer room environment data, a nonlinear dynamic compensation function is introduced to calculate the dynamic energy efficiency coefficient of the air conditioner. Then, the air conditioner cooling power is calculated by the dynamic energy efficiency coefficient of the air conditioner and the heat load demand of the computer room. Finally, the air conditioner compressor speed is calculated by referring to the air conditioner operating parameters. According to the speed of the air-conditioning compressor, a PID controller is used to perform preliminary frequency conversion on the computer room air conditioner. At the same time, the changes in the computer room equipment operation data and the computer room environment data are monitored. A multi-stage adjustment mechanism is used to make real-time corrections to the preliminary frequency conversion results based on the changes and frequency conversion conditions of the computer room equipment operation data and the computer room environment data.

2. The high-efficiency energy-saving intelligent control method for refrigeration room air conditioner according to claim 1, characterized in that: The collecting of air-conditioning operation parameters, equipment operation data of the computer room and environment data of the computer room, and preprocessing of the equipment operation data of the computer room and environment data of the computer room, specifically comprises the following steps: The air conditioner operating parameters include the maximum energy efficiency coefficient of the air conditioner, the current speed of the air conditioner compressor and the maximum speed of the air conditioner compressor, which are collected through the sensor interface of the air conditioner; The equipment room equipment operation data includes equipment power consumption, equipment load rate, equipment recommended operating temperature, equipment recommended operating humidity and equipment maximum load, which are collected through smart meters, power consumption metering equipment and equipment documents; The computer room environment data includes the internal temperature of the computer room, the external temperature of the computer room, the internal humidity of the computer room, and the external humidity of the computer room, which are collected through temperature and humidity sensors; The collected computer room equipment operation data and computer room environment data are cleaned, normalized and standardized, and timestamp calibrated, and transmitted to the data processing unit and stored in the database.

3. The high-efficiency energy-saving intelligent control method for refrigeration room air conditioner according to claim 2, characterized in that: Based on the pre-processed equipment operation data and environment data of the computer room, a deep neural network is built using a deep Q-learning reinforcement learning algorithm, and a heat load evaluation model is constructed by fitting the Q value function to obtain the heat load demand of the computer room. The specific steps are: Extract the pre-processed equipment operation data and environment data of the computer room from the database, and set the state space and action space; According to the state space and action space, the Q value function is set. The expression of the Q value function is: Among them, Q(s, a) is the expected return of taking action a in state s, R(s, a) is the immediate reward after taking action a in state s, γ is the discount factor, is the maximum value after taking all possible actions a' in the next state s', where a is the action taken in the action space and s is the state in the state space. is the weighted average symbol; A multi-layer perceptron structure is used to construct a feedforward neural network, the input features of the input layer, the number of layers and neurons of the hidden layer and the output layer are set, and the ReLU nonlinear activation function is used to improve the fitting ability of the feedforward neural network to build a heat load assessment model; The heat load assessment model is trained using the Q-value function. The Q-value is updated through the error between the actual observed instant reward and the predicted value to output the heat load demand of the computer room.

4. The high-efficiency energy-saving intelligent control method for a refrigeration room air conditioner according to claim 3, characterized in that: The nonlinear dynamic compensation function is introduced to calculate the dynamic energy efficiency coefficient of the air conditioner according to the air conditioner operation parameters, the equipment operation data of the computer room and the computer room environment data. The specific steps are: Extract the internal temperature and humidity of the computer room from the computer room environment data, extract the recommended operating temperature and humidity of the equipment from the equipment operation data of the computer room, and extract the maximum energy efficiency coefficient of the air conditioner from the air conditioner operation parameters; According to the above extracted parameters, the dynamic energy efficiency coefficient of the air conditioner is calculated, and the nonlinear dynamic compensation function expression is: Among them, η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time, η max is the maximum energy efficiency coefficient of the air conditioner, T p (t) is the internal temperature of the computer room at the current time, H p (t) is the humidity inside the computer room at the current time, T o Recommended operating temperature for the equipment, H o is the recommended operating humidity for the equipment, α1 is the dynamic energy efficiency temperature adjustment coefficient, α2 is the dynamic energy efficiency humidity adjustment coefficient, and t is the current time.

5. The high-efficiency energy-saving intelligent control method for a refrigeration room air conditioner according to claim 4, characterized in that: The air conditioning refrigeration power is calculated by the air conditioning dynamic energy efficiency coefficient and the heat load demand of the computer room, and finally the air conditioning compressor speed is calculated by referring to the air conditioning operation parameters. The specific steps are: The air conditioning cooling power is calculated based on the dynamic energy efficiency coefficient of the air conditioner and the heat load demand of the computer room. The expression is: P = E·η(t); Where P is the cooling power of the air conditioner, E is the heat load demand of the computer room, and η(t) is the dynamic energy efficiency coefficient of the air conditioner at the current time; The maximum speed of the air conditioner compressor is extracted from the air conditioner operating parameters. Combined with the heat load demand of the computer room, the cooling power of the air conditioner and the environmental data of the computer room, the speed of the air conditioner compressor is calculated through the speed adjustment function. The expression is: Where N is the air conditioning compressor speed, that is, the target speed, N max is the maximum speed of the air-conditioning compressor, β1 is the influence factor of temperature change on the cooling power, β2 is the nonlinear adjustment coefficient caused by temperature fluctuation, and β3 is the influence factor of humidity on the cooling power.

6. The high-efficiency energy-saving intelligent control method for a refrigeration room air conditioner according to claim 5, characterized in that: The PID controller is used to perform preliminary frequency conversion on the computer room air conditioner according to the air conditioner compressor speed. The specific steps are as follows: The current speed of the air-conditioning compressor is extracted from the air-conditioning operating parameters, and the control signal is calculated using the PID controller to adjust the current speed of the air-conditioning compressor. The expression is: e(t)=N-N now (t); Among them, u(t) is the control signal, K1 is the proportional coefficient, e(t) is the speed error at the current time t, N now (t) is the current speed of the air-conditioning compressor, K2 is the integral coefficient, is the error accumulation, τ is the integral variable of time, K3 is the differential coefficient, is the rate of change of error; Adjust and optimize PID parameters through manual adjustment method; The control signal is input to the inverter, and the inverter adjusts the speed of the compressor by changing the frequency of the current input to the compressor.

7. The high-efficiency energy-saving intelligent control method for a refrigeration room air conditioner according to claim 6, characterized in that: The multi-stage adjustment mechanism is used, and the specific steps are: The operating status of the computer room is divided into the initial adjustment stage, the stable adjustment stage and the fine adjustment stage; In the initial adjustment stage, the speed of the air conditioning compressor is adjusted to a large extent through the PID controller; Reduce the adjustment range of the PID controller during the stable regulation stage; In the fine adjustment stage, the speed of the air conditioning compressor is fine-tuned through the PID controller.

8. A high-efficiency refrigeration room air conditioner energy-saving intelligent control system, based on the high-efficiency refrigeration room air conditioner energy-saving intelligent control method according to any one of claims 1 to 7, characterized in that: Including data acquisition module, heat load assessment module, compressor speed calculation module and PID multi-stage control module; The data acquisition module is used to collect air-conditioning operating parameters, computer room equipment operating data and computer room environment data, and pre-process the computer room equipment operating data and computer room environment data; The heat load assessment module is used to build a deep neural network based on the pre-processed equipment operation data and environment data of the computer room using a deep Q-learning reinforcement learning algorithm, and to construct a heat load assessment model by fitting a Q value function to obtain the heat load demand of the computer room; The compressor speed calculation module is used to calculate the dynamic energy efficiency coefficient of the air conditioner by introducing a nonlinear dynamic compensation function according to the air conditioner operating parameters, the equipment operating data of the computer room and the computer room environment data, and then calculate the air conditioner refrigeration power by the air conditioner dynamic energy efficiency coefficient and the heat load demand of the computer room, and finally calculate the air conditioner compressor speed with reference to the air conditioner operating parameters; The PID multi-stage control module is used to perform preliminary frequency conversion on the computer room air conditioner using a PID controller based on the speed of the air conditioner compressor, while monitoring changes in the computer room equipment operating data and computer room environmental data, and using a multi-stage adjustment mechanism to perform real-time corrections on the preliminary frequency conversion results based on changes in the computer room equipment operating data and computer room environmental data and frequency conversion conditions.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the high-efficiency refrigeration room air conditioner energy-saving intelligent control method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the high-efficiency refrigeration room air conditioner energy-saving intelligent control method according to any one of claims 1 to 7 are implemented.

Images